Configuration Scheme Selection Method

Abstract

A configuration scheme selection method includes: obtaining measurement results from a measurement target, the measurement target being positioned on a semiconductor wafer; obtaining measurement results from an in-device target, the in-device target being positioned on the semiconductor wafer; and using both the measurement target measurement results and the in-device measurement measurement results to determine a configuration scheme for accurate measurement.

Description

Technical Field

[0001] The present specification relates to methods of determining information about a patterning process, such as values ​​of parameters of interest, such as overlay. The specification also relates to methods of reducing errors in measurement data, methods of calibrating a metrology apparatus, and methods of selecting a metrology target to be used in a metrology process. Background Art

[0002] A lithographic apparatus is a machine that applies a desired pattern to a substrate (usually, to a target portion of a substrate). A lithographic apparatus can be used, for example, for the manufacture of an integrated circuit (IC) or other device designed to be functional. In this example, a patterning device (which may alternatively be referred to as a mask or mask) can be used to generate a circuit pattern to be formed on a single layer of a device designed to be functional. The pattern can be transferred to a target portion (e.g., a portion including one or several dies) on a substrate (e.g., a silicon wafer). The transfer of the pattern is usually performed via imaging onto a layer of a radiation-sensitive material (resist) disposed on the substrate. In general, a single substrate will contain a network of adjacent target portions that are continuously patterned. Known lithographic apparatuses include so-called steppers, in which each target portion is irradiated by exposing the entire pattern to the target portion at one time; and so-called scanners, in which the pattern is scanned by a radiation beam along a given direction (the "scanning" direction), while simultaneously scanning the substrate in parallel or antiparallel to the direction to irradiate each target portion. It is also possible to transfer a pattern from a patterning device to a substrate by imprinting the pattern onto the substrate. Summary of the invention

[0003] Manufacturing equipment (such as semiconductor equipment) generally includes: using multiple manufacturing processes to process a substrate (e.g., a semiconductor wafer) to form various features, and often multiple layers of the equipment. Typically, such layers and / or features are manufactured and processed using, for example, deposition, lithography, etching, chemical mechanical polishing, and ion implantation. Multiple devices can be made on multiple dies on a substrate and then separated into individual devices. The device manufacturing process can be considered a patterning process. The patterning process includes: a pattern transfer step of providing a pattern on a substrate, such as optical and / or nanoimprint lithography using a lithography device; and typically but optionally includes: one or more related pattern process steps, such as resist development by a developing device, baking the substrate using a baking tool, etching the pattern by an etching device, etc. Further, one or more measurement processes are included in the patterning process.

[0004] A metrology process is used at various steps during the patterning process to monitor and / or control the process. For example, the metrology process is used to measure one or more characteristics of the substrate, such as the relative position (e.g., registration, overlay, alignment, etc.) or size (e.g., line width, critical dimension (CD), thickness, etc.) of features formed on the substrate during the patterning process, so that, for example, the performance of the patterning process can be determined based on the one or more characteristics. If one or more characteristics are unacceptable (e.g., outside a predetermined range of one or more characteristics), one or more variables of the patterning process can be designed or changed, for example, based on the measurement of the one or more characteristics, so that the substrate manufactured by the patterning process has one or more acceptable characteristics.

[0005] With the development of lithography and other patterning process technologies, the size of functional elements continues to decrease, while the number of functional elements (such as transistors) in each device has steadily increased over decades. At the same time, the requirements for precision in terms of overlay, critical dimension (CD), etc. have become increasingly stringent. Errors such as overlay error, CD error, etc. are inevitably generated in the patterning process. For example, imaging errors can be generated by optical aberrations, heated patterning devices, patterning device errors, and / or heated substrates, and can be characterized in terms of, for example, overlay, CD, etc. Additionally or alternatively, errors can be introduced in other parts of the patterning process (such as in etching, developing, baking, etc.), and can also be characterized in terms of, for example, overlay, CD, etc. Errors may cause problems with the function of the device, including one or more electrical problems with the device that cannot operate or a functional device. Therefore, it is desirable to be able to characterize one or more of these errors and adopt patterning processes such as design, modification, and control to reduce or minimize the steps of one or more errors in these errors.

[0006] According to one aspect, a method for determining information about a patterning process is provided, the method comprising: obtaining measurement data from a measurement process of each measurement target among a plurality of measurement targets applied to a substrate, wherein: the measurement data for each measurement target comprises at least a first contribution and a second contribution, the first contribution being from a parameter of interest in a patterning process for forming the measurement target, and the second contribution being from an error in the measurement process for measuring the measurement target; and the method further comprising: obtaining information about the error in the measurement process using the measurement data obtained from all the measurement targets among the plurality of measurement targets; and extracting a value of the parameter of interest for each measurement target using the obtained information about the error in the measurement process.

[0007] According to one aspect, a method for reducing errors in measurement data is provided, the method comprising: obtaining measurement data from a measurement process for each measurement target among a plurality of measurement targets applied to a substrate, wherein: the measurement data for each measurement target comprises at least a first contribution and a second contribution, the first contribution coming from a parameter of interest of a patterning process for forming the measurement target, and the second contribution coming from an error in the measurement process for measuring the measurement target; and the method further comprising: using the measurement data obtained from all the measurement targets among the plurality of measurement targets to obtain information about the error in the measurement process; and using the obtained information to reduce the error in the measurement data.

[0008] According to one aspect, a method for calibrating a measurement process is provided, the method comprising: obtaining measurement data from a measurement process for each measurement target among a plurality of measurement targets applied to a substrate, wherein: the measurement data for each measurement target comprises at least a first contribution and a second contribution, the first contribution being from a parameter of interest of a patterning process for forming the measurement target, and the second contribution being from an error in the measurement process for measuring the measurement target; and the method further comprising: using the measurement data obtained from all the measurement targets among the plurality of measurement targets to obtain information about the error in the measurement process; and using the obtained information to calibrate the measurement process to reduce the error in the measurement data generated by the measurement process.

[0009] According to one aspect, a method for selecting measurement targets for use in determining information about a patterning process is provided, wherein: the method for determining information about the patterning process includes: obtaining measurement data from a measurement process applied to each of a plurality of selected measurement targets; and using the obtained measurement data to extract values ​​of one or more parameters of interest of the patterning process used to form each measurement target; and the measurement targets are selected so that the number of measurement targets is greater than the number of parameters of interest. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Embodiments will now be described, by way of example only, with reference to the accompanying drawings, in which

[0011] Figure 1 One embodiment of a lithographic apparatus is schematically depicted;

[0012] Figure 2 One embodiment of a lithographic cell or cluster is schematically depicted;

[0013] Figure 3 (a) is a schematic diagram of a measurement apparatus for measuring a target using a first pair of illumination apertures providing certain illumination patterns according to one embodiment;

[0014] Figure 3(b) is a schematic detail of the diffraction spectrum of a target for a given illumination direction;

[0015] Figure 3 (c) is a schematic diagram of a second pair of illumination apertures providing an alternative illumination mode when using the measurement device to perform diffraction-based overlay measurement;

[0016] Figure 3 (d) is a schematic diagram of a third pair of illumination apertures in combination with the first pair of apertures and the second pair of apertures providing other illumination modes when using the measurement device for diffraction-based overlay measurement;

[0017] Figure 4 schematically depicts the form of a plurality of periodic structure (e.g., a plurality of grating) targets and the contours of the measurement spots on the substrate;

[0018] Figure 5 Schematically depicts the Figure 3 obtained from the device Figure 4 An image of the target;

[0019] Figure 6 schematically depicts example metrology apparatus and metrology techniques;

[0020] Figure 7 schematically depicts an example metrology device;

[0021] Figure 8 The diagram shows the relationship between the illumination spot of the measuring device and the measuring target;

[0022] Fig. 9 schematically depicts a process for deriving one or more variables of interest based on measured data;

[0023] Fig. 10A schematically depicts an example unit cell, an associated pupil representation, and an associated derived pupil representation;

[0024] Fig. 10B schematically depicts an example unit cell, an associated pupil representation, and an associated derived pupil representation;

[0025] Fig. 10C schematically depicts an example target comprising one or more physical instances of a unit cell;

[0026] Fig.11 A high-level process for obtaining weightings for determining patterning process parameters based on measured radiation is depicted;

[0027] Fig.12 A high-level process for determining patterning process parameters based on measured radiation is depicted;

[0028] Fig.13Describes a high-level flow of one embodiment of a data-driven technique;

[0029] Fig.14 Describes the high-level flow of one embodiment of a data-driven technique incorporating physical geometry models;

[0030] Fig.15 Describes the high-level flow of one embodiment of a data-driven technique incorporating physical geometry models;

[0031] Fig.16 Describes the high-level flow of one embodiment of a data-driven technique incorporating physical geometry models;

[0032] Fig.17 Describes the high-level flow of one embodiment of a data-driven technique incorporating physical geometry models;

[0033] Fig.18 One embodiment of a plurality of overlay unit cells of a target is schematically depicted;

[0034] Fig.19 One embodiment of a multi-element unit cell of a target is schematically depicted;

[0035] Fig. 20 An example illustration of two vectors corresponding to two different overlays is depicted;

[0036] Fig.21 A method of determining information about a patterning process is described;

[0037] Fig. 22 Depicted in Fig.21 A plurality of metrology targets on a substrate usable in the method;

[0038] Fig.23 Schematically depicts the Fig.21 the measured intensity in the detected pupil representation of radiation redirected by the first metrology target used in the method;

[0039] Fig.24 Schematically depicts the Fig.21 the measured intensity in the detected pupil representation of radiation redirected by a second metrology target used in the method;

[0040] Fig.25 A computer system is schematically depicted on which embodiments of the present disclosure may be implemented. DETAILED DESCRIPTION

[0041] Before describing the embodiments in detail, it is instructive to present an example environment in which the embodiments may be implemented.

[0042] Figure 1A lithographic apparatus LA is schematically depicted. The apparatus comprises:

[0043] - an illumination system (illuminator) IL configured to condition a radiation beam B (e.g. UV radiation or DUV radiation);

[0044] a support structure (e.g., mask table) MT constructed to support a patterning device (e.g., mask) MA and connected to a first positioner PM configured to accurately position the patterning device according to certain parameters;

[0045] a substrate table (e.g. wafer stage) WT configured to hold a substrate (e.g. resist coated wafer) W and connected to a second positioner PW configured to accurately position the substrate according to certain parameters; and

[0046] - a projection system (e.g., a refractive projection lens system) PS configured to project the pattern imparted to the radiation beam B by the patterning device MA onto a target portion C (e.g., comprising one or more dies) of the substrate W, the projection system being supported on a reference frame (RF).

[0047] The illumination system may include various types of optical components, such as refractive, reflective, magnetic, electromagnetic, electrostatic, or other types of optical components, or any combination thereof, for directing, shaping, or controlling radiation.

[0048] The support structure supports the patterning device in a manner that depends on the orientation of the patterning device, the design of the lithographic apparatus, and other conditions, such as, for example, whether the patterning device is held in a vacuum environment. The support structure can use mechanical, vacuum, electrostatic, or other clamping techniques to hold the patterning device. The support structure can, for example, be a frame or a stage, which can be fixed or movable as desired. The support structure can ensure that the patterning device is in a desired position, for example, relative to a projection system. Any use of the term "reticle" or "mask" herein may be considered synonymous with the more general term "patterning device".

[0049] The term "patterning device" as used herein should be broadly interpreted as referring to any device that can be used to impart a pattern in a target portion of a substrate. In one embodiment, the patterning device is any device that can be used to impart a pattern in a cross-section of a radiation beam so as to form a pattern in the target portion of the substrate. It should be noted that the pattern imparted to the radiation beam may not exactly correspond to the desired pattern in the target portion of the substrate, for example if the pattern includes phase-shifting features or so-called assist features. Typically, the pattern imparted to the radiation beam will correspond to a specific functional layer in a device produced in the target portion such as an integrated circuit.

[0050] Patterning devices may be transmissive or reflective. Examples of patterning devices include masks, programmable mirror arrays, and programmable LCD panels. Masks are well known in lithography and include mask types such as binary, alternating phase shift, and attenuated phase shift, as well as various hybrid mask types. An example of a programmable mirror array employs a matrix arrangement of small mirrors, each of which can be individually tilted so as to reflect an incident radiation beam in different directions. The tilted mirrors impart a pattern in the radiation beam, which is reflected by the matrix of mirrors.

[0051] The term "projection system" as used herein should be broadly interpreted to cover any type of projection system including refractive, reflective, catadioptric, magnetic, electromagnetic, and electrostatic optical systems, or any combination thereof, depending on the exposure radiation being used or other factors (such as the use of an immersion liquid or the use of a vacuum). Any use of the term "projection lens" herein may be considered synonymous with the more general term "projection system".

[0052] The projection system PS has an optical transfer function that may be non-uniform, which affects the pattern imaged onto the substrate W. For unpolarized radiation, this effect can be well described by two scalar diagrams that describe the transmittance (apodization) and relative phase (aberration) of the radiation emerging from the projection system PS as a function of its position in the pupil plane. These scalar diagrams, which may be referred to as transmittance diagrams and relative phase diagrams, can be expressed as linear combinations of a complete set of basis functions. A particularly convenient set is the Zernike polynomials, which form a set of orthogonal polynomials defined on the unit circle. The determination of each scalar diagram may involve determining the coefficients in such an expansion. Since the Zernike polynomials are orthogonal on the unit circle, the Zernike coefficients can be determined by calculating the inner product of the measured scalar diagram using each Zernike polynomial in turn and dividing it by the square of the norm of the Zernike polynomial.

[0053] The transmittance diagram and the relative phase diagram depend on the field and the system. That is, in general, each projection system PS has a different Zernike expansion for each field point (i.e., for each spatial position in its image plane). The relative phase of the projection system PS in its pupil plane can be determined by, for example, projecting radiation from a point source in the object plane of the projection system PS (i.e., the plane of the patterning device MA) through the projection system PS and using a shearing interferometer for measuring the wavefront (i.e., the trajectory of points with the same phase). The shearing interferometer is a common path interferometer and therefore advantageously, no secondary reference beam is required to measure the wavefront. The shearing interferometer may include a diffraction grating (e.g., a two-dimensional grid) in the image plane of the projection system (i.e., the substrate table WT) and a detector arranged to detect an interference pattern in a plane conjugate with the pupil plane of the projection system PS. The interference pattern is related to the derivative of the phase of the radiation in the shear direction with respect to the coordinates in the pupil plane. The detector may include an array of sensing elements (e.g., a charge coupled device (CCD)).

[0054] The projection system PS of the lithographic apparatus may not produce visible fringes, so a phase step technique (such as, for example, moving a diffraction grating) can be used to enhance the accuracy of the wavefront determination. Stepping can be performed in the plane of the diffraction grating and in a direction perpendicular to the scanning direction of the measurement. The stepping range can be one grating period, and at least three (evenly distributed) phase steps can be used. Thus, for example, three scanning measurements can be performed in the y direction, each for a different position in the x direction. Such stepping of the diffraction grating effectively transforms the phase changes into intensity changes, thereby determining the phase information. The grating can be stepped in a direction perpendicular to the diffraction grating (z direction) to calibrate the detector.

[0055] The transmittance (apodization) of the projection system PS in its pupil plane can be determined by the projection system PS projecting radiation, for example from a point source in the object plane of the projection system PS (i.e. the plane of the patterning device MA) and measuring the intensity of the radiation in a plane conjugate to the pupil plane of the projection system PS using a detector. The same detectors as those used to measure the wavefront to determine the aberrations can be used.

[0056] The projection system PS may include a plurality of optical (e.g., lens) elements, and may further include an adjustment mechanism AM configured to adjust one or more of the optical elements in order to correct aberrations (phase variations across the pupil plane in the entire field). To achieve this, the adjustment mechanism is 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 in which its optical axis extends in the z direction. The adjustment mechanism is operable to perform any combination of the following: displacing one or more optical elements; tilting one or more optical elements; and / or deforming one or more optical elements. The optical element may be displaced in any direction (x, y, z, or a combination thereof). Although rotation about the z-axis may be used for non-rotationally symmetric aspheric optical elements, by rotating about the axis in the x and / or y directions, the optical element is typically tilted outside a plane perpendicular to the optical axis. The deformation of the optical element may include low-frequency shapes (e.g., astigmatism) and / or high-frequency shapes (e.g., free-form aspheric surfaces). The deformation of the optical element may be performed, for example, by using one or more actuators that apply forces on one or more sides of the optical element and / or by using one or more heating elements that heat one or more selected areas of the optical element. In general, it may not be possible to adjust the projection system PS to correct for apodization (transmittance variations across the pupil plane). When designing a patterning device (e.g., mask) MA for the lithographic apparatus LA, a transmittance map of the projection system PS may be used. Using computational lithography techniques, the patterning device MA may be designed to at least partially correct for apodization.

[0057] As depicted herein, the device is a transmissive type device (eg, employing a transmissive mask). Alternatively, the device may be a reflective type device (eg, employing a programmable mirror array of the type described above, or employing a reflective mask).

[0058] The lithographic apparatus may be of a type having two (dual stage) or more tables (e.g. two or more substrate tables WTa, WTb, two or more patterning device tables, a substrate table WTa and a table WTb located below the projection system, without a dedicated substrate table for example to facilitate measurement and / or cleaning etc.). In such a "multi-stage" machine, the additional tables may be used in parallel, or preparatory steps may be performed on one or more tables while one or more other tables are being used for exposure. For example, alignment measurements using an alignment sensor AS and / or level (height, tilt etc.) measurements using a level sensor LS may be performed.

[0059] The lithographic apparatus may also be of a type in which at least a portion of the substrate may be covered by a liquid having a relatively high refractive index (e.g., water) to fill the space between the projection system and the substrate. Immersion liquid may also be applied to other spaces in the lithographic apparatus, for example, between the patterning device and the projection system. Immersion techniques are well known in the art for increasing the numerical aperture of a projection system. The term "immersion" as used herein does not mean that structures such as substrates must be submerged in the liquid, but only means that the liquid is located between the projection system and the substrate during exposure.

[0060] refer to Figure 1 , the illuminator IL receives a radiation beam from a radiation source SO. The source and the lithographic apparatus may be separate entities, for example when the source is an excimer laser. In this case, the source is not considered to be part of the lithographic apparatus, and the radiation beam is transferred from the source SO to the illuminator IL with the aid of a beam delivery system BD comprising, for example, suitable directing mirrors and / or a beam expander. In other cases, for example when the source is a mercury lamp, the source may be an integral part of the lithographic apparatus. The source SO and the illuminator IL and, if necessary, the beam delivery system BD together are referred to as a radiation system.

[0061] The illuminator IL may include an adjuster AD configured to adjust the angular intensity distribution of the radiation beam. Typically, at least the outer radial extent and / or the inner radial extent (commonly referred to as σ-outer and σ-inner, respectively) of the intensity distribution in a pupil plane of the illuminator may be adjusted. In addition, the illuminator IL may include various other components, such as an integrator IN and a condenser CO. The illuminator may be used to condition the radiation beam to have a desired uniformity and intensity distribution in its cross-section.

[0062] The radiation beam B is incident on a patterning device (e.g. a mask) MA, which is held on a support structure (e.g. a mask table) MT and is patterned by the patterning device. After traversing the patterning device MA, the radiation beam B passes through a projection system PS, which focuses the beam onto a target portion C of the substrate W. With the aid of a second positioner PW and a position sensor IF (e.g. an interferometer device, a linear encoder, a two-dimensional encoder or a capacitive sensor), the substrate table WT can be precisely moved, for example, in order to position a different target portion C in the path of the radiation beam B. Similarly, the first positioner PM and the further position sensor ( Figure 1The support structure MT may be used to accurately position the patterning device MA with respect to the path of the radiation beam B, for example after mechanical retrieval from a mask library or during scanning. In general, movement of the support structure MT may be achieved with the aid of a long-stroke module (coarse positioning) and a short-stroke module (fine positioning) which form part of the first positioner PM. Similarly, movement of the substrate table WT may be achieved using a long-stroke module and a short-stroke module which form part of the second positioner PW. In the case of a stepper (as opposed to a scanner), the support structure MT may be connected to the short-stroke actuator only, or may be fixed. The patterning device MA and the substrate W may be aligned using patterning device alignment marks M1, M2 and substrate alignment marks P1, P2. Although the illustrated substrate alignment marks occupy dedicated target portions, they may be located in the space between target portions (these are called scribe alignment marks). Similarly, where more than one die is provided on the patterning device MA, the patterning device alignment marks may be located between the die.

[0063] The depicted device can be used in at least one of the following modes:

[0064] 1. In step mode, the support structure MT and the substrate table WT remain substantially stationary and the entire pattern imparted to the radiation beam is projected at once onto a target portion C (i.e. a single static exposure). The substrate table WT is then shifted in the X and / or Y direction so that a different target portion C can be exposed. In step mode, the maximum size of the exposure field limits the size of the target portion C imaged in a single static exposure.

[0065] 2. In scan mode, the support structure MT and substrate table WT are scanned synchronously while a pattern imparted to the radiation beam is projected onto a target portion C (i.e. a single dynamic exposure). The speed and direction of the substrate table WT relative to the support structure MT may be determined by the (or less) magnification and image reversal characteristics of the projection system PS. In scan mode, the maximum size of the exposure field limits the width (in the non-scanning direction) of the target portion in a single dynamic exposure, while the length of the scanning movement determines the height (in the scanning direction) of the target portion.

[0066] 3. In another mode, the support structure MT remains substantially stationary, thereby holding the programmable patterning device, and the substrate table WT is moved or scanned while a pattern imparted to the radiation beam is projected onto a target portion C. In this mode, the programmable patterning device is updated and the pulsed radiation source is employed as required after each movement of the substrate table WT or between successive radiation pulses during a scan. This mode of operation may be readily adapted for maskless lithography utilizing a programmable patterning device, such as a programmable mirror array of the type described above.

[0067] Combinations and / or variations on the above-described modes of use or entirely different modes of use may also be employed.

[0068] like Figure 2 As shown, the lithography apparatus LA may form part of a lithography cell LC, which is sometimes also referred to as a lithography cell or cluster, and which also includes apparatus for performing pre-exposure and post-exposure processes on a substrate. Conventionally, these apparatuses include one or more spin coaters SC for depositing one or more resist layers, one or more developers DE for developing the exposed resist, one or more chill plates CH and / or one or more baking plates BK. A substrate handler or robot RO picks up one or more substrates from input / output ports I / O1, I / O2, moves them between different processing apparatuses, and then transfers them to a feed station LB of the lithography apparatus. These apparatuses, which are often collectively referred to as tracks, are under the control of a track control unit TCU, which itself is controlled by a management control system SCS, which also controls the lithography apparatus via a lithography control unit LACU. Thus, different apparatuses may be operated to maximize throughput and processing efficiency.

[0069] In order to correctly and consistently expose a substrate exposed by a lithographic apparatus, it is desirable to inspect the exposed substrate to measure or determine one or more properties, such as overlay (which may be, for example, between structures in overlapping layers, or between structures in the same layer, the latter structures having been separately provided to that layer by, for example, a double patterning process), line thickness, critical dimension (CD), focus offset, material properties, etc. Thus, the manufacturing facility in which the lithocell LC is located typically also comprises a metrology system MET, which receives part or all of the substrates W that have been processed in the lithocell. The metrology system MET may be part of the lithocell LC, for example, it may be part of the lithographic apparatus LA.

[0070] The metrology results may be provided directly or indirectly to the supervisory control system SCS. If an error is detected, the exposure of subsequent substrates (especially if the check can be completed soon and quickly enough so that one or more other substrates of the batch are still to be exposed) and / or the subsequent exposure of the exposed substrate may be adjusted. In addition, already exposed substrates may be stripped and reprocessed to improve yield, or discarded to avoid further processing on substrates known to be defective. In the event that only some target portions of the substrate are defective, further exposures may be performed only on those good target portions.

[0071] Within the metrology system MET, a metrology device is used to determine one or more properties of a substrate, in particular, how one or more properties of different substrates vary or how different layers of the same substrate vary from layer to layer. The metrology device may be integrated into the lithography apparatus LA or the lithography cell LC, or may be a stand-alone device. In order to enable fast measurements, it is desirable that the metrology device measures one or more properties in the exposed resist layer immediately after exposure. However, the contrast of the latent image in the resist is very low—there is only a small difference in the refractive index between the parts of the resist that have been exposed to radiation and the parts that have not been exposed to radiation—and not all metrology devices are sensitive enough to make useful measurements of the latent image. Therefore, the measurement can be made after a post-exposure bake step (PEB), which is usually the first step performed on the exposed substrate and increases the contrast between the exposed and unexposed parts of the resist. At this stage, the image in the resist can be referred to as a semi-latent image. It is also possible to measure the resist image after development—at which point the exposed or unexposed parts of the resist have been removed—or after a pattern transfer step (such as etching). The latter possibility limits the possibilities for reworking defective substrates but can still provide useful information.

[0072] To enable the measurement, one or more targets may be provided on the substrate. In one embodiment, the target is specially designed and may include a periodic structure. In one embodiment, the target is part of a device pattern, for example, a periodic structure of the device pattern. In one embodiment, the device pattern is a periodic structure of a memory device (e.g., a bipolar transistor (BPT), a bit line contact (BLC), etc. structure).

[0073] In one embodiment, the target on the substrate may include one or more 1D periodic structures (e.g., gratings) that are printed such that after development, the periodic structural features are formed by solid resist lines. In one embodiment, the target may include one or more 2D periodic structures (e.g., gratings) that are printed such that after development, the one or more periodic structures are formed by solid resist posts or vias in the resist. The bars, posts, or vias may alternatively be etched into the substrate (e.g., etched into one or more layers on the substrate).

[0074] In one embodiment, a parameter of interest in the patterning process is overlay. Overlay can be measured using dark field scatterometry, in which the zero-order diffraction (corresponding to specular reflection) is blocked and only higher-order diffraction is processed. Examples of dark field measurements can be found in PCT patent application publications WO 2009 / 078708 and WO 2009 / 106279, the entire contents of which are hereby incorporated by reference herein. Other developments of the technology have been described in U.S. patent application publications US2011-0027704, US2011-0043791 and US2012-0242970, the entire contents of which are hereby incorporated by reference herein. Diffraction-based overlay using dark field detection of diffraction orders enables overlay measurements on smaller targets. These targets can be smaller than the irradiation spot and may be surrounded by device product structures on the substrate. In one embodiment, multiple targets can be measured in one radiation capture.

[0075] Figure 3 In (a) a metrology apparatus suitable for measuring, for example, overlay in one embodiment is schematically shown. Figure 3 The target T (comprising a periodic structure such as a grating) and the diffracted light are shown in more detail in (b). The metrology apparatus may be a stand-alone device or may be incorporated into a lithography apparatus LA, for example at a measurement station, or into a lithography cell LC. The optical axis - which has a number of branches throughout the apparatus - is indicated by the dotted line O. In the apparatus, radiation emitted by an output 11 (e.g. a source such as a laser or a xenon lamp, or an opening connected to a source) is directed via a prism 15 onto a substrate W by an optical system comprising lenses 12, 14 and an objective lens 16. These lenses are arranged in a dual sequence of a 4F arrangement. A different lens arrangement may be used as long as it still provides an image of the substrate onto the detector.

[0076] In one embodiment, the lens arrangement allows access to the intermediate pupil plane for spatial frequency filtering. Therefore, the angular range of radiation incident on the substrate can be selected by defining the spatial intensity distribution in a plane presenting the spatial spectrum of the substrate plane (referred to herein as the (conjugate) pupil plane). Specifically, this can be performed in the plane of the back-projected image of the objective lens pupil plane, for example by inserting an aperture plate 13 of a suitable form between lenses 12 and 14. In the illustrated example, the aperture plate 13 has different forms marked as 13N and 13S, thereby allowing different illumination modes to be selected. The illumination system in this example forms an off-axis illumination mode. In a first illumination mode, the aperture plate 13N provides off-axis illumination from a direction designated as "north" (for descriptive purposes only). In a second illumination mode, the aperture plate 13S is used to provide similar illumination, but from the opposite direction marked as "south". Other illumination modes are also possible by using different holes. Desirably, the rest of the pupil plane is dark, because any unnecessary radiation outside the desired illumination mode may interfere with the desired measurement signal.

[0077] like Figure 3 As shown in (b), the target T is placed with a substrate W substantially perpendicular to the optical axis O of the objective lens 16. The illumination light I that impinges on the target T at a certain angle off the axis O produces a zero-order ray (solid line 0) and two first-order rays (dotted chain line +1 and double-dotted chain line -1). In the case of an overfilled small target T, these rays are just one of many parallel rays that cover the substrate area (including the measurement target T and other features). Due to the limited width of the aperture in the plate 13 (a useful amount of radiation must be allowed), the incident light I actually occupies a certain angle range, and the diffracted rays 0 and +1 / -1 are slightly spread out. According to the point spread function of the small target, each order +1 and -1 is further spread out over a certain angle range, rather than a single ideal beam as shown. Note that the pitch of the periodic structure and the illumination angle can be designed or adjusted so that the first-order rays entering the objective lens are closely aligned with the central optical axis. Figure 3 (a) and Figure 3 The rays illustrated in (b) are shown somewhat off-axis purely to make them easier to distinguish in the figure. At least the 0th and +1st orders diffracted by the target on substrate W are collected by objective lens 16 and directed back through prism 15.

[0078] Return to Figure 3(a), both the first illumination mode and the second illumination mode are illustrated by specifying diametrically opposed apertures labeled as north (N) and south (S). When the incident light ray I comes from the north side of the optical axis, that is, when the first illumination mode is applied using aperture plate 13N, the +1 diffracted light ray labeled as +1 (N) enters objective lens 16. In contrast, when the second illumination mode is applied using aperture plate 13S, the -1 order diffracted light ray (labeled as -1 (S)) is the light that enters lens 16. Therefore, in one embodiment, the measurement result can be obtained by measuring the target twice under certain conditions (for example, after rotating the target or changing the illumination mode or changing the imaging mode to obtain the -1 diffraction order intensity and the +1 diffraction order intensity, respectively). Comparing these intensities for a given target provides a measure of the asymmetry in the target, and the asymmetry in the target can be used as an indicator of a parameter (for example, overlay) of the lithography process. In the situation described above, the illumination mode is changed.

[0079] The beam splitter 17 splits the diffracted beam into two measurement branches. In the first measurement branch, the optical system 18 uses the zero-order diffracted beam and the first-order diffracted beam to form a diffraction spectrum (pupil plane image) of the target on a first sensor 19 (e.g., a CCD or CMOS sensor). Each diffraction order hits a different point on the sensor so that image processing can compare and contrast the orders. The pupil plane image captured by the sensor 19 can be used for focusing a metrology device and / or normalized intensity measurement. As further described below, the pupil plane image can also be used for other measurement purposes, such as reconstruction.

[0080] In the second measurement branch, the optical system 20, 22 forms an image of the target on the substrate W on a sensor 23 (e.g., a CCD or CMOS sensor). In the second measurement branch, an aperture stop 21 is set in a plane conjugate to the pupil plane of the objective lens 16. The aperture stop 21 is used to block the zero-order diffracted beam, so that the image of the target formed on the sensor 23 is formed by the -1 or +1 first-order beam. Data about the image measured by the sensors 19 and 23 are output to the processor and controller PU, whose function will depend on the specific measurement type performed. Note that the term 'image' is used in a broad sense. If only one of the -1 and +1 orders is present, an image of such periodic structural features (e.g., grating lines) will not be formed.

[0081] Figure 3 The specific forms of aperture plate 13 and diaphragm 21 shown are examples only. In another embodiment, on-axis illumination of the target is used, and an aperture diaphragm with an off-axis aperture is used to pass substantially only one first-order diffracted radiation to the sensor. In still other embodiments, second-order beams, third-order beams and higher-order beams ( Figure 3 not shown).

[0082] To adapt the illumination to these different types of measurements, the aperture plate 13 may include several patterns of holes formed around the circumference of a disk that rotates to place the desired pattern in place. Note that aperture plates 13N or 13S are used to measure periodic structures of a target oriented in one direction (X or Y, depending on the setup). To measure orthogonal periodic structures, target rotations of 90° and 270° can be achieved. Figure 3 (c) and Figure 3 Different aperture plates are illustrated in (d). Figure 3 (c) illustrates two other types of off-axis illumination modes. Figure 3 In the first illumination mode of (c), aperture plate 13E provides off-axis illumination from a direction designated as 'East' relative to the previously described 'North' (for illustration purposes only). Figure 3 In the second illumination mode of (c), aperture plate 13W is used to provide similar illumination but from the opposite direction marked 'West'. Figure 3 (d) illustrates two other types of off-axis illumination modes. Figure 3 In the first illumination mode (d), aperture plate 13NW provides off-axis illumination from the directions designated as 'North' and 'West' as previously described. In the second illumination mode, aperture plate 13SE is used to provide similar illumination, but from the opposite directions labeled 'South' and 'East' as previously described. The use of these devices and many other variations and applications of the devices are described in, for example, the prior published patent application publications mentioned above.

[0083] Figure 4An example composite metrology target T formed on a substrate is depicted. The composite target comprises four periodic structures (gratings in this case) 32, 33, 34, 35 positioned closely together. In one embodiment, the periodic structure layout can be made smaller than the measurement spot (i.e., the periodic structure layout is overfilled). Thus, in one embodiment, the periodic structures are positioned closely enough together that they are all within the measurement spot 31 formed by the illumination beam of the metrology device. Thus, in that case, all four periodic structures are illuminated simultaneously and imaged simultaneously on the sensors 19 and 23. In an example dedicated to overlay measurement, the periodic structures 32, 33, 34, 35 are themselves composite periodic structures (e.g., composite gratings) formed by overlapping periodic structures, i.e., periodic structures are patterned in different layers of a device formed on a substrate W, and such that at least one periodic structure in one layer overlays with at least one periodic structure in a different layer. The outer dimensions of such a target may be within 20 μm×20 μm or within 16 μm×16 μm. Further, all periodic structures are used to measure the overlay between a particular pair of layers. To facilitate the target being able to measure more than a single pair of layers, the periodic structures 32, 33, 34, 35 may have overlay offsets with different biases to facilitate measuring the overlay between different layers forming different parts of the composite periodic structure. Thus, all periodic structures for a target on a substrate will be used to measure the layers of one pair, and all periodic structures for another identical target on a substrate may be used to measure the layers of another pair, with different biases facilitating differentiation between the layer pairs.

[0084] Return to Figure 4 As shown, the orientation of the periodic structures 32, 33, 34, 35 may also be different so as to diffract the incident radiation in the X and Y directions. In one example, the periodic structures 32 and 34 are X-direction periodic structures with biases of +d and -d, respectively. The periodic structures 33 and 35 may be Y-direction periodic structures with biases of +d and -d, respectively. Although four periodic structures are shown, another embodiment may include a larger matrix to obtain the desired accuracy. For example, the biases of a 3×3 array of nine composite periodic structures may be -4d, -3d, -2d, -d, 0, +d, +2d, +3d, +4d. Individual images of these periodic structures may be identified in the image captured by the sensor 23.

[0085] Figure 5 shows an example of an image that can be used Figure 3 In the device Figure 4 The goal is to use Figure 3(d) The aperture plate 13NW or 13SE is formed on the sensor 23 and detected by the sensor 23. Although the sensor 19 cannot resolve the different individual periodic structures 32 to 35, the sensor 23 can resolve them. The dark rectangle represents the field of the image on the sensor, within which the irradiation spot 31 on the substrate is imaged into the corresponding circular area 41. Within the circular area 41, the rectangular areas 42-45 represent the images of the periodic structures 32 to 35. Instead of being located in or in addition to the scribe lanes, the target can be located in the device product features. If the periodic structure is located in the device product area, the device features are also visible at the periphery of the image field. The processor and controller PU use pattern recognition to process these images to identify individual images 42 to 45 of the periodic structures 32 to 35. In this way, it is not necessary to align the image very accurately at a specific position within the sensor frame, which greatly improves the overall throughput of the measuring device.

[0086] Once individual images of a periodic structure have been identified, the intensities of those individual images can be measured, for example, by averaging or summing selected pixel intensity values ​​within the identified region. The intensities and / or other properties of the images can be compared to one another. These results can be combined to measure different parameters of the lithographic process. Overlay performance is an example of such a parameter.

[0087] In one embodiment, one parameter of interest in the patterning process is feature width (eg, CD). Figure 6 A highly schematic example metrology apparatus (e.g., a scatterometer) that may enable determination of feature widths is depicted. The example metrology apparatus comprises a broadband (white light) radiation projector 2 that projects radiation onto a substrate W. The redirected radiation is passed to a spectrometer detector 4 that measures a spectrum 10 (intensity as a function of wavelength) of the specularly reflected radiation, as shown, for example, in the lower left diagram. From this data, a processor PU may (e.g., by rigorous coupled wave analysis and nonlinear regression or by comparing with, for example, Figure 6The structure or profile of the detected spectrum is reconstructed using a simulated spectral library (compared to the simulated spectral library shown in the lower right of ) to produce the detected spectrum. In general, for the reconstruction, the general form of the structure is known, and some variables are assumed from knowledge of the process of making the structure, leaving some variables in the structure to be determined from the measurement data. This measurement device can be configured as a normal incidence measurement device or an oblique incidence measurement device. Moreover, in addition to measuring parameters by reconstruction, angular resolved scattering measurements are useful in measuring the asymmetry of features in products and / or resist patterns. Asymmetry measurements are particularly useful for measuring overlays, where the target includes a set of periodic features superimposed on another set of periodic features. The concept of asymmetry measurements performed in this manner is described, for example, in U.S. Patent Application Publication US2006-066855, the entire contents of which are incorporated herein.

[0088] Figure 7 An example of a metrology device 100 suitable for use in the embodiments of the invention disclosed herein is illustrated. The operating principle of this type of metrology device is explained in more detail in U.S. Patent Application Nos. US2006-033921 ​​and US 2010-201963, the entire contents of which are incorporated herein by reference. An optical axis with several branches throughout the device is represented by a dotted line O. In the device, radiation emitted by a source 110 (e.g., a xenon lamp) is directed onto a substrate W by an optical system comprising a lens system 120, an aperture plate 130, a lens system 140, a partially reflective surface 150, and an objective lens 160. In one embodiment, these lens systems 120, 140, 160 are arranged in a dual sequence of a 4F arrangement. In one embodiment, the lens system 120 is used to collimate the radiation emitted by the radiation source 110. If desired, different lens arrangements can be used. The angular range of radiation incident on the substrate can be selected by defining a spatial intensity distribution in a plane presenting a spatial spectrum of the substrate plane. In particular, this can be done by inserting an aperture plate 130 of suitable form between the lenses 120 and 140 in the plane of the back-projected image of the pupil plane of the objective. By using different apertures, different intensity distributions (e.g., annular, dipole, etc.) are possible. The angular distribution of the illumination in the radial and peripheral directions as well as properties such as the wavelength, polarization and / or coherence of the radiation can all be adjusted to obtain the desired result. For example, one or more interference filters 130 can be arranged between the source 110 and the partially reflecting surface 150 (see Fig. 9 ) to select the wavelength of interest in the range of, for example, 400 nm to 900 nm or even smaller, for example, 200 nm to 300 nm. The interference filter may be tunable, rather than comprising a set of different filters. A grating may be used instead of an interference filter. In one embodiment, one or more polarizers 170 (see FIG. 1 ) may be provided between the source 110 and the partially reflective surface 150. Fig. 9 ) to select the polarization of interest. The polarizer may be tunable rather than comprising a set of different polarizers.

[0089] like Figure 7 As shown, the target T is placed together with the substrate W perpendicular to the optical axis O of the objective lens 160. Therefore, the radiation from the source 110 is reflected by the partially reflective surface 150 and focused via the objective lens 160 into an illumination spot S on the target T on the substrate W (see Figure 8 In one embodiment, the objective lens 160 has a high numerical aperture (NA), which is desirably at least 0.9 or at least 0.95. The numerical aperture of an immersion measurement device (using a fluid with a relatively high refractive index, such as water) may even be greater than 1.

[0090] Illumination rays 170, 172 focused from an angle off axis O onto the illumination spot produce diffracted rays 174, 176. It should be remembered that these rays are only one of many parallel rays covering the area of ​​the substrate including the target T. Every element within the illumination spot is within the field of view of the metrology device. Due to the finite width of the aperture in plate 130 (necessary to allow for a useful amount of radiation), the incident rays 170, 172 in fact occupy a range of angles, and the diffracted rays 174, 176 will be somewhat spread out. Each diffraction order is further spread out over a range of angles, according to the point spread function of a small target, as opposed to a single ideal ray as shown.

[0091] The objective lens 160 collects at least the 0th order diffracted by the target on the substrate W and guides it back through the partially reflective surface 150. The optical element 180 provides at least a portion of the diffracted beam to the optical system 182, which uses the zeroth order diffracted beam and / or the first order diffracted beam to form a diffraction spectrum (pupil plane image) of the target T on the sensor 190 (e.g., CCD or CMOS sensor). In one embodiment, an aperture 186 is provided to filter out certain diffraction orders so that specific diffraction orders are provided to the sensor 190. In one embodiment, the aperture 186 substantially or mainly allows only zeroth order radiation to reach the sensor 190. In one embodiment, the sensor 190 can be a two-dimensional detector so that a two-dimensional angular scattering spectrum of the substrate target T can be measured. The sensor 190 can be an array of, for example, CCD or CMOS sensors, and can use, for example, an integration time of 40 milliseconds per frame. The sensor 190 can be used to measure the intensity of the redirected radiation at a single wavelength (or a narrow wavelength range), the intensity at multiple wavelengths respectively, or the intensity integrated over a certain wavelength range. Furthermore, the sensor may be arranged to measure the intensity of radiation having transverse magnetic polarization and / or transverse electric polarization and / or the phase difference between transverse magnetic polarization radiation and transverse electric polarization radiation, respectively.

[0092] Optionally, the optical element 180 provides at least a portion of the diffracted beam to a measurement branch 200 for forming an image of a target on the substrate W on a sensor 230 (e.g. a CCD or CMOS sensor). The measurement branch 200 may be used for various auxiliary functions, such as focusing a metrology device (i.e. enabling the substrate W to be focused using the objective lens 160), and / or for dark field imaging of the type mentioned in the introduction.

[0093] In order to provide a customized field of view for gratings of different sizes and shapes, an adjustable field stop 300 is provided in the path from the source 110 to the objective lens 160 within the lens system 140. The field stop 300 includes an aperture 302 and is located in a plane conjugate to the plane of the target T so that the illumination spot is an image of the aperture 302. The image can be scaled according to the magnification, or the aperture and illumination spot can be in a 1:1 size relationship. In order to enable the illumination to be suitable for different types of measurements, the aperture plate 300 can include a number of aperture patterns formed around a disk that rotates to place the desired pattern in place. Alternatively or in addition, a set of plates 300 can be set and swapped to achieve the same effect. Additionally or alternatively, a programmable aperture device such as a deformable mirror array or a transmissive spatial light modulator can also be used.

[0094] Typically, the target will be aligned with its periodic structural features parallel to the Y axis or parallel to the X axis. With respect to diffraction behavior, a periodic structure having features extending in a direction parallel to the Y axis has periodicity in the X direction, while a periodic structure having features extending in a direction parallel to the X axis has periodicity in the Y direction. In order to measure performance in two directions, two types of features are typically provided. Although reference will be made to lines and spaces for simplicity, a periodic structure need not be formed by lines and spaces. Moreover, each line and / or space between lines may be a structure formed by smaller substructures. Further, a periodic structure having periodicity in two dimensions may be formed at once, for example, where the periodic structure includes pillars and / or through holes.

[0095] Figure 8 The diagram shows a typical target T plan and Figure 7 The range of the illumination spot S in the device is . In order to obtain a diffraction spectrum without interference from surrounding structures, the target T is in one embodiment a periodic structure (e.g., a grating) that is larger than the width (e.g., diameter) of the illumination spot S. The width of the spot S can be smaller than the width and length of the target. In other words, the illumination "underfills" the target, and the diffraction signal is substantially unaffected by any signals from product features, etc., outside the target itself. This simplifies the mathematical reconstruction of the target, as it can be considered infinite.

[0096] Fig. 9An example process of determining values ​​of one or more variables of interest for a target pattern 30' based on measurement data obtained using metrology is schematically depicted.The radiation detected by the detector 190 provides a measured radiation distribution 108 for the target 30'.

[0097] For a given target 30', a radiation distribution 208 may be calculated / simulated from a parameterized mathematical model 206 using, for example, a numerical Maxwell solver 210. The parameterized mathematical model 206 shows example layers of various materials that make up and are associated with the target. The parameterized mathematical model 206 may include one or more variables in the variables of the features and layers of the portion of the target under consideration, which may be varied and derived. Fig. 9 As shown, one or more of the variables may include the thickness t of one or more layers, the width w (e.g., CD) of one or more features, the height h of one or more features, the sidewall angle α of one or more features, and / or the relative position between features (referred to herein as overlay). Although not shown, one or more of the variables may also include, but are not limited to, the refractive index of one or more layers in the layers (e.g., real or complex refractive index, refractive index tensor, etc.), the extinction coefficient of one or more layers, the absorption of one or more layers, the resist loss during development, the footing of one or more features, and / or the line edge roughness of one or more features. One or more values ​​of one or more parameters of a 1D periodic structure or a 2D periodic structure (such as values ​​of width, length, shape, or 3D profile features) can be input into the reconstruction process based on knowledge of the patterning process and / or other measurement processes. For example, the initial values ​​of the variables may be those expected values ​​for one or more parameters of the target of measurement, such as values ​​of CD, pitch, etc.

[0098] In some cases, the target may be divided into multiple instances of a unit cell. To facilitate easy calculation of the target's radiation distribution in that case, model 206 may be designed to perform calculations / simulations using a unit cell of the target's structure, with the unit cell repeated as an instance across the entire target. Thus, model 206 may perform calculations using one unit cell, and copy the results to fit the entire target using appropriate boundary conditions in order to determine the target's radiation distribution.

[0099] In addition or as an alternative to calculating the radiation distribution 208 at reconstruction time, multiple radiation distributions 208 may be pre-calculated for multiple variations of the variable of the considered target portion to generate a library of radiation distributions for use at reconstruction time.

[0100] Then, at 212, the measured radiation distribution 108 is compared to the calculated radiation distribution 208 (e.g., a radiation distribution calculated around the time or obtained from a library) to determine a difference between the two. If a difference exists, the value of one or more of the variables of the parameterized mathematical model 206 may be varied to obtain a new calculated radiation distribution 208 (e.g., calculated or obtained from a library), and the new calculated radiation distribution 208 is compared to the measured radiation distribution 108 until a sufficient match is found between the measured radiation distribution 108 and the radiation distribution 208. At this point, the values ​​of the variables of the parameterized mathematical model 206 provide a good match or best match to the geometry of the actual target 30'. In one embodiment, a sufficient match exists when the difference between the measured radiation distribution 108 and the calculated radiation distribution 208 is within a tolerance threshold.

[0101] In these metrology devices, a substrate support may be provided to hold the substrate W during the measurement operation. The substrate support may be in the form of Figure 1 The substrate table WT of the optical system is similar or identical to that of the substrate table WT. In the example where the metrology apparatus is integrated with the lithography apparatus, it can even be the same substrate table. Coarse and fine positioners can be provided to accurately position the substrate relative to the measuring optical system. For example, various sensors and actuators are provided to obtain the position of the target of interest and place it under the objective lens. Typically, multiple measurements are performed on target instances at different positions across the substrate W. The substrate support can be moved in the X-direction and the Y-direction to obtain different target instances, and in the Z-direction to obtain the desired position of the target relative to the focus of the optical system. When, for example in practice, the optical system can remain essentially stationary (typically, in the X-direction and the Y-direction, and perhaps also in the Z-direction) and only the substrate moves, it is convenient to think of and describe the operation as if the objective lens is placed in different positions relative to the substrate. If the relative position of substrate and optical system is correct, in principle it does not matter which of the substrate and the optical system is moving in the real world, or whether both are moving, or a combination of one part of the optical system moving (e.g., in the Z direction and / or in a tilt direction) and the rest of the optical system remaining stationary and the substrate moving (e.g., in the X and Y directions and, optionally, also in the Z direction and / or in a tilt direction).

[0102] In one embodiment, the measurement accuracy and / or sensitivity of a target may vary with respect to one or more properties of a radiation beam provided onto the target, such as the wavelength of the radiation beam, the polarization of the radiation beam, the intensity distribution (i.e., angular or spatial intensity distribution) of the radiation beam, etc. Thus, a particular measurement strategy may be selected that desirably obtains, for example, good measurement accuracy and / or sensitivity of a target.

[0103] In order to monitor a patterning process (e.g., a device manufacturing process) that includes at least one pattern transfer step (e.g., a lithography step), the patterned substrate is inspected and one or more parameters of the patterned substrate are measured / determined. The one or more parameters may include, for example, overlay between consecutive layers formed in or on the patterned substrate, critical dimensions (CD) of features formed in or on the patterned substrate (e.g., critical line width), focus or focus error of the lithography step, dose or dose error of the lithography step, optical aberration of the lithography step, placement error (e.g., edge placement error), etc. The measurement may be performed on a target on the product substrate itself and / or on a dedicated metrology target provided on the substrate. The measurement may be performed after resist development but before etching, or may be performed after etching.

[0104] In one embodiment, the parameter obtained from the measurement process is a parameter derived from a parameter determined directly from the measurement process. As an example, the derived parameter obtained from the measured parameter is an edge placement error of the patterning process. The edge placement error provides a change in the edge position of the structure produced by the patterning process. In one embodiment, the edge placement error is derived from the overlay value. In one embodiment, the edge placement error is derived from a combination of the overlay value and the CD value. In one embodiment, the edge placement is derived from a combination of the overlay value, the CD value, and a value corresponding to a local variation (e.g., edge roughness, shape asymmetry, etc. of each structure). In one embodiment, the edge placement error includes an extreme value (e.g., 3 standard deviations, i.e., 3σ) of the combined overlay error and CD error. In one embodiment, in a multi-patterning process involving the generation of a structure and involving "cutting" a structure by removing a portion of the structure by etching a pattern provided by the patterning process relative to the structure, the edge placement error has the following form (or includes one or more of the following): where σ is the standard deviation, σ overlay Corresponding to the standard deviation of the overlay, σ CDU structures corresponds to the standard deviation of the critical dimension uniformity (CDU) of the structures produced during the patterning process, σ CDU cuts Corresponding to the standard deviation of the critical dimension uniformity (CDU) of the cuts (generated during the patterning process, if any), σ OPE,PBA corresponds to the standard deviation of the optical proximity effect (OPE) and / or the proximity bias average (PBA), which is the difference between the CD at the pitch and the reference CD, and σ LER,LPE corresponds to the standard deviation of the line edge roughness (LER) and / or the local placement error (LPE). Although the above formula is related to the standard deviation, it can be expressed in a different comparable statistical way, such as the variance.

[0105] There are various techniques for measuring the structure formed in the patterning process, including the use of scanning electron microscopes, image-based measurement tools and / or various special tools. As discussed above, a fast and non-invasive form of a special measurement tool is a tool that directs a radiation beam to a target on the surface of a substrate and measures the properties of the scattered (diffraction / reflection) beam. By evaluating one or more properties of the radiation scattered by the substrate, one or more properties of the substrate can be determined. This can be referred to as diffraction-based measurement. One such application of the measurement based on the diffraction is in the measurement of the characteristic asymmetry within the target. For example, this can be used as a measurement of overlay, but other applications are also known. For example, asymmetry can be measured by comparing relative parts of the diffraction spectrum (for example, comparing the -1 order and +1 order in the diffraction spectrum of a periodic grating). This can be performed as described above and, for example, in U.S. Patent Application Publication US2006-066855, the entire content of which is incorporated herein by reference. Another application of diffraction-based measurement is in the measurement of the characteristic width (CD) within the target. Such techniques can be used as described above Figures 6 to 9 The described apparatus and method.

[0106] Now, while these techniques are effective, it is desirable to provide a new type of measurement technique that yields feature asymmetries within a target, such as overlay, CD asymmetry, sidewall angle asymmetry, etc. Such a technique may be effective for specially designed metrology targets, or more significantly, for determining feature asymmetry directly on a device pattern.

[0107] Referring to Figure 10, the principles of this measurement technique are described in the context of an overlay embodiment. Fig. 10A In the figure, the geometric symmetry unit cell of the target T is shown. Fig. 10C As shown, the target T may include only a single physical instance of the unit cell, or may include multiple physical instances of the unit cell.

[0108] The target T may be a specially designed target. In one embodiment, the target is used for scribing. In one embodiment, the target may be an intra-die target, i.e., the target is within the device pattern (and therefore between scribing lanes). In one embodiment, the target may have a feature width or spacing comparable to the device pattern features. For example, the target feature width or spacing may be less than or equal to 300% of the minimum feature size or spacing of the device pattern, less than or equal to 200% of the minimum feature size or spacing of the device pattern, less than or equal to 150% of the minimum feature size or spacing of the device pattern, or less than or equal to 100% of the minimum feature size or spacing of the device pattern.

[0109] The target T may be a device structure. For example, the target T may be a portion of a memory device (which typically has one or more structures that are or may be geometrically symmetrical, as discussed further below).

[0110] In one embodiment, a physical instance of a target T or unit cell may have an area less than or equal to 2400 square microns, an area less than or equal to 2000 square microns, an area less than or equal to 1500 square microns, an area less than or equal to 1000 square microns, an area less than or equal to 400 square microns, an area less than or equal to 200 square microns, an area less than or equal to 100 square microns, an area less than or equal to 50 square microns, an area less than or equal to 25 square microns, an area less than or equal to 10 square microns, an area less than or equal to 5 square microns, an area less than or equal to 1 square micron, an area less than or equal to 0.5 square microns, or an area less than or equal to 0.1 square microns. In one embodiment, the physical instance of the target T or unit cell has a cross-sectional dimension parallel to the plane of the substrate that is less than or equal to 50 microns, less than or equal to 30 microns, less than or equal to 20 microns, less than or equal to 15 microns, less than or equal to 10 microns, less than or equal to 5 microns, less than or equal to 3 microns, less than or equal to 1 micron, less than or equal to 0.5 microns, less than or equal to 0.2 microns, or less than or equal to 0.1 microns. In the case of a unit cell, the area and / or size can be 0.75 to 0.05 times the lower limit of these ranges.

[0111] In one embodiment, the structural pitch of the physical instance of the target T or unit cell is less than or equal to 5 microns, less than or equal to 2 microns, less than or equal to 1 micron, less than or equal to 500nm, less than or equal to 400nm, less than or equal to 300nm, less than or equal to 200nm, less than or equal to 150nm, less than or equal to 100nm, less than or equal to 75nm, less than or equal to 50nm, less than or equal to 32nm, less than or equal to 22nm, less than or equal to 16nm, less than or equal to 10nm, less than or equal to 7nm, or less than or equal to 5nm.

[0112] In one embodiment, the target T has multiple physical instances of the unit cell. Thus, the target T may generally have the higher dimensions listed herein, while the physical instances of the unit cell will have the lower dimensions listed herein. In one embodiment, the target T includes 50,000 or more physical instances of the unit cell, 25,000 or more physical instances of the unit cell, 15,000 or more physical instances of the unit cell, 10,000 or more physical instances of the unit cell, 5,000 or more physical instances of the unit cell, 1000 or more physical instances of the unit cell, 500 or more physical instances of the unit cell, 200 or more physical instances of the unit cell, 100 or more physical instances of the unit cell, 50 or more physical instances of the unit cell, 10 or more physical instances of the unit cell.

[0113] Desirably, the physical instance of the unit cell or multiple physical instances of the unit cell together fill the beam spot of the metrology device. In this case, the measurement result substantially only includes information from the physical instance of the unit cell (or multiple instances thereof). In one embodiment, the cross-sectional width of the beam spot is 50 microns or less, 40 microns or less, 30 microns or less, 20 microns or less, 15 microns or less, 10 microns or less, 5 microns or less, 2 microns or less, 1 micron or less, or 500 nanometers or less.

[0114] Fig. 10A The unit cell in includes at least two structures that are instantiated on a substrate or will be physically instantiated on a substrate. The first structure 1000 includes a line, while the second structure 1005 has an elliptical shape. Of course, the first structure 1000 and the second structure 1005 can be structures different from the depicted structures.

[0115] Further, in this example, since the first structure 1000 and the second structure 1005 are transferred to the substrate separately, there may be a relative offset between the first structure 1000 and the second structure 1005 relative to their intended positions, thereby having an overlay error. In this example, the first structure 1000 is located in a higher layer on the substrate compared to the second structure 1005. Therefore, in one embodiment, the second structure 1005 can be produced in a first lower layer in a first execution of the patterning process, and the first structure 1000 can be produced in a second layer higher than the first lower layer in a second execution of the patterning process. Now, the first structure 1000 and the second structure 1005 do not have to be located in different layers. For example, in a double patterning process (including, for example, an etching process as part thereof), the first structure 1000 and the second structure 1005 can be produced in the same layer to essentially form a single pattern, but there may still be an "overlay" problem with respect to their relative placement within the same layer. In this single-layer example, both the first structure 1000 and the second structure 1005 can have, for example Fig. 10A 1000 , but the lines of the second structure 1005 (which have been provided on the substrate by a first pattern transfer process) may be staggered with the lines of the structure 1000 provided in the second pattern transfer process.

[0116] Notably, the unit cell has or can have geometric symmetry about an axis or point. For example, Fig. 10A The unit cell in has reflection symmetry with respect to, for example, axis 1010, and point / rotation symmetry with respect to, for example, point 1015. Similarly, it can be seen that Fig. 10C The physical instances of the unit cells in (and hence the combinations of physical instances of the unit cells) have geometric symmetries.

[0117] In one embodiment, the unit cell has a geometric symmetry for a certain feature (such as overlay). When it is geometrically symmetric, the embodiments herein focus on unit cells with zero overlay. However, in contrast, for certain geometric asymmetries, the unit cell can have zero overlay. Then, when it has certain geometric asymmetries, appropriate offsets and calculations can be used to illustrate the unit cell with zero overlay. Specifically, the unit cell should be able to change symmetry (e.g., become asymmetric, or further become asymmetric, or change from an asymmetric situation to symmetric) according to a certain feature value.

[0118] exist Fig. 10AIn the example of FIG. 1 , the unit cell has a geometric symmetry of zero overlay (although it need not be zero overlay). This is represented by arrows 1020 and 1025, which show that the lines of the first structure 1000 are uniformly aligned relative to the elliptical shape of the second structure 1005 (and this uniform alignment at least in part enables the unit cell to have a geometric symmetry such as Fig. 10A ). Thus, in this example, when the unit cell has geometric symmetry, there is zero overlay. However, when there is an error in the overlay (e.g., non-zero overlay), the unit cell is no longer geometrically symmetric, and by definition, the target is no longer geometrically symmetric.

[0119] Further, in the case where the target includes multiple physical instances of a unit cell, the instances of the unit cell are arranged periodically. In one embodiment, the instances of the unit cell are arranged in a lattice. In one embodiment, the periodic arrangement has geometric symmetry within the target.

[0120] Thus, as further discussed below, in the present technology, a change in geometric symmetry (e.g., a change to geometric asymmetry, or a change to further geometric asymmetry, or a change from geometric asymmetry to geometric symmetry) related to a feature asymmetry of interest (e.g., non-zero overlay) is utilized to determine the feature asymmetry (e.g., non-zero overlay).

[0121] You can use, for example Figure 7 The measuring device includes Fig. 10A 1030 . The target is a physical instance of a unit cell of a light source. The radiation redirected by the target can be measured, for example, by a detector 190. In one embodiment, the pupil of the redirected radiation is measured, i.e., the Fourier transform plane. An example measurement of such a pupil is depicted as a pupil image 1030. Although the pupil image 1030 has a diamond-like shape, it need not have such a shape. Unless the context requires otherwise (e.g., where the pupil plane of a particular optical system is being identified), the terms pupil and pupil plane herein include any conjugate thereof. Pupil image 1030 is effectively an image of the pupil of the redirected radiation specified in terms of an optical property (in this case, intensity).

[0122] For convenience, the discussion herein will focus on intensity as the optical property of interest. However, the techniques herein may be used with one or more alternative or additional optical properties, such as phase and / or reflectivity.

[0123] Further, for convenience, the discussion herein focuses on detecting and processing images of redirected radiation, and in particular pupil images. However, the optical properties of the redirected radiation may be measured and represented in a manner other than as an image. For example, the redirected radiation may be processed according to one or more spectra (e.g., intensity as a function of wavelength). Thus, detected images of the redirected radiation may be considered as examples of optical representations of the redirected radiation. Thus, in the case of a pupil plane image, the pupil image is an example of a pupil representation.

[0124] Further, the redirected radiation may be polarized or unpolarized. In one embodiment, the measurement beam radiation is polarized radiation. In one embodiment, the measurement beam radiation is linearly polarized.

[0125] In one embodiment, the pupil representation has primarily or substantially redirected radiation of one diffraction order from the target. For example, the radiation may be 50% or more, 70% or more, 80% or more, 85% or more, 90% or more, 95% or more, 98% or more, or 99% or more of a particular order. In one embodiment, the pupil representation has primarily or substantially zero-order redirected radiation. This occurs, for example, when the spacing of the target, the wavelength of the measured radiation, and optionally one or more other conditions, cause the target to redirect primarily zero-order (although one or more higher-order radiation may be present). In one embodiment, the majority of the pupil representation is zero-order redirected radiation. In one embodiment, the pupil representation has zero-order radiation, and separately has first-order radiation, which can then be linearly combined (superimposed). Figure 7 The aperture 186 in may be used to select a particular radiation order, for example, zeroth order radiation.

[0126] Considering the pupil image 1030 corresponding to the geometrically symmetrical unit cells of the first structure 1000 and the second structure 1005, it can be seen that the intensity distribution is substantially symmetrical within the pupil image (e.g., where the type of symmetry is the same as the type of symmetry of the geometric structure). This is further confirmed by removing the symmetrical intensity distribution portion from the pupil image 1030, thereby producing the derived pupil image 1035. To remove the symmetrical intensity distribution portion, the symmetrical intensity distribution portion of a particular pupil image pixel (e.g., a pixel) can be removed by subtracting the intensity at the symmetrically placed pupil image pixel from the intensity at the particular pupil image pixel, or vice versa. In one embodiment, the pixel can correspond to a pixel of a detector (e.g., detector 190), but this is not required; for example, the pupil image pixel can be a plurality of pixels of the detector. In one embodiment, the symmetry point or symmetry axis across which the pixel intensity is subtracted corresponds to the symmetry point or symmetry axis of the unit cell. Thus, for example, considering the pupil image 1030, the symmetrical intensity distribution portion can be removed by, for example, subtracting the intensity at the particular pixel as shown.i Subtract the intensity I from the symmetrically placed (ie, symmetrically positioned with respect to axis 1032) pixel i ' to remove. Therefore, the intensity S at a specific pixel where the symmetric intensity part is removed i Then S i =I i -I i '. This operation may be repeated for a plurality of pixels of the pupil image (e.g., all pixels in the pupil image). As can be seen in the derived pupil image 1035, the intensity distribution corresponding to the symmetric unit cell is substantially completely symmetric. Thus, a symmetric target having a symmetric unit cell geometry (and, if applicable, a certain periodicity of instances of the unit cell) produces a symmetric pupil response as measured by the metrology device.

[0127] Now, refer to Fig. 10B , relative to Fig. 10A An example of overlay error is depicted in the unit cell depicted in FIG. In this case, the first structure 1000 is displaced along the X direction relative to the second structure 1005. Specifically, Fig. 10B , axis 1010 centered on the line of first structure 1000 has been shifted to the right to axis 1045. Thus, there is an error in the overlay 1040 along the X direction; that is, an X-direction overlay error. Of course, the second structure 1005 can be shifted relative to the first structure 1000, or both can be shifted relative to each other. In any case, the result is an X-direction overlay error. However, as will be understood from the unit cell arrangement, a pure relative displacement between the first structure 1000 and the second structure 1005 along the Y direction does not change the geometric symmetry of the unit cell. However, with appropriate geometric arrangements, overlays along two directions or between different combinations of portions of unit cells can change the symmetry and can also be determined, as discussed further below.

[0128] As the physical configuration of the unit cell from Fig. 10A As a result of the change in the nominal physical configuration of the unit cell in the pupil image 1030 and as represented by the error in the overlay 1040, the result is that the unit cell has become geometrically asymmetric. This can be seen by the different length arrows 1050 and 1055, which show that the elliptical shape of the second structure 1005 is unevenly positioned relative to the lines of the first structure 1000. The symmetry is checked about the symmetry point or symmetry axis of the pupil image 1030 (i.e., in that case axis 1032, which is now shown as axis 1034).

[0129] For example, you can use Figure 7 The measuring device is exposed to radiation Fig. 10BA pupil image of the redirected radiation may be recorded, for example, by detector 190. An example of such a pupil image is depicted as pupil image 1060. Pupil image 1060 is effectively an image of intensity. Although pupil image 1060 has a diamond-type shape, it need not have such a shape; pupil image 1060 may be circular in shape or any other shape. Moreover, pupil image 1060 has substantially the same axis or coordinate positions as pupil image 1030. That is, in this embodiment, Fig. 10A The symmetry axes 1010 and Fig. 10B The same axis in the unit cell of is aligned with the symmetry axis 1032 of the pupil images 1030 , 1060 .

[0130] Considering the pupil image 1060 corresponding to the geometrically asymmetric unit cell of the first structure 1000 and the second structure 1005, visually, the intensity distribution is substantially symmetrical within the pupil image. However, there is an asymmetric intensity distribution portion within the pupil image. The asymmetric intensity distribution portion is attributed to the asymmetry in the unit cell. Moreover, the asymmetric intensity distribution is significantly lower in magnitude than the symmetric intensity distribution portion in the pupil image.

[0131] Thus, in one embodiment, to more effectively isolate the asymmetric intensity distribution portion, the symmetric intensity distribution portion may be removed from pupil image 1060, which results in derived pupil image 1065. As with obtaining derived pupil image 1035, the symmetric intensity distribution portion of a particular pupil image pixel (e.g., pixel) may be removed by subtracting the intensity of the symmetrically placed pupil image pixel from the intensity at the particular pupil image pixel, or vice versa, as discussed above. Thus, for example, considering pupil image 1060, the symmetric intensity distribution portion may be removed by, for example, subtracting the intensity I at the particular pixel as shown. i Subtract the intensity I from the symmetrically placed (ie, symmetrically positioned with respect to axis 1032) pixel i 'To get S i This operation can be repeated for multiple pixels of the pupil image (e.g., all pixels in the pupil image). Fig. 10A and Fig. 10B For explanatory purposes, S i The fully derived pupil image of Fig. 10A or Fig. 10B Thus, in one embodiment, only values ​​from half of the pupil image may be used for further processing discussed herein, and thus the derived image pupil used in further processing herein may be only S of the pupil. i half of its value.

[0132] As can be seen in the derived pupil image 1065, the intensity distribution measured using the physical instance of the asymmetric unit cell is asymmetric. Once the symmetric intensity distribution portion is removed, the asymmetric intensity distribution portion is visible, as can be seen in regions 1075 and 1080. As noted above, the fully derived pupil image 1065 is shown, so the asymmetric intensity distribution portion is shown on both halves (even though they are equal to each other in amplitude and distribution on their respective halves).

[0133] Thus, the asymmetry of the geometric domain corresponds to the asymmetry of the pupil. Thus, in one embodiment, a method is provided that uses the optical response of a periodic target that has or can have inherent geometric symmetry in its physical instances of unit cells to determine parameters corresponding to a physical configuration change that causes a change in the geometric symmetry of the physical instances of the unit cells (e.g., causes an asymmetry, or causes other asymmetries, or causes an asymmetric unit cell to become symmetric). Specifically, in one embodiment, the asymmetry (or the lack thereof) caused by overlay in the pupil as measured by a metrology device can be used to determine the overlay. That is, the pupil asymmetry is used to measure the overlay within the physical instances of the unit cells and therefore within the target.

[0134] To consider how to determine a parameter corresponding to a physical configuration change that causes geometric asymmetry in a unit cell, the intensity of a pixel in a pupil image can be considered based on the physical properties of the target that affects the pixel. To this end, an overlay example will be considered, but the techniques and principles can be extended to another parameter corresponding to a physical configuration change that causes geometric asymmetry in a unit cell (e.g., asymmetric side wall angles, asymmetric bottom wall tilts, ellipticity in a contact hole, etc.).

[0135] Return to Fig. 10A and Fig. 10B The unit cell, pixel I in pupil image 1060 i ,I′ i The intensity of can be analytically evaluated as a combination of intensity components attributable to different physical properties of the unit cell. Specifically, a physical configuration change from a symmetric unit cell to an asymmetric unit cell can be evaluated to determine in what way the intensity distribution (and specifically within the pupil image) changes.

[0136] Thus, in a very simple example to illustrate the principle, some changes in the physical configuration of the unit cell profile may be evaluated (although, of course, more or different physical configuration changes may occur). One of the physical configuration changes to be considered is the change in height of the structure 1000 along the Z direction, referred to as Δx h. However, remarkably, this height change is usually uniform across the physical instance of the unit cell. That is, Δx h will result in the same change in the physical configuration of the unit cell on one side of the symmetry point or axis as on the other side of the symmetry axis or point. Similarly, other physical configuration changes (such as changes in CD, sidewall angle, etc.) are generally consistent across the physical instances of the unit cell, so it follows that the change in the physical configuration of the unit cell on one side of the symmetry point or axis is the same as the change in the physical configuration of the unit cell on the other side of the symmetry axis or point. Therefore, for convenience, only Δx is considered h , but it represents many other physical configuration changes that are consistent throughout the unit cell.

[0137] Another physical configuration change in the unit cell of interest is the relative shift between structure 1000 and structure 1005, i.e., the change in overlay 1040. This overlay shift is referred to as Δx ov Of course, overlays in different or additional directions can be considered. Notably, Δx ov The physical configuration of the unit cell on one side of the axis of symmetry or the point of symmetry is different from the physical configuration of the unit cell on the other side of the axis of symmetry or the point of symmetry. Each symmetrical pixel pair has information about overlay. Significantly, although changes in most target profile parameters (CD, height, etc.) can cause symmetrical changes in the pupil (and can therefore be considered as symmetrical parameters), overlay changes can cause asymmetric changes in the measured pupil. Therefore, overlay changes produce asymmetric pupil responses. Further, most (if not all) other unit cell profile parameters do not produce asymmetry in the unit cell or pupil response. However, they may affect the measured overlay value. As discussed below, for the first order, other unit cell profile parameters may have no effect. In one embodiment, for the second order or higher order, other unit cell profile parameters have an influence on the determination of the overlay value. Therefore, as discussed in more detail below, by measuring the pupil asymmetry, overlay can be determined thereby.

[0138] Specifically, to evaluate how to determine overlay based on the measured pupil asymmetry, the intensity I of pixel i in pupil image 1060 is i It can be defined as follows:

[0139]

[0140] where I0 is the fundamental intensity attributable to the illuminating radiation, and a, e, f, and g are coefficients. Thus, similarly, the complementary symmetric pixel I′ in pupil image 1060 i The strength can be defined as follows:

[0141]

[0142] where coefficients a', b', c', d', e' and f' are specific to the complementary symmetric pixel I' i The intensity of the pupil image 1060 is the same as the intensity of the pixel I i The strength coefficients a, b, c, d, e and f are related.

[0143] Then, the difference S in intensity between symmetric pixels in pupil image 1060 is i =I i -I′ i can be evaluated as follows:

[0144]

[0145] It has been found that due to, for example, symmetry, as can be seen in equation (3), it is possible to include only all terms of symmetric parameters (such as eΔx h ) are discarded. Further, due to, for example, symmetry, it has been found that terms with even powers of overlay are equal for symmetrically positioned pixels, so terms such as Term like 1 is also discarded. This leaves only terms that are combinations of overlays with symmetry parameters and terms that are overlays with only odd powers (e.g., powers of 1, 3, 5, 7, etc.).

[0146] In equation (3) above, it has been found that the difference in intensity S i Mainly depends on aΔx ov That is, the difference in intensity S i depends largely linearly on the overlay, or more significantly, depends largely linearly on the intensity (specifically, on the difference in intensity S i ). Therefore, the combination of pixel intensities, when linearly combined with an appropriate conversion factor, can give a good estimate of the overlay.

[0147] Therefore, in one embodiment, it has been discovered that the overlay can be determined from a combination of appropriately weighted pixel intensities (where the weighting itself serves as or can be combined with the intensity to overlay conversion factor). In one embodiment, the overlay signal can be described as follows:

[0148] M=∑ i w i S i (4)

[0149] The overlay signal M is the signal component S in the measurement pupil i The weighted combination of i is the signal component S iThe weights w are the corresponding weights for each signal component in (and the weights act as conversion factors between signal components and overlays; as noted above, conversion factors can be used in combination with weights that do not convert signal components into overlay signals. In one embodiment, the weights w i is a vector whose amplitude is related to the overlay. As indicated above, the signal component S for half of the measured pupil can be determined i In one embodiment, if the signal component S i For all pairs (N / 2) of symmetrical pixels (N) to have substantially the same amplitude, the signal component S i The signal component S is averaged and converted into i With the total signal component S i to the overlay conversion factor C to give the total overlay: Thus, in one embodiment, the weights may have two roles: one as a measure of the trust of each pair of pixels with respect to their overlay, and another role is to convert the value of the optical property of the signal component (e.g., intensity level, e.g., grayscale) to an overlay value (in nanometers). As discussed above, the second role may be delegated to a conversion factor.

[0150] However, for example, for all paired symmetric pixels, the signal component S i In the case where the amplitudes of the pixels in the measurement pupil are all substantially different, weighting all pixels in the measurement pupil may result in a low signal-to-noise ratio (poor accuracy). Therefore, it is desirable to weight those pixels that are sensitive to overlay to make a greater contribution to the calculation of overlay. Therefore, in one embodiment, the weights of pixels that are sensitive to overlay are different (e.g., higher) than the weights of those pixels that are less sensitive to overlay (effectively, inactive pixels). As noted above, the pixels in regions 1075 and 1080 of the derived pupil 1065 are relatively highly sensitive to overlay, while the remaining pixels in the derived pupil 1065 (having low to no intensity relative to the pixels in regions 1075 and 1080) are less sensitive to overlay (and thus should be weighted to contribute less to the overlay determination). Therefore, in one embodiment, a weighting scheme can be created to increase or maximize the signal-to-noise ratio (e.g., to obtain better accuracy). In one embodiment, a weighting scheme can be created to increase or maximize stacking sensitivity (e.g., to provide better tolerance to systematic errors).

[0151] In one embodiment, the aΔx of equation (3) is effectively determined ov In one embodiment, it can be extended to determine aΔx ov term and bΔx ov Δx h(and other comparable terms for other parameters such as CD, sidewall angle, etc.). However, unlike the case where only aΔx of Eq. (3) is effectively determined ov The calculation may be more complex than the weights of the terms. Moreover, there is a trade-off between the robustness of the nonlinear process (for the symmetry parameters) and the accuracy of the determination of the overlay (i.e., how close the determined values ​​are for each determination of the same actual overlay). Thus, using this calculation may sacrifice accuracy to enhance robustness. Thus, optimizations may be performed to enhance accuracy (e.g., maximize the impact of the linear terms and suppress the nonlinear terms), enhance robustness (e.g., maximize the nonlinear terms), or find a balance between the two. In any case, however, the combination of intensities using linear combinations with associated weights can result in a fast determination of the overlay because it requires only pupil acquisition and a simple calculation of equation (4).

[0152] In one embodiment, in the case where high-order terms become significant, nonlinear solution techniques can be used to solve the problem with and / or other higher order terms. It will be appreciated that nonlinear solution techniques may be more efficient than simply measuring each signal component S in the pupil. i With each signal component S i The respective weights w i It is more complex to multiply and then add them all up. Moreover, there is again a trade-off between the robustness of the nonlinear process and the accuracy of the determination of the overlay (i.e., how close the determined values ​​are for each determination of the same actual overlay). Therefore, using this calculation may sacrifice accuracy to enhance robustness. Thus, optimizations may be performed to enhance accuracy and / or enhance robustness.

[0153] Therefore, utilizing the realization of an asymmetric intensity distribution resulting from the geometric asymmetry of the unit cell caused by overlay, the error of overlay can be determined by focusing on the analysis of the asymmetric intensity distribution. Therefore, a technique for determining overlay based on an asymmetric intensity distribution caused by a change in the physical configuration of the target associated with the overlay will now be discussed.

[0154] Reference Fig.11 , schematically depicting a method for determining weights. In order to enable the determination of weights, it is advantageous to use the above Fig. 9 That is, in one embodiment, CD reconstruction is used to isolate the pupil image and overlay signal of a physical instance of an asymmetric unit cell.

[0155] Fig.11The method comprises two processes. The first process 1100 involves: using a reconstruction technique for the CD and / or one or more other profile parameters of the target to derive a nominal profile of the target (and therefore one or more physical instances of a unit cell therein) exposed on the substrate as part of the patterning process. Using the nominal profile of the target, in process 1110, a basic engine of the reconstruction technique is used to obtain weights. The weights can then be used to derive the overlay from the measured pupil, such as in combination with Fig.12 as further described.

[0156] Thus, at process 1100, measurements 1130 are obtained of a substrate on which one or more physical instances of a unit cell of interest are provided as a target. In one embodiment, the measurements are of the target after etching. In one embodiment, the measurements are of the target after development but before etching. In one embodiment, the target is a device structure. In one embodiment, a measurement such as Figure 7 For example, targets may include Fig. 10A or Fig. 10B A physical instance of a unit cell is, for example, Fig. 10C A single instance or multiple adjacent instances as shown. In one embodiment, measurements of multiple instances of the target (and therefore multiple physical instances of the unit cell) are obtained. In one embodiment, the measurements are of target instances distributed across the substrate. In one embodiment, multiple substrates are measured, each substrate having one or more target instances (each having one or more physical instances of the unit cell). Thus, in one embodiment, a radiation distribution 108 is obtained for each measured target.

[0157] Then, using the reconstruction process at 1100 (such as in Fig. 9 In and about Fig. 9 The reconstruction process described in Fig. 9 The nominal profile of the physical instance of the unit cell is compared to the profile 206 of the unit cell. The reconstruction process obtains the expected profile 1120 of the physical instance of the unit cell to initiate and facilitate the reconstruction process. In one embodiment, the derived nominal profile is obtained from the average of the profiles of the target instances across one or more substrates. For example, the radiation distribution 108 of each target can be processed to derive a specific profile of that instance of the target, and then the profiles of multiple instances of the target can be averaged together to derive the nominal profile. In one embodiment, the nominal profile includes at least a geometric profile of the target. In one embodiment, the geometric profile is a 3D profile. In one embodiment, the nominal profile includes information about one or more material properties of one or more layers that constitute the physical target.

[0158] Thus, in one embodiment, for various parameter values ​​of the profile of a target (and, therefore, a unit cell) obtained by measuring multiple instances of the target across a substrate (and optionally more than one substrate), a nominal profile may be considered to be the center of gravity. However, in one embodiment, the nominal profile may have a different form and be more specific. For example, the nominal profile may be defined for one or more specific instances of the target (e.g., by using values ​​from the same target location (or locations) from multiple substrates). As another example, the nominal profile may be defined for a particular substrate (e.g., by using values ​​from only that substrate). In one embodiment, as Fig.12 As part of the process, the nominal profile can be adjusted for a specific target and / or substrate. Fig.12 As part of the process, when a target and / or substrate is measured, a reconstruction technique can be used with the measurement data to fine-tune a nominal profile of the target and / or substrate, which can then be used as the nominal profile in this article to determine weights, and the weights can then be used together with the same measurement data to produce one or more overlay values.

[0159] The reconstructed nominal profile 1140 is then provided to process 1110. Thus, in one embodiment, process 1110 uses a derived nominal profile of a target, e.g., a geometric post-etch profile of a unit cell of a device derived from measurement data. In one embodiment, the nominal profile may be in the form of a parameterized model, such as model 206 parameterized from the measured unit cell. Thus, in one embodiment, process 1110 uses a derived profile model of a unit cell, e.g., a model of a geometric post-etch profile of a physical instance of a unit cell of a device derived from measurement data.

[0160] The basic engine of the reconstruction techniques described herein is used with a derived profile or a derived profile model to derive weights in process 1110. In one embodiment, the derived profile model or a derived profile model derived from the derived profile is used to determine pupil pixels that are sensitive to overprint in a unit cell. In particular, in one embodiment, the sensitivity to overprint of the pupil response is determined by using simulations (e.g., a Maxwell solver) to determine changes in the pupil response due to induced changes in overprint with respect to the nominal profile.

[0161] This can be achieved by causing the derived profile model to change in such a way that an overlay change of a certain amount (e.g., 1 nm) is induced in the model, leaving all other parameters / variables of the derived profile model unchanged. This effectively causes a symmetric unit cell to become asymmetric, or an already asymmetric unit cell to be symmetric to change symmetry (which includes becoming further asymmetric or changing from an asymmetric case to a symmetric case).

[0162] Then, the pupil as expected in the metrology device (e.g., for radiation at a certain measurement beam wavelength, measurement beam polarization, measurement beam intensity, etc.) can be derived based on the derived profile model by inducing overlay changes (e.g., using a Maxwell solver, library search, or other reconstruction technique). In the case where the physical instance of the unit cell is smaller than the beam spot, the reconstruction can treat the beam spot as being filled with the physical instance of the unit cell. In one embodiment, the derived pupil can be a simulated pupil image 1060 and / or a derived pupil image 1065 based on the simulated pupil image.

[0163] The derived pupil can then be used to determine the sensitivity of the intensity in the plurality of pupil pixels to changes in overlay, for example, by comparison with a derived pupil of a unit cell without induced overlay (e.g., the derived pupil of the unit cell without induced overlay can be a simulated pupil image 1030 and / or a derived pupil image 1035 based on the simulated pupil image). In one embodiment, these sensitivities form the basis of the weights.

[0164] In one embodiment, the pixels of the pupil (and therefore the pixel intensity, signal component S i etc.) can be expressed as a vector. In one embodiment, the weights can then be derived from the Jacobian matrix generated in the modeling. In one embodiment, the weights can be derived from the Moore-Penrose pseudo-inverse of the Jacobian matrix generated in the modeling. Thus, aΔx in equation (3) is effectively determined ov The weights derived from the Jacobian matrix or the Moore-Penrose pseudo-inverse of the Jacobian matrix appear to work well for relatively modest overlay variations (e.g., within ±3 nm or within ±4 nm or within ±5 nm).

[0165] In one embodiment, it can be extended to determine aΔx ov term and bΔx ov Δx h(and typically other comparable terms for other parameters such as CD, sidewall angle, etc.). In this case, in addition to the Jacobian, the weights are, or can be derived from, a Hessian matrix generated from the modeling. A Hessian shows how the response to overlay changes due to a change in another (symmetric) parameter such as CD by a certain amount. Thus, for each such parameter, there is a column in the Hersian. In one embodiment, in order to be (more) robust, the weights can be altered so that the weights become more orthogonal to the columns (parameters) to which the unit cell is sensitive. In order to become more orthogonal, one or more sensitive columns can be cascaded to the Jacobian, and then the Moore-Penrose pseudo-inverse can be calculated from the Jacobian using one or more columns from the Hersian cascaded with it. The weights are derived from this calculation. However, this calculation may be more complex, and may therefore be applicable to those cases where the overlay values ​​expected in practice are outside the range of overlay variation, within which the weights derived from the (Moore-Penrose pseudo-inverse) of the Jacobian matrix show good results.

[0166] In one embodiment, it can be extended to determine the weights of other terms in equation (3). In that case, in addition to the Jacobian matrix, the weights are third-order derivatives generated in the modeling or can be derived therefrom. In addition, different types of series expansions can be used to determine overlay sensitivity and nonlinearity (e.g., Taylor series, Fourier series, etc.).

[0167] As noted above, the nominal profile may be a fine-tuned nominal profile for each target or substrate. Fig.12 When measuring a specific target or substrate, the reconstruction technique can be used together with the measurement data to fine-tune the nominal profile of that target or substrate as part of the process. Now, depending on the fine-tuning, the weights can be (re)determined and / or a choice can be made between the types of weights made (e.g., Jacobian or a combination of Jacobian and Herschel). For example, weights based on the un-fine-tuned nominal profile may have been previously selected to suppress Δx h However, if the fine-tuning identifies and updates the Δx of the target and / or substrate h , then there is no need to suppress Δx h Therefore, weights can be chosen that favor accuracy over robustness.

[0168] Therefore, the weight w can be output from process 1110 i A collection of (e.g., a vector). The weight w iThe weights 1065 may be used as intensity-to-overlay conversion factors by themselves, or they may be combined with a conversion factor from intensity to overlay (which may be derived as part of the same modeling). From pupil image 1065, it will be appreciated that pixels in regions 1075 and 1080 have relatively higher overlay sensitivities than pixels outside regions 1075 and 1080, and thus their weights are significantly different (e.g., higher) than the weights of pixels outside regions 1075 and 1080 (which have relatively lower overlay sensitivities). Thus, when the weights are combined with measured intensity values ​​of a target having a physical instance of one or more unit cells (such as according to equation (4)), an overlay signal for a particular target (such as a device pattern having a physical instance of a unit cell) may be obtained.

[0169] Further, one or more measurement parameters may be determined to form a measurement strategy for obtaining a measured intensity value for a target. One or more measurement parameters may affect the overlay sensitivity of a pixel. For example, overlay sensitivity varies across different wavelengths of the measurement beam. Thus, in one embodiment, as part of the modeling process 1110, one or more measurement parameters (such as wavelength, polarization, dose, multiple optical property readings obtained by a detector sensor for a particular irradiation of a target (these readings are typically averaged to provide an average optical property value measured for the target)) may be varied. For example, for a particular induced overlay change, one or more measurement parameters may be examined to determine a value for the one or more measurement parameters that reduces the residual error (e.g., between the overlay obtained when the weight is for one value of the one or more parameters and the overlay obtained when the weight is for another value of the one or more parameters) to a minimum or below a particular threshold. Thus, then, values ​​of one or more measurement parameters that improve accuracy may be obtained.

[0170] Further, the robustness to process variations varies between different values ​​of one or more measurement parameters. For example, in particular, the robustness to process variations varies between different values ​​of measurement beam wavelength and / or measurement polarization. Therefore, in one embodiment, the weighting scheme should at least address the dominant contributors that result in a lack of robustness to process variations. Therefore, as a supplement or alternative to determining the values ​​of one or more measurement parameters to improve accuracy, for different specific induced overlay change values ​​(and / or specific induced changes in one or more other parameters of the derived profile model (such as changes in CD, sidewall angle, etc.)), one or more measurement parameters can be examined to obtain the following values ​​of the one or more measurement parameters that support the results of using the weights to have enhanced robustness to process variations. For example, for different amounts of induced overlay change, various values ​​of one or more measurement parameters can be evaluated to determine the following values ​​of the one or more measurement parameters that result in a minimal (or below a threshold) change in overlay determined using the weights associated with the values ​​of the one or more measurement parameters. Of course, a balance can be used between accuracy and enhanced robustness when selecting the values ​​of one or more measurement parameters. For example, a weighting may be applied between the values ​​of one or more measurement parameters determined for accuracy (e.g., weights of performance indicators applicable to measurement accuracy) and the values ​​of one or more measurement parameters determined for enhanced robustness (e.g., weights of performance indicators applicable to measurement robustness), and then the combination of the largest, highest ranking, etc. may be selected. And of course, multiple values ​​of one or more measurement parameters may be determined so that there are actually multiple different measurement strategies in the overall measurement strategy. Multiple values ​​may be ranked according to one or more performance indicators. Thus, optionally, a measurement strategy may be output from process 1110 for obtaining measurement intensity values ​​for a target having one or more physical instances of a unit cell.

[0171] Further, one or more non-overlay parameters (such as CD, side wall angle, etc.) can affect the weights used to map the intensity signal to the overlay. As noted above, an example way to determine the weights in this context is to use a Herschel matrix and / or a third-order derivative. Therefore, in one embodiment, various possible weighting schemes for taking into account one or more non-overlay parameters are possible so as to still maintain a good overlay value. In one embodiment, overlay-informative pixels and their weights can be optimized for overlay determination accuracy. This may require good model quality, i.e., a good estimate of the non-overlay parameters. In one embodiment, overlay-informative pixels and their weights can be optimized to improve robustness to process variations such as non-overlay parameters. This may be at the expense of accuracy.

[0172] In one embodiment, the estimation of one or more non-overlay parameters can be performed using, for example, Fig. 9The CD reconstruction may be performed using the reconstruction techniques described herein and feed forward to adjust the derived profile or derived profile model. For example, the CD reconstruction may estimate the CD of the target at a specific combination of patterning process settings (e.g., exposure dose, exposure focus, etc.) and / or a specific location of the substrate and use the CD estimate to adjust the CD parameters of the derived profile or derived profile model. In one embodiment, an iterative reconstruction of the precise derived profile or derived profile model parameters may be performed.

[0173] Reference Fig.12 , a method for determining an overlay value of a target having one or more physical instances of a unit cell capable of geometric symmetry. The method includes two processes 1200 and 1210. Process 1200 includes: obtaining a measurement of a target having one or more physical instances of a unit cell. Process 1210 includes: determining an overlay value of the measured target based on the measurement of the target from process 1200.

[0174] Process 1200 takes as input a target 1220 to be measured, which includes one or more physical instances of a unit cell capable of geometric symmetry as described herein. In one embodiment, a target 1220 is provided to a user, such as a user. Figure 7 A metrology device such as a metrology device provides a substrate having one or more instances of a target.

[0175] Optionally, the process 1200 takes as input a particular measurement strategy 1230 specified for a target. In one embodiment, the measurement strategy may specify values ​​for one or more measurement parameters, such as one or more selected from the group consisting of: measurement beam wavelength, measurement beam polarization, measurement beam dose, and / or a plurality of optical property readings obtained by a detector sensor of a metrology device for a particular exposure of the target. In one embodiment, the measurement strategy may include a plurality of measurement strategies, each specifying values ​​for one or more measurement parameters. The measurement strategy may be used to measure a target.

[0176] Then, process 1200 uses the metrology device to measure the target according to the optional measurement strategy. In one embodiment, the metrology device obtains a pupil representation of the redirected radiation. In one embodiment, the metrology device can generate a pupil representation, such as pupil image 1030 (e.g., in the case where the target has no overlay error) or pupil image 1060 (e.g., in the case where the target has overlay error). Thus, in one embodiment, process 1200 outputs optical information 1240 about the redirected radiation from the target, such as a pupil representation of the radiation.

[0177] Then, process 1210 receives optical information 1240 and processes the optical information to determine an overlay value of the target 1260. In one embodiment, process 1210 receives optical information from Fig.11The weight 1250 determined by the method of is taken as input, which is then combined with one or more optical characteristic values ​​(e.g., intensity) obtained or derived from the optical information 1240.

[0178] In one embodiment, process 1210 (or process 1200) can process the optical information to derive a raw overlay signal from the optical information. In one embodiment, the raw overlay signal includes a difference in the optical information, i.e., a difference in the value of an optical property between symmetric pixels across a symmetry axis or symmetry point. In one embodiment, a derived pupil image 1035 (e.g., in the case where the target has no overlay error) or a derived pupil image 1065 (e.g., in the case where the target has an overlay error) can be obtained.

[0179] In one embodiment, weighted and optical information about the radiation redirected by the target (e.g., optical information from process 1200 or a processed version of optical information from process 1200 (such as an original overlay signal)) is combined to determine an overlay value. In one embodiment, overlay can be quickly determined using a combination of redirected measurement beam intensities in a linear combination with associated weights. For example, in one embodiment, an overlay value can be derived using equation (4), where the signal component S of the original overlay signal is used. i The corresponding weight w of each signal component in i The overlay value M is used as the signal component S i A weighted combination of .

[0180] In one embodiment, additionally, the optical information collected from process 1200 can be used to derive one or more parameters related to the target in addition to overlay. For example, the optical information collected from process 1200 can be used in a reconstruction process to derive any one or more geometric profile parameters of the target, such as CD, sidewall angle, floor tilt, etc. Thus, in one embodiment, the same set of optical information collected from a target, such as a post-etch target within a die, can be used to determine overlay, CD, and / or one or more other geometric profile parameters of a target, such as a device structure.

[0181] Although as noted above, the focus is on intensity, in one embodiment the optical property may be reflectivity, the radiation may be polarized and the measurement may be a cross-polarization measurement. For example, a target exposed to a particular linear polarization may be measured across that polarization or at a different polarization. Thus, for a symmetric pixel p i and p i ′ (where the prime sign indicates a symmetrical position), the reflectance R of those pixels can be measured as follows:

[0182]

[0183] Where s represents s polarization and p represents p polarization. Therefore, the reflectivity R ss corresponds to the reflectivity R of s-polarized radiation measured when the target is illuminated with s-polarization, the reflectivity R sp corresponds to the reflectivity R of s-polarized radiation measured when the target is illuminated with p-polarization, and so on. Moreover, these measurements can be made at different wavelengths. Furthermore, it has been found that in some embodiments, the congruence R ps and R sp An overlay of a symmetric unit cell that changes its symmetry in response to the overlay change is found and determined.

[0184] Further, nonlinearities may be caused by overlay and / or other parameters. As discussed above, specific nonlinearities may be addressed by appropriately selecting weights (e.g., by deriving weights using a Herschel matrix and / or third-order derivatives). In one embodiment, nonlinearities may be addressed by deriving overlay using a nonlinear solution based on measured optical information of the redirected radiation from the target.

[0185] In one embodiment, overlay can be determined by using a reconstruction engine for deriving a nominal profile as described above. For example, a nonlinear solver operating from a model based on a derived nominal profile and / or a derived nominal profile model can be used to derive a simulated version of optical information of redirected radiation expected from a target of interest, which can be compared with measured optical information of the target of interest. As noted above, the target of interest includes one or more physical instances of a unit cell, which can be symmetrical and change its symmetry when subjected to overlay. Then, if agreement is not reached within a particular threshold, the geometric profile parameters (e.g., overlay) can be varied, and the simulated version of the optical information recalculated and compared with the measured optical information until agreement is reached within the threshold. Similarly, the measured optical information of the target of interest can be compared with a library of optical information of redirected radiation expected from the target of interest (which is typically derived using a nonlinear solver). Then, if agreement is not reached within a certain threshold, the geometric profile parameters (e.g., overlay) can be varied and the library can be queried again to obtain a simulated version of the optical information, which is compared to the measured optical information until agreement is reached within the threshold.

[0186] In one embodiment, the use of a reconstruction engine together with measured optical information from an object of interest uses measured optical information from which a symmetric radiation distribution has been removed as described above, for example by subtracting the optical property value at each pixel symmetrically located across a point or axis of symmetry from the optical property value at that pixel. Thus, the optical information is substantially related only to the asymmetric radiation distribution. Likewise, the simulated or library version of the optical information is substantially related only to the asymmetric radiation distribution. This will aid in the speed of calculations and / or comparisons, since a large portion of the optical information need not be calculated or evaluated, since this portion of the optical information will be eliminated by the differencing.

[0187] In another embodiment of a nonlinear solution, the expansion of equation (3) can be solved using a nonlinear solver to derive Δx ov Specifically, the values ​​of (aa′), (bb′), (cc′), etc. (as appropriate) in equation (3) may be determined as part of determining a derived nominal profile and / or a derived nominal profile model for the unit cell of interest. For example, once the derived nominal profile is determined as part of the nonlinear reconstruction, a model corresponding to the derived nominal profile (e.g., with a particular overlay change (e.g., Δx)) may be obtained. ov ), then the values ​​of a, b, c, etc. (as appropriate) for each pixel in the pupil can be determined by a nonlinear solver that iterates, for example, by solving (e.g., in response to an overlay (e.g., Δx ov ) to minimize the residual. The result is a vector of a values ​​for the pupil (each a value corresponds to a pixel in the pupil), a vector of b values ​​for the pupil (each b value corresponds to a pixel in the pupil), a vector of c values ​​for the pupil (each c value corresponds to a pixel in the pupil), and so on, as appropriate. These vectors can then be compared with S i A vector combination of values, the S i The vector of values ​​is determined from the measured pupil of the target with the unit cell of interest. For example, a nonlinear solver that iterates through the solution to minimize the residual can take these input vectors and then calculate the overlay Δx ov To solve.

[0188] While the discussion above focuses on using a model that models the physical profile of a unit cell, in one embodiment, the weights may be derived using a data-driven technique without physical profile modeling, or may be derived using a data-driven technique supplemented with physical profile modeling. Thus, in one embodiment, advantageously, the data-driven technique may not require a physical profile model; this may be useful, for example, for limiting the sharing of confidential information, because physical profile modeling starts with and determines details about the unit cell (and therefore the target), which may be sensitive information if the unit cell is a device pattern structure. In one embodiment, the data-driven technique may enable relatively rapid determination of weights, such as those discussed above, to transform measured optical information (e.g., pupil intensity) into patterning process parameters (e.g., overlay). In one embodiment, the data-driven technique enables determination of patterning process parameters at an early stage, because, as discussed below, the data-driven technique may require only measurement data and associated references.

[0189] Thus, in one embodiment, a data-driven technique includes processing data measured from one or more substrates ("get" data) having physical instances of a unit cell of interest patterned thereon as one or more targets, using one or more specific set values ​​of a patterning process parameter (e.g., overlay) of interest. The combination of "set" intentional values ​​of specific patterning process parameters (e.g., overlay) used to produce a pattern together with data measured from those patterns ("get" data) is referred to as a "set-get" process. For example, a specific amount of overlay of a physical instance of a unit cell is created as part of a patterning process, and then a target having the physical instance of the unit cell is measured to obtain, for example, its pupil image (i.e., "get" data). In one embodiment, multiple substrates may be patterned and measured in this manner. In one embodiment, multiple different set values ​​of overlay are created, and these different values ​​of overlay may be on one substrate, may be across different substrates, etc. In one embodiment, each substrate will have multiple measured target instances, resulting in, for example, multiple pupil images. In one embodiment, overlay may be produced by inducing a magnification change from a design magnification between patterning different portions of a physical instance of a unit cell. In one embodiment, overlay can be produced by providing an intentional shift from the designed positioning between patterning different portions of the physical instance of the unit cell. Thus, the result is an intentionally imposed overlay in a target induced, for example, by a lithographic apparatus.

[0190] Typically, in one embodiment, measured data and associated reference values ​​are obtained. Thus, in one embodiment, intentional overlay need not be provided in the case of different overlays, but those overlays are determined by another means (e.g., from a scanning electron microscope). In one embodiment, a critical dimension uniformity substrate with corresponding reference data (e.g., from a CD-SEM) can be used as input data. With the measured data and reference values, the data-driven approach can find weights as discussed above so that the inferred overlay values ​​are similar to the reference values. Thus, while the discussion of data-driven techniques will focus on measured optical information and pupil representations obtained at intentionally set overlay values, they can generally be applied to more general measurement data and associated reference values ​​(whether measured or intentionally set).

[0191] Further, although the techniques herein relate to a particular overlay (e.g., overlay in the X direction), it will be appreciated that the techniques herein may be repeated using corresponding measurement data and reference values ​​for different overlays (e.g., overlay in the Y direction, overlay between structures in different layers, etc.). Therefore, different sets of weights may be determined for different overlays.

[0192] Therefore, reference Fig.13 , depicting a high-level flow of one embodiment of a data-driven technique. At 1300, calculations are performed to derive weights as discussed above to transform measured optical information (e.g., pupil intensity) into patterning process parameters (e.g., overlay). In particular, the calculation uses several inputs. One of the inputs is a setting value 1320 for a setup-acquisition process of a target having a physical instance of a unit cell of interest. As noted above, multiple instances of a target may be measured across one or more substrates, wherein the values ​​of the intentional setting values ​​of the patterning process parameters of one or more instances of the target are different from the values ​​of the intentional setting values ​​of the patterning process parameters of one or more other instances of the target. Another input is the measured optical information 1310 of those instances of the target at different setting values. In one embodiment, the optical information 1310 is a plurality of pupil representations, each pupil representation corresponding to an instance of the target. Then, the inputs 1310 and 1320 are processed in the data-driven technique to obtain weights 1330. An example of such a data-driven technique is described below.

[0193] In one embodiment, the weights are found w An example of a data-driven technique for a vector of is to minimize the following objective function or merit function to obtain the weights w:

[0194]

[0195] in wis a vector of weights that are used to combine with the measured value of the optical property (e.g., intensity) to determine the patterning process parameters (e.g., overlay), each weight corresponding to a pixel value of the pupil, P i is a matrix with the following columns, where each column contains pixel values ​​of a measured optical property of a measured pupil of an instance of a target obtained from a patterned substrate i in order to obtain a particular setting value of a patterning process parameter (the matrix is ​​then transposed so that the columns become pixels of the pupil, the rows become one or more instances of the target on the substrate, and the values ​​in the matrix are the values ​​of the measured optical property at each pixel), s i is a vector containing corresponding settings of a patterning process parameter for one or more instances of a target on one or more substrates i, each setting corresponding to a patterning process parameter value, 1 is a unit vector with a magnitude of set number of values, and c i is the setting value of the patterning process parameter of each substrate and the patterning process parameter The matrix P is the offset difference between the inferred values ​​and D is the number of substrates measured. i There may be a combination of different results for each instance of the target. For example, the target may be measured at different wavelengths, different polarizations, etc. Thus, these results may be cascaded to each column, so that, for example, a single column may have pixel values ​​for the pupil of a target measured at a first wavelength and a first polarization, followed by values ​​in a column for pixels of the pupil of a target measured at a second, different wavelength, or followed by values ​​in a column for pixels of the pupil of a target measured at a second, different polarization (which may then be followed by other values ​​at one or more different polarizations and / or wavelengths).

[0196] So, in fact, the function finds the weight vector w , so that the inferred value for each substrate i Looks the same except for the offset c i Setting values ​​other than s i As similar as possible (in the sense of L2 regularization norm). In principle, the optimal weights and offsets can be calculated by matrix inversion. Since the pixel values ​​of the measured optical properties are obtained by one or more specific measurement devices, the obtained weights can be normalized by calibration data to reduce the influence of the specific measurement device itself on the result.

[0197] Instead of or in addition to using an objective function or merit function as a data-driven technique to find weights as described above, the data-driven technique may use a machine learning algorithm (such as a neural network) or a nonlinear approach to determine the weights based on a measured pupil of a target in which differences in patterning process parameters of interest (e.g., overlay) are intentionally provided.

[0198] In one embodiment, after training (i.e., using an objective function or merit function or machine learning algorithm), other data may be used to check the weights. Training may result in overfitting; the data-driven approach "just" fits the data to the set values. Thus, cross-validation is done. New data with known set values ​​is used to check the weights. This new data may also be a subset of the substrates at hand. Thus, in one embodiment, training is performed on a subset of substrates, and validation is performed on another (extracted) subset of substrates.

[0199] Fig.14 A high-level flow diagram of an embodiment of a data-driven technique incorporating a physical geometry model is described. Fig.13 The described data-driven techniques may be used to derive weights that are used to adjust the physical-geometry model (e.g., by using Herschel to obtain better model nominal values, by changing model nominal values, etc.) such that the weights from the physical-geometry model (e.g., the Jacobian (Moore-Penrose pseudo-inverse) of the physical-geometry model) are the same or similar (e.g., in value, statistics, etc.) to the weights determined by the data-driven techniques. Thus, in one embodiment, the (scaled) weight vector w It can be used to fine-tune the physical geometry model so that the physical geometry model is adjusted so that the Jacobian (the Moore-Penrose pseudo-inverse) is consistent with the (scaled) weight vector w similar.

[0200] Thus, in one embodiment, at 1400, a data-driven technique (an example of which is described above) is performed to derive weights as discussed above. The calculation uses several inputs. One of the inputs is a setting value 1420 of a setup-acquisition process of a target having a physical instance of a unit cell of interest. As noted above, multiple instances of a target may be measured across one or more substrates, wherein the values ​​of the intentional settings of the patterning process parameters of one or more instances of the target are different from the values ​​of the intentional settings of the patterning process parameters of one or more other instances of the target. Another input is measured optical information 1310 of those instances of the target at different settings. In one embodiment, the optical information 1410 is a plurality of pupil representations, each pupil representation corresponding to an instance of the target. Inputs 1410 and 1420 are then processed in a data-driven technique to obtain weights 1430.

[0201] The weights 1430 are input to a process 1440 to fine-tune the physical geometry model using the weights 1430. The process 1440 obtains a physical contour 1450 of the unit cell (which the process 1440 uses to derive a physical contour model) or obtains a physical contour model 1450 of the unit cell (which the process 1440 uses to derive a physical contour model). In one embodiment, the physical contour is a derived nominal contour and / or a derived nominal contour model of the unit cell as discussed above.

[0202] Process 1440 uses the physical geometry model to derive weights corresponding to weights 1430. Those weights are then compared to weights 1430. The comparison may include: amplitude matching, statistical analysis, fit assessment, etc. If there is a significant difference (e.g., by evaluating the comparison relative to a threshold), one or more parameters of the physical profile may be adjusted. For example, one or more physical profile parameters (e.g., CD, sidewall angle, material height, etc.) may be adjusted so that the result of the comparison is closer to or equal to, for example, a specific threshold. In one embodiment, a Herschel may be used to make such fine-tuning, or a nonlinear solver including one or more forward calls (e.g., a Maxwell solver) may be used to accomplish this. Adjustments and comparisons may be iterated until a threshold is met or exceeded. The adjusted physical geometry model may then output updated weights 1460 for use in combination with measured optical information of the target of interest to derive patterning process parameter values.

[0203] Fig.15 A high-level flow diagram of another embodiment of a data-driven technique incorporating a physical geometry model is depicted. When the behavior of the physical geometry model is similar to the measured data, the physical geometry model can be used to predict the impact of process variations. Thus, in one embodiment, the Herschel of the physical geometry model can be used to adjust the weights so that the weights become (more) orthogonal to process variations in the data that are not used in the data-driven technique to obtain weights for adjusting the physical geometry model.

[0204] This approach of using Herschel to adjust weights can also be done without data-driven techniques. That is, this technique of using Herschel to update weights can be used with Fig.11 The physical geometry model approach described in connection with the physical geometry model may be performed. In this case, for example, the weights may be adjusted so that the weights become (more) orthogonal to process variations that are not in the following data used to obtain the derived nominal profile of the unit cell and / or the derived nominal profile model as discussed above. By such adjustment, the weights become more robust to process variations that are not observed in the measurement data used to generate the physical geometry model.

[0205] Thus, in one embodiment, at 1500, a data-driven technique (an example of which is described above) is performed to derive weights as discussed above. The calculation uses several inputs. One of the inputs is a setting value 1510 of a setup-acquisition process of a target having a physical instance of a unit cell of interest. As noted above, multiple instances of a target may be measured across one or more substrates, wherein the values ​​of the intentional settings of the patterning process parameters of one or more instances of the target are different from the values ​​of the intentional settings of the patterning process parameters of one or more other instances of the target. Another input is optical information 1505 of the measurements of those instances of the target at different settings. In one embodiment, the optical information 1505 is a plurality of pupil representations, each pupil representation corresponding to an instance of the target. Then, the inputs 1505 and 1510 are processed in a data-driven technique to obtain weights 1515.

[0206] The weights 1515 are input to a process 1520 to fine-tune the physical geometry model using the weights 1515. The process 1520 obtains a physical contour 1525 of the unit cell (which the process 1520 uses to derive a physical contour model) or obtains a physical contour model 1525 of the unit cell (which the process 1520 uses to derive a physical contour model). In one embodiment, the physical contour is a derived nominal contour and / or a derived nominal contour model of the unit cell as discussed above.

[0207] Process 1520 uses the physical geometry model to derive weights (corresponding to weights 1515) and then compares to weights 1515. The comparison may include: amplitude matching, statistical analysis, fit assessment, etc. If there is a significant difference (e.g., by evaluating the comparison relative to a threshold), one or more parameters of the physical profile may be adjusted. For example, one or more physical profile parameters (e.g., CD, sidewall angle, material height, etc.) may be adjusted so that the result of the comparison is closer to or equal to, for example, a particular threshold. In one embodiment, a Herschel may be used to perform such fine-tuning, or it may be accomplished using a nonlinear solver including one or more forward calls (e.g., a Maxwell solver). Adjustments and comparisons may be iterated until a threshold is met or exceeded.

[0208] However, it will be appreciated that the patterning process may vary during execution and may vary differently for different executions of the patterning process. Therefore, the data obtained for the data-driven technique does not account for all possible patterning process variations. However, when the physical geometry model is adjusted so that its behavior is similar to the measured data, the physical geometry model can be used to predict the impact of process variations and the weights adjusted accordingly.

[0209] Thus, in one embodiment, at 1535, the adjusted physical geometry model 1530 is used to calculate a Herschel of the adjusted physical geometry model. Then, at 1545, the Herschel 1540 is used to adjust the weights so that the weights become (more) orthogonal to (i.e., robust to) process variations in the data that are not used in the data-driven technique to obtain weights for adjusting the physical geometry model. In other words, the weights are adjusted to be more likely to yield accurate results when combined with measured data from the substrate (even when the substrate is subject to process variations).

[0210] A non-limiting example of how to use Herschel's fine-tuning weights is described herein in the context of overlay; different patterning process parameters may be used as appropriate. In this example, it is assumed that only one overlay type (e.g., overlay in the X direction) is evaluated. It is also possible to use multiple overlay types for fine-tuning.

[0211] In this embodiment using Herschel-tuned weights, the overlay response is estimated from data measured from one or more setup-acquire substrates by applying single value decomposition to the data. Assume that the eigenvector d (whose length is 1) corresponds to the overlay response. Then, the following equation is solved to find the vector Δp :

[0212]

[0213] in J is the Jacobian with respect to the overlay parameters, and Herssian H is a matrix where the columns contain the partial derivatives with respect to process variations (e.g., variations in CD, material height, etc.) and the overlay parameters (both the Jacobian and Herssian are obtained from the model as described above). Then, the vector determined Δp Corresponds to the delta parameters that are to be applied to the non-overlay parameters in the model to obtain an updated (eg, better) model.

[0214] To make the weights robust to process variations (i.e., orthogonal to process variations), the following technique can be used. I It can be defined by the following second-order Taylor expansion:

[0215] I ＝ J o+H Δp o (8)

[0216] in J is the Jacobian with respect to the overlay parameters, and H is a matrix where the columns contain the partial derivatives and overlay parameters with respect to process variations (e.g., variations in CD, material height, etc.). The vector Δpcontains the corresponding process variation. Therefore, for a given structure and a given process variation instance Δp (which has an overlay value of ο), the pupil is equal to (approximately) I It will be appreciated that by adding these contributions the above formula can also be extended to more overlay parameters. Moreover, the formula is an approximation since higher orders in the Taylor expansion are neglected.

[0217] Now, if the effect of the process change is small, the Jacobian J The weight is calculated by the Moore-Penrose inverse of . In the case of only one overlay parameter, the weight is equal to And in fact, the weighted average (inner product) with the pupil produces the overlay value ο( Δp = 0 ),Right now,

[0218]

[0219] However, when process variations have a large effect, the overlay response changes:

[0220]

[0221] To make the weights robust to these changes,

[0222] H w = 0 (11)

[0223] This can be done by making the weight w Equal to the matrix [ J H]. Or in other words, the Herssey matrix H is concatenated to the Jacobian before inversion. In this way, the weights become orthogonal to the process variation (but at the expense of some accuracy).

[0224] Thus, from the adjustment 1545, the adjusted weights 1550 are output for use in combination with the measured optical information of the target of interest to derive the patterning process parameter values.

[0225] Fig.16 A high-level flow diagram of another embodiment of a data-driven technique incorporating a physical geometry model is depicted. In this embodiment, the data input to the data-driven technique is expanded by including synthetic optical information (e.g., pupil representation) that contains process variations for the patterning process (e.g., patterning process variations can be obtained from CD measurements). The synthetic optical information alone or in combination with measured optical information can be used to find new weights using the data-driven technique.

[0226] Thus, in one embodiment, at 1500, a data-driven technique (an example of which is described above) is performed to derive weights as discussed above. The calculation uses several inputs. One of the inputs is a setting value 1510 of a setup-acquisition process of a target having a physical instance of a unit cell of interest. As noted above, multiple instances of a target may be measured across one or more substrates, wherein the values ​​of the intentional settings of the patterning process parameters of one or more instances of the target are different from the values ​​of the intentional settings of the patterning process parameters of one or more other instances of the target. Another input is measured optical information 1505 of those instances of the target at different settings. In one embodiment, the optical information 1505 is a plurality of pupil representations, each pupil representation corresponding to an instance of the target. Then, the inputs 1505 and 1510 are processed in a data-driven technique to obtain weights 1515.

[0227] The weights 1515 are input to a process 1520 to fine-tune the physical geometry model using the weights 1515. The process 1520 obtains a physical contour 1525 of the unit cell (which the process 1520 uses to derive a physical contour model) or obtains a physical contour model 1525 of the unit cell (which the process 1520 uses to derive a physical contour model). In one embodiment, the physical contour is a derived nominal contour and / or a derived nominal contour model of the unit cell as discussed above.

[0228] Process 1520 uses the physical geometry model to derive weights corresponding to weights 1515. Those weights are then compared to weights 1515. The comparison may include: amplitude matching, statistical analysis, fit assessment, etc. If there is a significant difference (e.g., by evaluating the comparison relative to a threshold), one or more parameters of the physical profile may be adjusted. For example, one or more physical profile parameters (e.g., CD, sidewall angle, material height, etc.) may be adjusted so that the result of the comparison is closer to or equal to, for example, a particular threshold. The adjustments and comparisons may be iterated until the threshold is met or exceeded.

[0229] Thus, in one embodiment, at 1535, the adjusted physical geometry model 1530 is used to calculate a Herschel of the adjusted physical geometry model. Then, at 1610, the Herschel 1600 is used to generate synthetic optical information (e.g., one or more pupil representations). The synthetic optical information is simulated optical information. The synthetic optical information is intended to mimic one or more expected process variations in the patterning process. In one embodiment, data 1620 about one or more process variations in the patterning process can be combined with the Herschel 1600 to derive the synthetic optical information. In one embodiment, the synthetic pupil can be generated by substituting different overlay values ​​o and different parameter changes Δp into equation (8) above I, where the weights are Corresponding. Although equation (8) described above relates to a single overlay parameter, the technique can also be extended to more overlay parameters by adding these contributions. Further, because the higher order in the Taylor expansion is ignored, the technique using equation (8) is approximate. Data 1620 may include information such as the type and range of process variation (e.g., an indication that overlay, CD, etc. can vary by a certain percentage). Data 1620 may be obtained by measurements in the patterning process (e.g., measurements such as overlay, CD, etc.). Therefore, data 1620 is used together with Herschel 1600 to generate simulated optical information 1630 including expected process variation. Synthetic optical information 1630 may also include one or more associated estimated settings associated with the synthetic optical information 1630. Then, the synthetic optical information 1630 (and any associated settings) is input to the data-driven technique 1500 to be analyzed individually or in combination with the measured optical information to find new weights using the data-driven technique.

[0230] Fig.17 Describes the high-level flow of another embodiment of the data-driven technology combined with physical geometric models. Fig.16 The embodiment of is similar to that of , except that instead of computing the Herschel, for each process variation, a forward call is made to a nonlinear solver (eg, a Maxwell solver) to obtain the synthetic optical information.

[0231] Thus, in one embodiment, at 1500, a data-driven technique (an example of which is described above) is performed to derive weights as discussed above. The calculation uses several inputs. One of the inputs is a setting value 1510 of a setup-acquisition process of a target having a physical instance of a unit cell of interest. As noted above, multiple instances of a target may be measured across one or more substrates, wherein the values ​​of the intentional settings of the patterning process parameters of one or more instances of the target are different from the values ​​of the intentional settings of the patterning process parameters of one or more other instances of the target. Another input is measured optical information 1505 of those instances of the target at different settings. In one embodiment, the optical information 1505 is a plurality of pupil representations, each pupil representation corresponding to an instance of the target. Then, the inputs 1505 and 1510 are processed in a data-driven technique to obtain weights 1515.

[0232] The weights 1515 are input to a process 1520 to fine-tune the physical geometry model using the weights 1515. The process 1520 obtains a physical contour 1525 of the unit cell (which the process 1520 uses to derive a physical contour model) or obtains a physical contour model 1525 of the unit cell (which the process 1520 uses to derive a physical contour model). In one embodiment, the physical contour is a derived nominal contour and / or a derived nominal contour model of the unit cell as discussed above.

[0233] Process 1520 uses the physical geometry model to derive weights corresponding to weights 1515. Those weights are then compared to weights 1515. The comparison may include: amplitude matching, statistical analysis, fit assessment, etc. If there is a significant difference (e.g., by evaluating the comparison relative to a threshold), one or more parameters of the physical profile may be adjusted. For example, one or more physical profile parameters (e.g., overprint, CD, sidewall angle, etc.) may be adjusted so that the result of the comparison is closer to or equal to, for example, a particular threshold. The adjustments and comparisons may be iterated until the threshold is met or exceeded.

[0234] Thus, in one embodiment, at 1720, the adjusted physical geometry model 1700 is used to calculate synthetic optical information as discussed above. As discussed above, data 1710 about one or more process variations in the patterning process can be combined with the adjusted physical geometry model 1700 to derive synthetic optical information. For example, the data 1710 may include information describing the type and range of process variations (e.g., an indication that overlay, CD, etc. can vary by a certain percentage). The data 1710 may be obtained by measurements in the patterning process (e.g., overlay, CD, etc. measurements). As noted above, the process at 1720 may use a forward call for a nonlinear solver (e.g., a Maxwell solver) for process variations to obtain synthetic optical information. Thus, the data 1710 is used together with the adjusted physical geometry model 1700 to generate simulated optical information 1730 including expected process variations. The synthetic optical information 1730 may also include one or more associated estimated setting values ​​associated with the synthetic optical information 1730. The synthesized optical information 1730 (and any associated settings) is then input to the data-driven technique 1500 for analysis, either alone or in combination with the measured optical information, to find new weights using the data-driven technique.

[0235] exist FIG. 10A to FIG. 10C In , a relatively simple example of a unit cell is given, where essentially only the overlay in one direction causes a change in the symmetry of the unit cell. In particular, FIG. 10A to FIG. 10CIn the unit cell, the change of overlay in the X direction leads to the change of symmetry / asymmetry of the unit cell, while the change of overlay in the Y direction does not lead to the change of symmetry in the unit direction. FIG. 10A to FIG. 10C The unit cell has the result of the following two structures 1000, 1005, which are configured in a specific geometrical manner so that the overlay in substantially only one direction causes a change in the symmetry of the unit cell. Of course, this can be designed in this way by appropriate selection of structures. However, it is possible that existing structures (such as device structures) with specific geometric shapes can be identified, so that the overlay in substantially only one direction causes a change in the symmetry of the unit cell. Therefore, various unit cells can be selected or designed so that the overlay in substantially only one direction (which does not have to be along the X direction) can be determined.

[0236] However, advantageously, the unit cell may be identified or designed, the unit cell being configured so that a change in the symmetry of the unit cell is produced for two or more different overlays. In one embodiment, the different overlays may be in different directions. Specifically, in one embodiment, the first overlay may be in the X direction, while the second overlay may be in the Y direction. In one embodiment, the different overlays may be between different combinations of structures or parts of the unit cell, respectively. In one embodiment, those structures may be in the same layer and / or different layers of the target. Specifically, in one embodiment, the first overlay may be between the first structure and the second structure of the unit cell, while the second overlay may be between the first structure (or the second structure) and the third structure of the unit cell or between the third structure and the fourth structure of the unit cell. In this case, the first overlay and the second overlay may be in the same direction. Naturally, there may be a combination of different overlays in different directions and different overlays from a combination of structures of the unit cell. For example, for a first structure in the first layer and a second structure in the second lower layer, the first overlay may be in the X direction; and for a first structure in the first layer and a third structure in the third layer lower than the second layer, the second overlay may be in the Y direction. Therefore, various combinations of overlays can be determined by appropriate identification or design of the unit cells (and thus the targets).

[0237] Furthermore, it will be appreciated that determining the overlay in the X and Y directions may enable the determination of the total overlay (in the X and Y directions) by appropriate combination. Likewise, in order to enable the determination of the total overlay for a plurality of different structures between which overlay may occur, the overlay for each of those structures needs to be determined. Thus, as an example, for a unit cell having 4 different structures in 4 layers (one of which is a reference layer) between which overlay may occur, 6 overlays may be determined (one for each layer in the X and Y directions) to enable the determination of the total overlay for the unit cell. Of course, sub-combinations may be determined as desired to derive one or more different overlays of interest among the 4 layers.

[0238] Fig.18 An example embodiment of a plurality of overlay unit cells of a target is depicted. FIG. 10A to FIG. 10C Similar to the unit cell of , the unit cell includes a first structure 1000 and a second structure 1005. Additionally, the unit cell has a third structure 1800, which in this embodiment is in a layer above the first structure 1000 and the second structure 1005 in the Z direction. In this embodiment, the asymmetry of the unit cell can be formed by one or more different overlays. For example, the relative displacement between the structure 1005 and the structure 1800 in the X direction can produce overlays in the X direction, which leads to asymmetry. As another example, the relative displacement between the structure 1005 and the structure 1000 in the Y direction can produce overlays in the Y direction, which leads to asymmetry. As another example, the relative displacement between the structure 1000 and the structure 1800 in the Y direction can produce other overlays in the Y direction, which leads to asymmetry.

[0239] Fig.19 Another example embodiment of a plurality of overlay unit cells of a target is depicted. FIG. 10A to FIG. 10C Similar to the unit cell of , the unit cell includes a first structure 1000 and a second structure 1005. Fig.18 Similar to the unit cell of , the unit cell has a third structure 1800, which in this embodiment is in a layer above the first structure 1000 and the second structure 1005 in the Z direction. Further, the unit cell has a fourth structure 1900, which in this embodiment is in a layer above the first structure 1000, the second structure 1005 and the third structure 1800 in the Z direction. Fig.18Similar to the unit cell of , in this embodiment, the asymmetry of the unit cell can be formed by one or more different overlays. For example, the relative displacement between the structure 1005 and the structure 1800 in the X direction can produce overlay in the X direction, which leads to asymmetry. As another example, the relative displacement between the structure 1005 and the structure 1900 in the X direction can produce overlay in the X direction, which leads to asymmetry. As another example, the relative displacement between the structure 1005 and the structure 1000 in the Y direction can produce overlay in the Y direction, which leads to asymmetry. As yet another example, the relative displacement between the structure 1000 and the structure 1800 in the Y direction can produce other overlays in the Y direction, which leads to asymmetry.

[0240] Therefore, in one embodiment, if there are actually multiple different overlays, Fig.18 or Fig.19 Measurement of the illumination physical instance of a unit cell will produce optical information that may potentially include multiple different overlays. For example, reference Fig.18 ,if Fig.18 The symmetry of the unit cell of indicates zero overlay, and there is a displacement in X and Y of structure 1005 from its zero overlay position relative to the structure it overlaps (e.g., a displacement in a direction other than 0, 90, 180, or 270 degrees), then the displacement results in asymmetry due to the relative displacement between structure 1005 and structure 1800 in the X direction and the relative displacement between structure 1005 and structure 1000 in the Y direction. Therefore, it is desirable to determine the overlay of structure 1005 in both the X and Y directions (which combine to produce the total overlay of structure 1005).

[0241] As discussed below, a technique is presented that can separately determine a first set of values ​​of a physical instance of a unit cell based on an optical property value relative to a second set of values ​​of the physical instance of the unit cell, where the second set of values ​​can also be obtained from the same optical property value, wherein the first set is in a different direction than the second set (e.g., X-direction set and Y-direction set), or the first set is between different combinations of parts of the unit cell relative to the second set (e.g., a first set between structure 1005 and structure 1800 and a second set between structure 1005 and structure 1000 or a second set between structure 1000 and structure 1800, where the first set and the second set may be in the same direction).

[0242] That is, in one embodiment, weights are determined to decouple a first overlay information in an optical property value from a second (or more) overlay information in the same optical property value. Thus, in one embodiment, by applying a specifically selected weight, the combination of the weight and the optical property value will produce a specific overlay of interest that is distinguishable from other possible overlay information in the same optical property value. In effect, the weight will highlight the overlay of interest and reduce one or more other overlays. Of course, for each overlay of interest, a different set of weights can be constructed so that the optical property values ​​can be processed to produce different values ​​for each overlay of interest in different overlays of interest.

[0243] against Fig. 20 The technique is described in the following figure. Fig. 20 The diagram of presents a graphical representation of the technique, but in reality, there is no need to construct such a diagram, as all processing can be done mathematically without generating such a diagram. Fig.11 However, the models described for other figures in this document (and associated other techniques) may be used.

[0244] Further, the example is presented in terms of deriving a linear version of the weights from the model. That is, in one embodiment, the weights are derived from (the Moore-Penrose pseudo-inverse of) the Jacobian.

[0245] Thus, in this linear case, the Jacobian can be inverted to reconstruct a particular parameter (such as the overlay in a certain direction). However, how the column for the parameter of interest is related to the rest of the columns determines how easy it is to reconstruct that parameter.

[0246] Thus, having, for example, a unit cell of interest (e.g., Fig.18 The nominal contour model of the unit cell can generate at least two vectors. The first set of vectors p1 represents the first interesting overlay within the unit cell (e.g., the X-direction overlay), and the second overlay vector p2 represents a second overlay of interest (eg, Y direction overlay). It will be appreciated that other vectors may be generated for additional overlays of interest.

[0247] Further, for each of the two overlay vectors, one or more pixels of the pupil representation corresponding to the expected measurement of the physical instance of the unit cell are selected. In this embodiment, a pair of pixels is selected for each overlay vector, wherein each pair of pixels includes symmetrically positioned pixels as previously described. Desirably, the pixel pairs are selected from an asymmetric radiation distribution portion of the pupil representation, as discussed above.

[0248] Now, the first set of tick vectors p1The response (in this case, the asymmetric signal between a pair of pixels) corresponds to a change in a first overlay of interest to a first overlay vector (leaving all other parameters constant, i.e., the second overlay of interest constant) in a pixel pair. This response can be generated using a nominal profile model by inducing a change in the first overlay of interest (e.g., a 1 nm change) and then calculating the optical response (e.g., intensity) in the pixel pair to the change.

[0249] Similarly, the second set of tick vectors p2 The response (in this case, the asymmetric signal between the pixels in a pair) corresponds to a change in the second overlay of interest of the second overlay vector (leaving all other parameters unchanged, i.e., the first overlay of interest unchanged) in the pixel pair. The response can be generated using a nominal profile model by inducing a change in the second overlay of interest (e.g., a 1 nm change) and then calculating the optical response (e.g., intensity) in the pixel pair to the change.

[0250] The resulting vector is Fig. 20 , where the horizontal axis u and the asymmetric intensity (I i -I i '), while the vertical axis v corresponds to the asymmetric intensity (I i -I i '). Therefore, Fig. 20 shows two highly correlated vectors p1 and p2 .

[0251] Therefore, to decouple and separate the contributions of the first overlay of interest and the second overlay of interest to the pixel pair, the vector p1 Backprojection to vector to form a vector p1′, which is the vector p2 Orthogonal vectors, and the length of the projected vector p1′ is divided by the vector p1 and This vector then helps isolate the intensity of the first overlay and pixel pair of interest (and, by extension, other pixel pairs in the pupil representation).

[0252] Additionally or alternatively, the vector p2 Backprojection to vector to form a vector p2′, which is the vector p1 Orthogonal vectors, and the length of the projected vector p2′ is divided by the vector p2 and This vector then helps isolate the intensity of the second overlay and pixel pair of interest (and, by extension, other pixel pairs in the pupil representation).

[0253] Therefore, returning to equations (3) and (4), S i The asymmetric intensity (I i -I i '). Therefore, the first set of tick vectors p1 Can be used with S with U0 i The first pixel pair and S with V0 i The response of the second pixel pair to the change of the first overlay of interest corresponds to that of the second pixel pair. Similarly, the second overlay vector p2 may correspond to the responses of those first and second pixel pairs to the change in the second overlay of interest. Thus, vector p1′ and / or vector p2′ may be constructed; both are here for explanation purposes. Vector p1′ and vector p2′ are defined in terms of intensity u corresponding to the first pixel pair corresponding to U0 and in terms of intensity v corresponding to the second pixel pair corresponding to V0. Thus, vector p1′ and vector p2′ may be specified as follows:

[0254] p1 ′=(u1′,v1′) (12)

[0255] p2 ′=(u2′,v2′) (13)

[0256] Therefore, now in the linear context described above and referring to equation (4), the overlay value of the first overlay of interest can be defined based on U0, V0 and vectors p1' and p2' as follows:

[0257]

[0258] Additionally or alternatively, then, an overlay value of a second overlay of interest may be defined based on U0, V0 and vectors p1′ and p2′ as follows:

[0259]

[0260] Therefore, according to equation (14), for U0 and V0 respectively, the weights used to determine the first interesting overlay are as follows:

[0261]

[0262] Further, according to equation (15), for U0 and V0 respectively, the weights used to determine the second interesting overlay are as follows:

[0263]

[0264] It will therefore be appreciated that this operation may be repeated for all or substantially all pairs of pixels in the pupil representation to arrive at the first set of pixels of interest. The weight set w i and / or get a second set of interest The weight set w i One or both of these weights may then be applied to the measured optical property values ​​to derive an overlay value for the corresponding overlay of interest, in accordance with equation (4). Of course, one or more other overlays of interest may be evaluated and one or more appropriate sets of weights determined therefor. It will be appreciated that in one embodiment, the sensitivities (e.g., Jacobians) for all different overlays of interest are included in the definition of weights for a particular overlay of interest.

[0265] Thus, for example, for a unit cell having 4 layers (one of which is a reference layer), where shifts in the X and Y directions of each of the layers may result in a change in symmetry (e.g., resulting in asymmetry, or resulting in other asymmetries, or resulting in an asymmetric unit cell becoming symmetric), 6 vectors (each associated with a different pair of pixels) may be generated, the 6 vectors including an X-direction overlay vector for each layer and a Y-direction overlay vector for each layer. Thus, there may be 6 sets of weights to derive the corresponding overlays. Of course, if one of the vectors is not a vector of interest, all weight sets do not need to be derived (but in one embodiment, the sensitivities (e.g., Jacobians) for all different overlays of interest are included in the weight definition for a particular overlay of interest). Any other overlays may then be determined by appropriate mathematical combinations of two or more of these overlays.

[0266] It will be appreciated that certain shifts of the layers in the unit cell do not result in a change in symmetry and therefore the overlay corresponding to the shift cannot be determined from the unit cell. It is therefore apparent that no vector is defined for such shifts. Fig.18 For example, three vectors may be defined for the unit cell - one vector for overlay in the X direction and two vectors for different overlays in the Y direction. Thus, a set of weights may be determined which, when combined with the measured optical property values, gives an overlay in the X direction. Alternatively, a set of weights may be determined which, when combined with the measured optical property values, gives one of the overlays in the Y direction; and / or a set of weights may be determined which, when combined with the measured optical property values, gives another of the overlays in the Y direction. Of course, all three sets of weights may be determined, or only two sets of weights may be determined.

[0267] The above discussion focuses on a target formed by one or more instances of a symmetrical unit cell composed of the structure of the device. Such a target can make it possible to determine the non-product value of a patterning process parameter by non-product measurement of radiation redirected to the non-product target. However, as described above, the target need not consist only of the device structure. In other words, a non-product target can be provided, the structure of which does not exclusively include the device structure. For example, in one embodiment, a target can be specially generated by a structure that is not used to form the device but only for measurement. Such a target can be provided, for example, in a scribe away from the device (and therefore, provided in a portion of a device patterning pattern away from the device pattern). In one embodiment, a target can be provided in the device pattern (and therefore, in the features of the device pattern of the patterned device pattern). Non-product targets can include one or more device structures and one or more specially generated structures in appropriate cases, which are not used to form the device, but only for measurement.

[0268] Non-product targets may be used if, for example, patterning process parameters are determined for a device pattern that does not present instances of symmetric unit cells. As another example, non-product targets may be useful if, for example, patterning process parameters are determined for a portion of a device pattern that does not have a symmetric unit cell as described above, which can give a measurement of the patterning process parameters. For example, there may be a situation where it is desired to determine the structure of the overlay after etching using the symmetric unit cell method described above, but the overlay structure after etching does not have symmetry. For example, a logic circuit or structure has many process layers / steps, each of which can introduce different overlay components that can destroy the symmetry of the structure. In the case of, for example, a logic circuit, measurements of the device pattern cannot generally be performed due to the lack of symmetric unit cells of the logic circuit structure.

[0269] As another example, non-product targets may be used in association with device patterns that may present symmetric unit cell instances (and even if the unit cells may give measurements of all patterning process parameters of interest). This may be the case, for example, if the device pattern is complex, which may require a significant amount of computation time. Further, the device pattern may present potential crosstalk with signals of patterning process parameters that are not of interest. As an example, pupil correlations of different overlay components may be so large that different overlay errors cannot be separated.

[0270] Therefore, non-product targets can be used with device patterns for instances where the beam spot has symmetrical unit cells or device patterns for instances where the beam spot cannot present symmetrical unit cells. Details of the configuration, design, measurement, and use of non-product targets are described in detail in U.S. Patent Application No. 15 / 445,612 filed on February 28, 2017, the entire contents of which are incorporated herein by reference.

[0271] Obtaining accurate information about the patterning process from the radiation redirected by the measurement target generally requires careful calibration of the measurement process. If the information is contained in an asymmetric component of the detected radiation, asymmetries caused by other sources need to be reduced. Such asymmetries may be referred to as system asymmetries. System asymmetries are asymmetries that occur independently of asymmetries in the measurement target. Imperfections in the optics or sensors of the measurement device used to perform the measurement process may contribute to the system asymmetry. This contribution may be referred to as sensor asymmetry. Alternatively or additionally, errors in the mathematical modeling used to infer the parameter of interest from the detected radiation (e.g., as described above with reference to Fig. 9 The contribution of ) may contribute to the system asymmetry. This contribution can be called model asymmetry. Figures 13 to 17 As described, data-driven techniques may be used in place of or in addition to mathematical modeling. Data-driven techniques may involve processing data measured from one or more substrates having one or more measurement targets ("obtaining" data) using one or more intentionally set values ​​for a parameter of interest (e.g., overlay). This combination of "setting" intentional values ​​for the parameters of interest and measurements of the corresponding measurement targets is referred to as a "set-obtain" process. In one embodiment, multiple different set values ​​for overlay are created across one or more substrates. Measurements of measurement targets using different set values ​​for the parameters of interest can be used to determine the sensitivity of the detected radiation from the measurement targets to changes in the values ​​of the parameters of interest. In embodiments based on data-driven techniques, model asymmetry may not contribute at all (in the case where modeling is not used), or may contribute less (in the case where data-driven techniques supplement mathematical modeling). Therefore, system asymmetry may include sensor asymmetry or a combination of sensor asymmetry and model asymmetry.

[0272] One method for reducing the effects of system asymmetry is to perform multiple measurements on the substrate W at different rotational positions of the substrate W, including, for example, measurements of the substrate W at θ degrees and θ+180 degrees (for any θ). This method allows the effects of system asymmetry to be at least partially cancelled but significantly increases the time required to perform the metrology process. Furthermore, drifts in the state of the metrology apparatus between measurements performed at different rotational positions of the substrate W (which must be performed at different times) can adversely affect the extent to which the effects of system asymmetry are cancelled.

[0273] Before the metrology target itself is measured, the system asymmetry can be determined using calibration measurements. The determined system asymmetry can then be subtracted from the measurements from the metrology target to obtain the parameter of interest. The extent to which the system asymmetry can be eliminated in this way is limited by the drift of the state of the metrology device between the calibration measurement and the measurement of the metrology target, and by the difference in the excitation of the system asymmetry by the target used for calibration compared to the target used for metrology.

[0274] An arrangement is described below which reduces the real-time impact of errors such as system asymmetry and does not require measurements to be performed at different rotational positions of the substrate W (although measurements may still be performed at different rotational positions of the substrate W if desired to further improve accuracy). As a result, information about the patterning process may be obtained faster, more accurately and / or more reliably.

[0275] Fig.21 An exemplary method of determining information about a patterning process is depicted. The method includes a data acquisition step 320, which includes obtaining measurement data 340 from a metrology process. In one embodiment, using the above reference Figure 7 and Figure 8 The measurement process is performed using the measurement device described.

[0276] like Fig. 22 As schematically depicted in , the metrology process is applied to each of a plurality of metrology targets T on a substrate W. In the example shown, the metrology process is applied to three different metrology targets T and therefore includes three units of measurement data 340 (labeled 340A to 340C) corresponding to each of the three metrology targets T, respectively.

[0277] The measurement data 340 for each measurement target T includes at least a first contribution 341 and a second contribution 342. The first contribution comes from the parameters of interest of the patterning process used to form the measurement target T. The second contribution 342 comes from errors in the measurement process used to measure the measurement target T. In this embodiment, the measurement data includes a symmetric component of the detected pupil representation as a third contribution 343. The role of the symmetric component of the detected pupil representation is described below.

[0278] The method further comprises a data processing step 321. The data processing step 321 comprises: using measurement data obtained from all measurement targets in a plurality of measurement targets T from which the measurement data was obtained in step 320 to obtain information about errors in the measurement process. The data processing step 321 further comprises: using the obtained information about errors in the measurement process to extract a value of a parameter of interest for each measurement target T. In one embodiment, the obtained information about the errors is used to reduce or eliminate the effect of the errors on the accuracy of the obtained values ​​of the parameters of interest. In one embodiment, the parameters of interest include overlay errors between different layers in the measurement target T.

[0279] The method is capable of extracting a value of a parameter of interest with reduced or eliminated errors by using measurement data from a plurality of metrology targets T, which are optionally spaced apart from one another on a substrate W. Using measurement data from a plurality of metrology targets T effectively provides a plurality of equations equal to or greater than the number of unknowns (including the errors), thereby allowing the errors to be at least partially decorrelated from information about the value of the parameter of interest in the measurement data.

[0280] In one embodiment, for each metrology target, the contribution to the measurement data from the parameter of interest is known (e.g., sensitivity to overlay) based on the expected variation in the value of the parameter of interest (e.g., from modeling or from calibration data in data-driven techniques, as described above). The expected variation may be the same for different metrology targets T, or may be different for different metrology targets T.

[0281] In one embodiment, the plurality of measurement targets T are nominally identical to one another. This simplifies the data processing required to implement the data processing step 321 (e.g., when used, as described above with reference to Fig. 9 modeling as described).

[0282] In one embodiment, the error in the measurement process is substantially the same for each measurement target in the plurality of measurement targets T. This improves the accuracy of the extracted value of the parameter of interest by increasing the extent to which the error in the measurement process can be removed.

[0283] A detailed example will now be described. The example illustrates the application of the method in the case where the first contribution and the second contribution to the measurement data comprise contributions to an asymmetric component of the detected pupil representation of radiation redirected by the measurement target T. In embodiments of this type, the measurement data may comprise (or be derived from) measurements of optical characteristic values ​​of pixel pairs in the detected pupil representation. In one embodiment, at least two pixel pairs are used for each measurement target (i.e., the asymmetry is evaluated in at least two different parts of the detected pupil representation). The pixels in each pixel pair may be point symmetric with respect to a common symmetry point in the pupil plane, or may be mirror symmetric with respect to a common mirror symmetry line in the pupil plane. The optical characteristic values ​​may comprise intensity or phase. As described above with reference to Fig. 10A and Fig. 10B As described, the asymmetric component of the detected pupil representation can be obtained, for example, by subtracting the intensity of a symmetrically located pixel from the intensity at a particular pixel in the detected pupil representation, or vice versa. In this type of embodiment, each measurement target T can be in a nominal physical configuration (e.g., Fig. 10A The physical configuration of the measurement target T that differs from the nominal physical configuration (e.g. due to overlay, such as Fig. 10B ) contributes to the asymmetric component of the detected pupil representation (a first contribution to the measurement data). The parameter of interest (eg, overlay) measures the change in the physical configuration.

[0284] In the case where the parameter of interest is overprint, the asymmetric component of the detected pupil representation for each measurement target T can be described by the sum of: 1) the sensitivity to overprint multiplied by the overprint (e.g., derived from the Jacobian matrix and / or data-driven techniques as described above); and 2) the asymmetric component of the error multiplied by the symmetric component of the detected pupil representation. The number of equations provided by the measurement data is equal to half the number of detected pixels in the detected pupil representation (a 1 / 2 factor in the number of pixels occurs since only the asymmetric component of the detected pupil representation is used). The number of unknowns is greater than or equal to half the number of detected pixels + the number of overprints to be detected. Therefore, the system of equations is underdetermined and cannot be solved uniquely.

[0285] Using data from multiple measurement targets T increases the number of equations that make a solution possible. In one embodiment, the number of measurement targets T used is greater than the number of parameters of interest (e.g., overlays) to be extracted from each measurement target T. This ensures that the number of equations is equal to or greater than the total number of unknowns, even if the number of unknowns increases due to the need to extract a greater number of parameters of interest.

[0286] The accuracy is high when the asymmetric error from the measurement process is relatively stable in the measurement. Assuming that the asymmetric error from the measurement process is relatively stable, the following set of equations can be provided:

[0287]

[0288] Among them, aP t is the asymmetric component of the pupil representation measured for the measurement target t, is the overlay of the measurement target t, sP t is the symmetric component of the detected pupil representation, J t is the Jacobian matrix of the measurement target t (representing the expected sensitivity of the measurement process to overlay), and a ∈ is the unknown asymmetric error from the measurement process. In other embodiments, as described above, the Jacobian matrix of the measurement target t can be replaced with weights representing the expected sensitivity derived from data-driven techniques.

[0289] This embodiment is an example of a class of embodiments in which the contribution to the asymmetric component of the detected pupil representation resulting from the system asymmetry is a linear function of the symmetric component of the detected pupil representation (this may also be referred to as a linear mapping from the symmetric component of the detected pupil representation to the asymmetric component of the detected pupil representation). In this particular example, the linear function takes the form sP t ·a ∈ , but other forms are possible.

[0290] If for multiple measurement targets t, the asymmetric error a from the measurement process ∈ Basically the same, the above equations become A linear system of equations of the form . In this equation, is a vector containing measurement data for each metrology target (including a first contribution 341 and a second contribution 342 to the measurement data 340 ), is a vector containing, for each metrology target, the unknown values ​​of the parameter of interest derived from the first contribution to the measurement data and the unknown second contribution to the measurement data (the asymmetric error). A is a matrix representing the expected sensitivity of the metrology process to the parameter of interest (in this example, the overlay, and containing the Jacobian for each metrology target in this example) and the symmetric component of the detected pupil representation (provided by the third contribution 343 to the measurement data 340).

[0291] is a system of linear equations, where the number of equations is equal to half the number of detected pixels in the detected pupil representation multiplied by the number of measurement targets T. The number of unknowns is equal to the sum of the product of the number of overlays to be determined and the number of measurement targets and the number of asymmetric error patterns. As long as the number of equations is equal to or greater than the number of unknowns, the system of linear equations can be solved by various known techniques. In the case where the asymmetric errors between the detected pixels are uncorrelated, the number of asymmetric error patterns is equal to half the number of detected pixels in the detected pupil representation.

[0292] In one embodiment, the method takes into account the correlation of the asymmetric errors between the detected pixels of the detected pupil representation. will be a vector containing both overlay and asymmetry errors. Asymmetric errors may include contributions from one or both of sensor asymmetry and model asymmetry. Both types of errors have similar effects on the inference of the parameter of interest (overlay in this example). The method corrects for both errors jointly.

[0293] The method is not limited to detected pupil representations comprising pixels. The method is applicable to any method for measuring the asymmetry of a measurement target by measuring an asymmetric component of a signal, wherein the signal comprises an asymmetric error contribution interacting with a symmetric component of the signal, and wherein measurement data from different measurement targets are combined together according to a sampling scheme. As can be seen from the above equation, the method works when the number of input signals in the asymmetric part of the signal is greater than 1 and the sampling scheme comprises enough measurement targets. As in the above example, the multiple input signals may come from multiple detected pixels representing asymmetry (e.g., derived by taking the difference between pairs of symmetrically positioned detected pixels in the detected pupil representation), but may also be different wavelengths (e.g., in the context of spectral ellipsometry), different polarizations, or different signals obtained in a phase-based measurement. Thus, the measurement process may include one or both of the following: irradiating the measurement target T with radiation of different wavelength bands and detecting redirected radiation of each wavelength band separately; and irradiating the measurement target T with radiation with different polarizations and detecting redirected radiation with different polarizations separately.

[0294] Reference below Fig.23 and Fig.24 An illustrative simplified example is described.

[0295] Fig.23 and Fig.24 Detected pupil representations of two metrology targets T at different positions on a substrate W are depicted respectively. Fig.23The measurement data corresponding to the measurement target T includes intensity values ​​for two pairs of pixels (corresponding to parts of the intensity distribution in the pupil plane): I1 and I3, and I2 and I4, respectively. Fig.24 The measurement data of the measurement target T includes intensity values ​​for two corresponding pairs of pixels: I1 ′ and I3 ′, and I2 ′ and I4 ′, respectively.

[0296] Information about the parameter of interest is contained in the asymmetric component of the detected pupil representation, which in the ideal case would be equal to the difference between the measured intensities for each pair of pixels. In practice, the measured intensities are subject to an error ∈ i The influence of (i.e., including contributions from one or both of sensor asymmetry and model asymmetry) makes each measured intensity given as follows:

[0297] I1=(1+∈1)I N1

[0298] I2=(1+∈2)I N2

[0299] I3=(1+∈3)I N3

[0300] I4=(1+∈4)I N4

[0301] I1′=(1+∈1)I′ N1

[0302] I2′=(1+∈2)I′ N2

[0303] I3′=(1+∈3)I′ N3

[0304] I4′=(1+∈4)I′ N4

[0305] Among them I N1 Etc. represents the intensity that would be measured without any error.

[0306] Fig.23 The difference in measured intensity between the upper pixel pairs shown can be written as follows:

[0307]

[0308] The last term on the right-hand side of the above equation is usually much larger than the second term, in which case the second term can be ignored. represents the asymmetric component of the error in the measurement process, which can be expressed as E1. Corresponding expressions can be obtained for each of the other three pixel pairs.

[0309] First, in the absence of asymmetric errors, the parameter of interest α is linearly proportional to the asymmetric component of the measured intensity. Thus, for Fig.23 The upper pixel pair in (I N1 -I N3 )=S1·α, where S1 is a coefficient representing the ratio (or sensitivity). According to the above, the measurement of four pixel pairs results in the following four equations with four unknown quantities:

[0310] I1-I3=S1·α+E1·(I N1 +I N3 )

[0311] I2-I4=S2·α+E2·(I N2 +I N4 )

[0312] I1′-I3′=S1·α′+E1·(I N1 ′+I N3 ′)

[0313] I2′-I4′=S2·α′+E2·(I N2 ′+I N4 ′)

[0314] in and (I N1 +I N3 )≈(I1+I3), (I N2 +I N4 )≈(I2+I4), (I N1 ′+I N3 ′)≈(I1′+I3′), and (I N2 ′+I N4 ′)≈(I2+I4). Therefore, the system of equations can be solved to obtain four unknowns: α, α′, E1, and E2.

[0315] Reference above Figure 21 to Figure 24 The method described uses measurements from multiple measurement targets T to create a system of equations, where the number of equations is equal to or greater than the number of unknowns. Specific examples are given for situations where only a single parameter of interest is required (e.g., a single overlay), but the method can be extended to extract any number of parameters of interest by increasing the number of measurement targets T used appropriately. In one embodiment, a method of selecting measurement targets for a method for determining information about a patterning process is provided, wherein the measurement targets T are selected so that the number of measurement targets is greater than the number of parameters of interest. Thus, for a single overlay, a minimum of two measurement targets T are required. For two overlays, a minimum of three measurement targets T are required, and so on. Reference above Figures 18 to 20 A unit cell of an example metrology target T with multiple overlays of interest is discussed in detail.

[0316] In one embodiment, the plurality of metrology targets is selected to include all instances of a predetermined metrology target type on a substrate W, wherein all instances have the same nominal structure. The plurality of metrology targets may or may not include a subset of all metrology targets on a substrate W. The information obtained about errors in the metrology process may therefore include average errors (averaged spatially over a portion of the substrate surface containing the metrology targets T contributing to the obtained information, and / or averaged temporally over a period of time during which the metrology targets T contributing to the obtained information were measured). In one embodiment, sensor asymmetry may be monitored for drift and fed forward to improve the accuracy of subsequent measurements.

[0317] In an alternative embodiment, a plurality of measurement targets are composed of a subset of all instances of a predetermined measurement target type on a substrate W, wherein all instances have the same nominal structure. Therefore, a specific subset of the measurement targets is selected. A specific subset can be selected to provide a more accurate value of one or more parameters of interest. In one embodiment, the method is repeatedly performed using different subsets of the measurement targets. Therefore, information about errors in the measurement process can be repeatedly obtained using different subsets of the measurement targets. In this type of embodiment, information about errors in the measurement process can be updated step by step in real time. In one embodiment, different subsets of the measurement target T are obtained by gradually adding measurement targets T for which new measurement data is available to the subset, and removing measurement targets T for which only old measurement data is available from the subset. Therefore, at any given time, the subset contains only (or mainly contains) the most recently measured measurement targets T, and therefore will be less affected by errors caused by drift of sensor asymmetry between measurements of different measurement targets T, for example. Gradually changing the subset also helps to identify anomalies that can significantly change sensor asymmetry. The identified anomalies can be further investigated and processed.

[0318] In one embodiment, the above method is used to reduce errors in measurement data. A method is provided in which measurement data is obtained from a measurement process applied to each of a plurality of measurement targets T on a substrate W. The measurement data for each measurement target T includes at least a first contribution and a second contribution. The first contribution comes from a parameter of interest of a patterning process used to form the measurement target T. The second contribution comes from errors in the measurement process used to measure the measurement target T. The measurement data obtained from all the plurality of measurement targets is used to obtain information about errors in the measurement process. The information obtained is used to reduce errors in the measurement data.

[0319] In one embodiment, the above method is used to calibrate a metrology process. A method is provided in which measurement data is obtained from a metrology process applied to each of a plurality of metrology targets T on a substrate W. The measurement data for each metrology target T includes at least a first contribution and a second contribution. The first contribution comes from a parameter of interest of a patterning process used to form the metrology target T. The second contribution comes from an error in the metrology process used to measure the metrology target T. The measurement data obtained from all the plurality of metrology targets is used to obtain information about the error in the metrology process. The information obtained is used to calibrate the metrology process to reduce the error in the measurement data generated by the metrology process.

[0320] In another embodiment, a set of targets on a wafer are measured. Additionally, a subset of the set is measured with the wafer rotated 180°. A sensor asymmetry model is created by fitting the set of measurements and the subset of measurements. The sensor asymmetry model is used to correct the measurement of a parameter of interest (e.g., overlay), which correction is performed according to previous embodiments of the invention.

[0321] In another embodiment, a set of measurements are performed using a wafer rotated 180°. These points are used to create a correction model. Using the correction model, the sensor asymmetry is estimated at multiple locations that are different from the locations where the measurements were performed to obtain the input for creating the correction model. The locations where the correction model is used to estimate are called secondary points. Using the sensor asymmetry of these secondary points, a new system asymmetry correction model is generated, which is further used in the correction scheme required to estimate the accurate parameters of the lithography process (e.g., overlay). In addition, the sampling scheme required to select the position where the measurement is performed to create the correction model can be static (fixed value of the target position). In addition, the sampling scheme can be dynamic, where the information collected from the previous wafer is used to calibrate the measurement of the next wafer in the batch.

[0322] In another embodiment of the invention, it is assumed that the sensor asymmetry is described as a function of a symmetric pupil. The symmetric pupil is obtained from a measured pupil to which a symmetry operation is applied. Furthermore, the function may be a parametric model that is learned via a statistical learning method such as linear regression, principal component regression, partial least squares, or canonical correlation analysis. The parametric model may be a linear model (i.e., a linear transformation) or a nonlinear model (i.e., a Gaussian process or a neural network). The input to the learning process is a set of symmetric pupils as described above, and the output is an antisymmetric error in pupil or overlay space that has been measured for a wafer rotation target. On a single directional target, the measured symmetric pupil is fed to the learned model to obtain the sensor error.

[0323] Furthermore, a given measurement configuration scheme contains: a) an acquisition profile (a set of parameters of the optical system, such as wavelength, polarization, dose or aperture); b) a measurement profile that allows overlay measurements from the measured raw signal (e.g., from the measured pupil); and c) a wafer layout or sampling scheme that indicates the targets to be measured.

[0324] Furthermore, in one embodiment, a calibration is presented which comprises a method of obtaining a model comprising: measuring a set of targets at relative wafer-sensor angles, feeding these measurements to a learning algorithm as described above, wherein the output of the learning algorithm is a model in which a symmetric pupil can be used as input.

[0325] In addition, in one embodiment, a measurement step is proposed, including: measuring another subset of targets on the chip, which another subset may include all targets on the chip, or a set formed only by targets not used in the calibration step described in the previous embodiment; making the measured pupil obtained in the measurement symmetrical; and using the symmetrized pupil as input to the model determined in the calibration step.

[0326] In yet another embodiment, a method of determining sensor asymmetry comprises: an acquisition step, wherein all targets are measured, wherein a subset of the measurements are performed for two sensor-wafer orientation angles and another subset of the measurements are performed for a single orientation; a symmetrization step, as described above, wherein the targets measured in the wafer rotation mode are used as input to a learning algorithm or model; a calculation step, wherein the output of the learning algorithm is a model to be used for measurements obtained on targets in a single orientation, and wherein the sensor asymmetry is obtained by feeding a symmetry pupil to the model.

[0327] In another embodiment of the invention, an off-line calibration method is proposed. In a first step, data acquired using two wafer-sensor orientation angles is used to build a model, where the input to the learning method is the measured symmetric pupil and antisymmetric sensor errors, and the output is the model. In the measurement step (considered the "online" step or real-time step), only the measurements made on the target in a single acquisition are used, which are then fed back to the learning algorithm.

[0328] Furthermore, other sensor information carrying channels can be used in the above calibration method. These can be the target position, the full pupil (without the symmetrization step), the asymmetric pupil obtained from the full measurement pupil, or the temperature of the optical system.

[0329] In another embodiment of the present invention, it is assumed that the sensor asymmetry is described as a function that varies smoothly across the wafer. The sensor asymmetry is obtained in this assumption as an interpolation (such as a bilinear interpolation) of a set of measurements obtained at a single orientation, where the domain parameter is the target position.

[0330] Determining the correct configuration is critical to determining accurate overlay. Training an overlay model suitable for measuring in-device targets is obtained on a metrology target located on a scribe lane, for example, as known from U.S. Patent Application No. US16 / 178,638, which is incorporated herein in its entirety. However, a model trained on a metrology target can lead to possible metrology errors when used on an in-device target, because the processing conditions (one of many examples) can be different where the metrology target is formed and where the in-device target is to be measured.

[0331] The purpose of the following is to propose configuration scheme selection that allows accurate measurement on the in-device target. Configuration scheme selection includes: obtaining measurement results from the measurement target; obtaining measurement results from the in-device target; using both the measurement target measurement results and the in-device measurement measurement results to determine the configuration scheme for accurate measurement.

[0332] In one embodiment, a method is provided for correcting or obtaining a mapping from measurement results to parameters of interest (such as overlay) for metrology. The metrology targets include known deviations (e.g., in nanometers, among other parameters of interest), so by obtaining metrology measurements from these targets, one is able to train a model because the overlay values ​​corresponding to each target are known. This is not the case when one obtains metrology measurements from in-device metrology targets. The overlay values ​​(or values ​​of the parameters of interest) are unknown. In order to proceed with the step of determining the best configuration for precision measurement, it is necessary to first obtain the parameters of interest from the metrology measurements from the in-device targets.

[0333] In one embodiment, the in-device target used to provide the metrology measurement results in the configuration selection method mentioned in the previous paragraph is based on or placed or positioned in close proximity to the metrology target. In this way, it is assumed that the harmful processing conditions of both the metrology target and the in-device target are similar, so the overlay mapping (how to obtain the metrology parameters of interest, such as pupil images or image plane images, from the measurement results) is similar.

[0334] Furthermore, in one embodiment, the metrology measurements obtained from the metrology target and from the in-device target may be in the pupil plane or in the image plane. In one embodiment, the metrology measurements obtained from the metrology target and from the in-device target may be symmetric portions of the measurements in the pupil plane. In one embodiment, the metrology measurements obtained from the metrology target and from the in-device target may be antisymmetric portions of the measurements in the pupil plane. In one embodiment, the metrology measurements obtained from the metrology target and from the in-die target may be symmetric portions and antisymmetric portions of the measurements in the pupil plane.

[0335] Furthermore, the determining step includes using both the metrology target measurements and the intra-device metrology target measurements that are corrected, the correction being based on measurements based on a closely located intra-device target that is not necessarily located proximate to the metrology target.

[0336] Reference Fig.25 , a computer system 3200 is shown. The computer system 3200 includes a bus 3202 or other communication mechanism for conveying information, and a processor 3204 (or multiple processors 3204 and 3205) coupled to the bus 3202 for processing information. The computer system 3200 also includes a main memory 3206, such as a random access memory (RAM) or other dynamic storage device, which is coupled to the bus 3202 for storing information and instructions to be executed by the processor 3204. The main memory 3206 can also be used to store temporary variables or other intermediate information during the execution of instructions to be executed by the processor 3204. The computer system 3200 also includes a read-only memory (ROM) 3208 or other static storage device, which is coupled to the bus 3202 for storing static information and instructions for the processor 3204. A storage device 3210 such as a magnetic disk or optical disk is provided and coupled to the bus 3202 for storing information and instructions.

[0337] The computer system 3200 can be coupled to a display 3212 via bus 3202, such as a cathode ray tube (CRT) or a flat panel or touch panel display for displaying information to a computer user. An input device 3214 including alphanumeric keys and other keys is coupled to bus 3202 for communicating information and command selections to processor 3204. Another type of user input device is a cursor control 3216, such as a mouse, trackball, or cursor direction keys, for communicating direction information and command selections to processor 3204 and controlling cursor movement on display 3212. The input device typically has two degrees of freedom in two axes, a first axis (e.g., x) and a second axis (e.g., y), which allows the device to specify a position in a plane. A touch panel (screen) display can also be used as an input device.

[0338] The computer system 3200 may be adapted to function as a processing unit herein in response to the processor 3204 executing one or more sequences of one or more instructions contained in the main memory 3206. Such instructions may be read into the main memory 3206 from another computer-readable medium such as the storage device 3210. Executing the sequence of instructions contained in the main memory 3206 causes the processor 3204 to perform the processes described herein. One or more processors in a multi-processing arrangement may also be employed to execute the sequence of instructions contained in the main memory 3206. In alternative embodiments, hard-wired circuitry may be used in place of or in combination with software instructions. Therefore, the embodiments are not limited to any specific combination of hardware circuitry and software.

[0339] As used herein, the term "computer-readable medium" refers to any medium that participates in providing instructions to processor 3204 for execution. Such medium can take many forms, including but not limited to non-volatile media, volatile media and transmission media. Non-volatile media include, for example, optical disks or disks, such as storage device 3210. Volatile media include dynamic memory, such as main memory 3206. Transmission media include coaxial cables, copper wires and optical fibers, which include wires that constitute bus 3202. Transmission media can also take the form of sound waves or light waves, such as sound waves or light waves generated during radio frequency (RF) and infrared (IR) data communications. Common forms of computer-readable media include, for example, floppy disks, flexible disks, hard disks, magnetic tapes, any other magnetic media, CD-ROMs, DVDs, any other optical media, punch cards, paper tapes, any other physical media with hole patterns, RAMs, PROMs and EPROMs, FLASH-EPROMs, any other memory chips or boxes, carrier waves described below, or any other media that a computer can read from.

[0340] Various forms of computer-readable media may be involved in carrying one or more sequences of one or more instructions to the processor 3204 for execution. For example, the instructions may first be carried on a disk of a remote computer. The remote computer may load the instructions into its dynamic memory and then send the instructions over a telephone line using a modem. A modem local to the computer system 3200 may receive the data over the telephone line and convert the data into an infrared signal using an infrared transmitter. An infrared detector coupled to the bus 3202 may receive the data carried in the infrared signal and place the data on the bus 3202. The bus 3202 carries the data to the main memory 3206, from which the processor 3204 retrieves and executes the instructions. The instructions received by the main memory 3206 may optionally be stored on the storage device 3210 before or after execution by the processor 3204.

[0341] The computer system 3200 may also include a communication interface 3218 coupled to the bus 3202. The communication interface 3218 provides a two-way data communication coupled to a network link 3220 connected to a local area network 3222. For example, the communication interface 3218 may be an integrated services digital network (ISDN) card or a modem to provide a data communication connection with a corresponding type of telephone line. As another example, the communication interface 3218 may be a local area network (LAN) card to provide a data communication connection with a compatible LAN. A wireless link may also be implemented. In any such implementation, the communication interface 3218 sends and receives electrical, electromagnetic, or optical signals that carry digital data streams representing various types of information.

[0342] The network link 3220 typically provides data communication to other data devices through one or more networks. For example, the network link 3220 can provide a connection to a host 3224 or a data device operated by an Internet Service Provider (ISP) 3226 through a local network 3222. The ISP 3226, in turn, provides data communication services through a global packet data communication network (now, collectively referred to as the "Internet" 3228). Both the local network 3222 and the Internet 3228 use electrical, electromagnetic, or optical signals that carry digital data streams. The signals through the various networks and the signals on the network link 3220 and through the communication interface 3218 are exemplary forms of carrier waves that transport the information, and these signals carry the digital data to and from the computer system 3200.

[0343] The computer system 3200 can send messages and receive data including program code through one or more networks, network link 3220 and communication interface 3218. In the Internet example, server 3230 can transmit the requested code of the application program through Internet 3228, ISP 3226, local network 3222 and communication interface 3218. According to one or more embodiments, one such downloaded application provides, for example, the method disclosed herein. The received code can be executed by processor 3204 when received, and / or stored in storage device 3210 or other non-volatile memory for later execution. In this way, computer system 3200 can obtain application code in the form of a carrier wave.

[0344] Embodiments of the present disclosure may take the form of a computer program or a data storage medium, the computer program comprising one or more machine-readable instruction sequences describing the methods disclosed herein, the data storage medium (e.g., a semiconductor memory, a magnetic disk, or an optical disk) storing such a computer program. Further, the machine-readable instructions may be embodied in two or more computer programs. Two or more computer programs may be stored on one or more different memories and / or data storage media.

[0345] When one or more computer programs are read by one or more computer processors located in at least one component of the lithographic apparatus, any controller described herein, each or in combination, may be operable. Each or in combination of the controllers may have any suitable configuration for receiving, processing and sending signals. One or more processors are configured to communicate with at least one of the controllers. For example, each controller may include one or more processors for executing a computer program, which includes machine-readable instructions for the method described above. The controller may include a data storage medium for storing such a computer program, and / or hardware for receiving such a medium. Therefore, one or more controllers may operate according to the machine-readable instructions of one or more computer programs.

[0346] Although specific reference may be made herein to the use of metrology apparatus in the manufacture of ICs, it should be understood that the metrology apparatus and processes described herein may have other applications, such as the manufacture of integrated optical systems, for guiding and detecting patterns for magnetic domain memories, flat panel displays, liquid crystal displays (LCDs), thin film heads, etc. Those skilled in the art will understand that in the context of such alternative applications, any use of the terms "wafer" or "die" herein may be considered synonymous with the more general terms "substrate" or "target portion", respectively. Before or after exposure, the substrates referred to herein may be processed, for example, in a track (a tool that typically applies a resist layer on a substrate and develops the exposed resist), a metrology tool, and / or one or more other tools. Where applicable, the disclosure herein may be applicable to such and other substrate processing tools. Further, a substrate may be processed, for example, more than once, in order to produce a multi-layer IC, so the term substrate used herein may also refer to a substrate that already contains multiple processed layers.

[0347] Although the above may have made specific reference to the use of embodiments of the present disclosure in the context of optical lithography, it will be appreciated that the present disclosure may be used in other applications, such as nanoimprint lithography, and is not limited to lithography where the context permits. In the case of nanoimprint lithography, the patterning device is an imprint template or mold.

[0348] The terms "radiation" and "beam" as used herein encompass all types of electromagnetic radiation, including ultraviolet (UV) radiation (e.g., having a wavelength of or about 365 nm, 355 nm, 248 nm, 193 nm, 157 nm, or 126 nm) and extreme ultraviolet (EUV) radiation (e.g., having a wavelength ranging between 5 nm and 20 nm), as well as particle beams (such as ion beams or electron beams).

[0349] The term "lens," where the context permits, may refer to any one or combination of various types of optical components, including refractive, reflective, magnetic, electromagnetic, and electrostatic optical components.

[0350] References herein to crossing or passing a threshold include things having a value less than a particular value or less than or equal to a particular value, things having a value greater than a particular value or greater than or equal to a particular value, things being ranked higher or lower than other things (e.g., by sorting) based on (e.g.) a parameter, etc.

[0351] References to correcting or correction of errors herein include eliminating the error or reducing the error to within a tolerance range.

[0352] As used herein, the terms "optimizing" and "optimization" refer to or mean adjusting a lithographic apparatus, a patterning process, etc., so that the result and / or process of the lithographic or patterning process has more desired characteristics, such as higher projection accuracy of the design layout on the substrate, a larger process window, etc. Therefore, the term "optimization" as used herein refers to or means a process of identifying one or more values ​​of one or more variables, which provides an improvement in at least one related indicator, such as a local optimum, compared to an initial set of one or more values ​​of the one or more variables. "Optimal" and other related terms should be interpreted accordingly. In one embodiment, the optimization steps can be iteratively applied to provide other improvements in one or more indicators.

[0353] In the optimization process of the system, the quality factor of the system or process can be expressed as a cost function. The optimization process is reduced to the process of finding a set of parameters (design variables) of the system or process that optimizes (e.g., minimizes or maximizes) the cost function. The cost function can have any suitable form depending on the goal of the optimization. For example, the cost function can be the weighted root mean square (RMS) of the deviation of certain characteristics (evaluation points) of the system or process relative to the expected values ​​(e.g., ideal values) of these characteristics; the cost function can also be the maximum value (i.e., the worst deviation) among these deviations. The term "evaluation point" herein should be interpreted broadly to include any characteristic of the system or process. The design variables of the system can be limited to the interdependence and / or limited range of the practice due to the implementation of the system or process. In the case of a lithographic apparatus or patterning process, the constraints are usually associated with the physical properties and characteristics of the hardware (such as adjustable range, and / or design rules for manufacturability of the patterning device), and the evaluation points can include physical points on the resist image on the substrate and non-physical characteristics (such as dose and focus).

[0354] Although specific embodiments of the present disclosure have been described above, it will be appreciated that the present disclosure may be practiced in a manner other than as described. For example, the present disclosure may take the form of a computer program comprising one or more machine-readable sequences of instructions describing the methods disclosed above, or a data storage medium (e.g., a semiconductor memory, a magnetic disk, or an optical disk) in which such a computer program is stored.

[0355] In the block diagram, the illustrated components are depicted as discrete functional blocks, but embodiments are not limited to systems in which the functions described herein are organized as illustrated. The functions provided by each of the components can be provided by software modules or hardware modules organized in different ways from those currently depicted, for example, such software or hardware can be mixed, combined, replicated, decomposed, distributed (e.g., in a data center or geographically), or organized in different ways. The functions described herein can be provided by one or more processors of one or more computers executing code stored on a tangible, non-transient machine-readable medium. In some cases, a third-party content distribution network may host some or all of the information transmitted over the network, in which case, to the extent that information (e.g., content) is supplied or otherwise provided, the information can be provided by sending a command to retrieve the information from the content distribution network.

[0356] Unless otherwise expressly stated, it will be apparent from the discussion that it will be understood that throughout the specification, discussions utilizing terms such as "processing," "computing," "calculating," "determining," etc., refer to the actions or processes of a specific apparatus, such as a special purpose computer or similar special purpose electronic processing / computing device.

[0357] The reader will appreciate that this application describes several inventions. The applicant has not divided these inventions into multiple isolated patent applications, but has grouped these inventions into a single document because their related themes can make the application process more economical. However, the unique advantages and aspects of these inventions should not be lumped together. In some cases, the embodiments solve all the defects noted herein, but it should be understood that the present invention is independently useful, and some embodiments only solve a subset of these problems, or provide other unmentioned benefits that are obvious to those skilled in the art who view this disclosure. Due to cost constraints, some of the inventions disclosed herein may not currently be claimed for protection, and may be claimed in a later application (such as a continuation application or by amending the claims). Similarly, due to space limitations, the invention summary or the invention content section of this document should not be considered as a comprehensive listing of all such inventions or all aspects of such inventions.

[0358] It should be understood that the specification and drawings are not intended to limit the invention to the particular forms disclosed, but on the contrary, the intention is to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the invention as defined by the appended claims.

[0359] Other embodiments according to the present invention are further described in the following numbered clauses:

[0360] 1. A method of determining information about a patterning process, comprising:

[0361] Measurement data is obtained from a metrology process applied to each of a plurality of metrology targets on a substrate, wherein:

[0362] The measurement data for each metrology target includes at least a first contribution from a parameter of interest of a patterning process used to form the metrology target and a second contribution from an error in the metrology process used to measure the metrology target; and

[0363] The method further comprises:

[0364] obtaining information about errors in the measurement process using the measurement data obtained from all of the plurality of measurement targets; and

[0365] Using the information obtained about the errors in the measurement process, the value of the parameter of interest for each measurement target is extracted

[0366] 2. A method according to clause 1, wherein the plurality of measurement targets are nominally identical to each other.

[0367] 3. A method according to clause 1 or 2, wherein the error in the measurement process is substantially the same for each measurement target in the plurality of measurement targets.

[0368] 4. A method according to any of the preceding clauses, wherein each measurement target has geometric symmetry under a nominal physical configuration, wherein a physical configuration of the measurement target that is different from the nominal physical configuration causes an asymmetry in the measurement data, and the parameter of interest measures a change in the physical configuration.

[0369] 5. A method according to any of the preceding clauses, wherein the metrology process comprises illuminating each metrology target with radiation and detecting radiation redirected by each metrology target.

[0370] 6. A method according to clause 5, wherein the measurement data comprises a detected representation of the redirected radiation.

[0371] 7. A method according to clause 6, wherein the detected representation comprises a detected pupil representation.

[0372] 8. A method according to clause 7, wherein the first contribution to the measurement data comprises a contribution to an asymmetric component of the detected pupil representation.

[0373] 9. A method according to clause 8, wherein the second contribution to the measurement data comprises a contribution to an asymmetric component of the detected pupil representation.

[0374] 10. A method according to clause 9, wherein the contribution to the asymmetric component of the detected pupil representation from the second contribution to the measurement data comprises a linear function of the symmetric component of the detected pupil representation.

[0375] 11. A method according to any of clauses 7 to 10, wherein the measurement data comprises or is derived from measurements of optical characteristic values ​​of at least two pairs of pixels in the detected pupil representation.

[0376] 12. A method according to clause 11, wherein the optical property value comprises intensity or phase.

[0377] 13. A method according to clause 11 or 12, wherein the pixels in each pair of pixels are point symmetric with respect to a common point of symmetry in the pupil plane, or are mirror symmetric with respect to a common line of mirror symmetry in the pupil plane.

[0378] 14. A method according to any one of clauses 6 to 13, wherein:

[0379] The first contribution comprises a contribution to an asymmetric component of the detected representation, and the second contribution comprises a contribution to an asymmetric component of the detected representation; and

[0380] Obtaining information about errors in the measurement process and extracting the value of the parameter of interest for each measurement target includes:

[0381] The solution form is A system of linear equations, where

[0382] is a vector containing measurement data for each of the measurement targets;

[0383] is a vector comprising, for each of the metrology targets, unknown values ​​of the parameter of interest derived from the first contribution to the measurement data and the unknown second contribution to the measurement data; and

[0384] A is a matrix representing the expected sensitivity of the metrology process to the parameter of interest and the symmetric components of the detected representation.

[0385] 15. The method of any of the preceding clauses, wherein the plurality of metrology targets comprises all instances of a predetermined metrology target type on the substrate, all said instances of the predetermined metrology target type having the same nominal structure.

[0386] 16. A method according to any of clauses 1 to 14, wherein the plurality of metrology targets comprises a subset of all metrology targets present on the substrate.

[0387] 17. The method of clause 16, wherein the subset consists of a subset of all instances of a predetermined metrology target type on the substrate, all of the instances of the predetermined metrology target type having the same nominal structure.

[0388] 18. A method according to clause 16 or 17, wherein information about errors in the measurement process is repeatedly obtained using different subsets of measurement targets.

[0389] 19. The method of clause 18, wherein the different subsets of measurement targets are obtained by gradually adding measurement targets for which new measurement data are available to the subset and removing measurement targets for which only older measurement data are available from the subset.

[0390] 20. A method according to any of the preceding clauses, wherein the measurement process includes any one or both of the following: irradiating the measurement target with radiation in different bands and separately detecting the redirected radiation in each of the bands; and irradiating the measurement target with radiation with different polarizations and separately detecting the redirected radiation with different polarizations.

[0391] 21. A method according to any of the preceding clauses, wherein the parameter of interest comprises overlay error between different layers in the metrology target.

[0392] 22. The method of any preceding clause, wherein each measurement target of the plurality of measurement targets is spaced apart from each other measurement target of the plurality of measurement targets.

[0393] 23. A method for reducing errors in measurement data, comprising:

[0394] Measurement data is obtained from a metrology process applied to each of a plurality of metrology targets on a substrate, wherein:

[0395] The measurement data for each metrology target includes at least a first contribution from a parameter of interest of a patterning process used to form the metrology target and a second contribution from an error in the metrology process used to measure the metrology target; and

[0396] The method further comprises:

[0397] obtaining information about errors in the measurement process using the measurement data obtained from all of the plurality of measurement targets; and

[0398] The obtained information is used to reduce errors in the measurement data.

[0399] 24. A method for calibrating a measurement process, comprising:

[0400] Measurement data is obtained from a metrology process applied to each of a plurality of metrology targets on a substrate, wherein:

[0401] The measurement data for each metrology target includes at least a first contribution from a parameter of interest of a patterning process used to form the metrology target and a second contribution from an error in the metrology process used to measure the metrology target; and

[0402] The method further comprises:

[0403] obtaining information about errors in the measurement process using the measurement data obtained from all of the plurality of measurement targets; and

[0404] The obtained information is used to calibrate the metrology process to reduce errors in measurement data produced by the metrology process.

[0405] 25. A method of selecting a metrology target for use in a method of determining information about a patterning process, wherein:

[0406] The method of determining information about a patterning process includes: obtaining measurement data from a metrology process applied to each of a selected plurality of metrology targets; and using the obtained measurement data to extract values ​​of one or more parameters of interest of the patterning process used to form each metrology target; and

[0407] The measurement targets are selected such that the number of measurement targets is greater than the number of parameters of interest.

[0408] 26. A method according to clause 25, wherein the method of determining information about the patterning process comprises a method according to any one of clauses 1 to 22.

[0409] 27. A method according to clause 25 or 26, wherein the metrology targets are selected to include a subset of all metrology targets present on the substrate.

[0410] 28. The method of clause 27, wherein the subset consists of a subset of all instances of a predetermined metrology target type on the substrate, all of the instances of the predetermined metrology target type having the same nominal structure.

[0411] 29. A method according to clause 27 or 28, wherein information about the patterning process is repeatedly determined using different subsets of metrology targets as the selected plurality of metrology targets.

[0412] 30. The method of clause 29, wherein the different subsets of measurement targets are obtained by gradually adding measurement targets for which new measurement data are available to the subset and removing measurement targets for which only older measurement data are available from the subset.

[0413] 31. A computer program product comprising a computer non-transitory readable medium having instructions recorded thereon, the instructions, when executed by a computer, implementing the method according to any one of clauses 1 to 30.

[0414] 32. A system comprising:

[0415] Computer systems; and

[0416] A non-transitory computer-readable storage medium configured to store machine-readable instructions, wherein the machine-readable instructions, when executed, cause the computer system to perform the method according to any one of clauses 1 to 30.

[0417] 33. A metrology device for measuring an object of a patterning process, the metrology device being configured to perform a method according to any of clauses 1 to 30.

[0418] 34. A system comprising:

[0419] a metrology device configured to provide a radiation beam onto a surface of an object and detect radiation redirected by structures on the surface of the object; and

[0420] A computer program product according to clause 31.

[0421] 35. The system according to item 34 further includes a lithography device, which includes a support structure and a projection optical system, wherein the support structure is configured to hold a patterning device to modulate a radiation beam, and the projection optical system is arranged to project the modulated radiation beam onto a radiation-sensitive substrate, wherein the object is a substrate, and the lithography device is configured to control the settings of the lithography device based on information obtained using the measurement device and the computer program product.

[0422] In view of this description, modifications and alternative embodiments of various aspects of the present invention are obvious to those skilled in the art. Thus, the description and accompanying drawings should be interpreted as being illustrative only, and the purpose is to teach those skilled in the art the general way of implementing the present invention. It should be understood that the forms of the present invention shown and described herein will be considered as examples of embodiments. After benefiting from this description of the present invention, the following matters are obvious to those skilled in the art: for those elements and materials illustrated and described herein, elements and materials can be replaced, parts and materials can be reversed or omitted, certain features can be utilized independently, and the features of embodiments or embodiments can be combined. Without departing from the spirit and scope of the present invention as described in the appended claims, the elements described herein may be changed. The titles used herein are only for organizational purposes and are not intended to limit the scope of this specification.

[0423] As used throughout the application, the word "may" is used in a permissive sense (i.e., meaning potentially) rather than a mandatory sense (i.e., meaning must). The words "include, including, and includes" and the like mean including, but not limited to. As used throughout the application, unless the content clearly indicates otherwise, the singular forms "a", "an", and "the" include plural referents. Thus, for example, reference to "an" element or "an" element includes a combination of two or more elements, although other terms and phrases, such as "one or more", are used for one or more elements. Unless otherwise indicated, "or" is non-exclusive, i.e., covers both "and" and "or". Terms describing conditional relationships (e.g., "in response to X, Y," "at X, Y," "if X, Y," "when X, Y"), etc., encompass causal relationships where the antecedent is a necessary causal condition, the antecedent is a sufficient causal condition, or the antecedent is a contributing causal condition to the consequence, e.g., "state X occurs when condition Y is obtained" is common to "X occurs only when Y" and "X occurs when Y and Z." Such conditional relationships are not limited to consequences that are obtained immediately following the antecedent, as some consequences may be delayed, and in conditional statements, the antecedent is associated with its consequences, e.g., the antecedent is related to the likelihood of the consequence occurring. Unless otherwise specified, a statement in which multiple attributes or functions are mapped to multiple objects (e.g., one or more processors that perform steps A, B, C, and D) encompasses both that all of these attributes or functions are mapped to all of these objects, and that a subset of the attributes or functions are mapped to a subset of the attributes or functions (e.g., involving both that all of the processors each perform steps A to D, and that processor 1 performs step A, processor 2 performs step B and part of step C, and processor 3 performs part of step C and step D). Further, unless otherwise stated, a statement that a value or action is "based on" another condition or value encompasses both instances where the condition or value is the only factor and instances where the condition or value is one factor among multiple factors. Unless otherwise stated, a statement that "each" instance of a set has some property should not be read to exclude the case where some otherwise identical or similar members of the larger set do not have that property, i.e., each does not necessarily mean one by one.

[0424] To the extent that certain U.S. patents, U.S. patent applications, or other materials (e.g., articles) are incorporated by reference, the text of such U.S. patents, U.S. patent applications, and other materials is incorporated by reference only to the extent that there is no conflict between such materials and the language and figures set forth herein. In the event of such a conflict, any such conflicting text in such incorporated by reference U.S. patents, U.S. patent applications, and other materials is specifically not incorporated by reference herein.

[0425] The above description is intended to be illustrative rather than limiting. It will therefore be apparent to those skilled in the art that modifications may be made to the present disclosure as described without departing from the scope of the following claims as set forth.

Claims

1. A configuration scheme selection method, comprising: Obtaining measurement results from a metrology target positioned on a semiconductor wafer outside of a substrate area for devices, obtaining a measurement result from an in-device target, the in-device target being positioned on the semiconductor wafer, and Both the metrology target measurement results and the in-device metrology measurement results are used to determine acquisition profile aspects and / or substrate layout or sampling scheme aspects of a configuration scheme for accurate metrology.

2. The method of claim 1, wherein the metrology target is in close proximity to the in-device target.

3. The method of claim 1, wherein the metrology measurement is a symmetric portion of a measurement in a pupil plane.

4. The method of claim 1, wherein the metrology measurement is an antisymmetric portion of a measurement in a pupil plane.

5. The method of claim 1, wherein the metrology measurements are symmetric and antisymmetric parts of measurements in a pupil plane.

Images