[go: up one dir, main page]

WO2025229431A1 - Systèmes et procédés de prédiction d'impulsion subséquente dans une source de lumière - Google Patents

Systèmes et procédés de prédiction d'impulsion subséquente dans une source de lumière

Info

Publication number
WO2025229431A1
WO2025229431A1 PCT/IB2025/053611 IB2025053611W WO2025229431A1 WO 2025229431 A1 WO2025229431 A1 WO 2025229431A1 IB 2025053611 W IB2025053611 W IB 2025053611W WO 2025229431 A1 WO2025229431 A1 WO 2025229431A1
Authority
WO
WIPO (PCT)
Prior art keywords
time series
light source
light pulse
light
optical system
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
PCT/IB2025/053611
Other languages
English (en)
Inventor
Christopher James STEVENS
Nathan Gibson WELLS
Deepthi MYSORE NAGARAJ
Gian Paolo Custodio PASCO
Shashidhar MURTHY
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Cymer LLC
Original Assignee
Cymer LLC
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Cymer LLC filed Critical Cymer LLC
Publication of WO2025229431A1 publication Critical patent/WO2025229431A1/fr
Pending legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G03—PHOTOGRAPHY; CINEMATOGRAPHY; ANALOGOUS TECHNIQUES USING WAVES OTHER THAN OPTICAL WAVES; ELECTROGRAPHY; HOLOGRAPHY
    • G03F—PHOTOMECHANICAL PRODUCTION OF TEXTURED OR PATTERNED SURFACES, e.g. FOR PRINTING, FOR PROCESSING OF SEMICONDUCTOR DEVICES; MATERIALS THEREFOR; ORIGINALS THEREFOR; APPARATUS SPECIALLY ADAPTED THEREFOR
    • G03F7/00—Photomechanical, e.g. photolithographic, production of textured or patterned surfaces, e.g. printing surfaces; Materials therefor, e.g. comprising photoresists; Apparatus specially adapted therefor
    • G03F7/70—Microphotolithographic exposure; Apparatus therefor
    • G03F7/70008—Production of exposure light, i.e. light sources
    • G03F7/70041—Production of exposure light, i.e. light sources by pulsed sources, e.g. multiplexing, pulse duration, interval control or intensity control
    • G—PHYSICS
    • G03—PHOTOGRAPHY; CINEMATOGRAPHY; ANALOGOUS TECHNIQUES USING WAVES OTHER THAN OPTICAL WAVES; ELECTROGRAPHY; HOLOGRAPHY
    • G03F—PHOTOMECHANICAL PRODUCTION OF TEXTURED OR PATTERNED SURFACES, e.g. FOR PRINTING, FOR PROCESSING OF SEMICONDUCTOR DEVICES; MATERIALS THEREFOR; ORIGINALS THEREFOR; APPARATUS SPECIALLY ADAPTED THEREFOR
    • G03F7/00—Photomechanical, e.g. photolithographic, production of textured or patterned surfaces, e.g. printing surfaces; Materials therefor, e.g. comprising photoresists; Apparatus specially adapted therefor
    • G03F7/70—Microphotolithographic exposure; Apparatus therefor
    • G03F7/70483—Information management; Active and passive control; Testing; Wafer monitoring, e.g. pattern monitoring
    • G03F7/70491—Information management, e.g. software; Active and passive control, e.g. details of controlling exposure processes or exposure tool monitoring processes
    • G03F7/705—Modelling or simulating from physical phenomena up to complete wafer processes or whole workflow in wafer productions

Definitions

  • aspects of the present application relate to operations of laser systems, for example, models for predicting a next light pulse in a radiation source given a prior light pulse.
  • a lithographic apparatus is a machine that applies a desired pattern onto a substrate, usually onto a target portion of the substrate.
  • a lithographic apparatus can be used, for example, in the manufacture of integrated circuits (ICs).
  • a patterning device which can be a mask or a reticle, can be used to generate a circuit pattern to be formed on an individual layer of the IC.
  • This pattern can be transferred onto a target portion (e.g., comprising part of, one, or several dies) on a substrate (e.g., a silicon wafer). Transfer of the pattern is typically via imaging onto a layer of radiationsensitive material (photoresist or simply “resist”) provided on the substrate .
  • photoresist or simply “resist”
  • a single substrate will contain a network of adjacent target portions that are successively patterned.
  • lithographic apparatuses include so-called steppers, in which each target portion is irradiated by exposing an entire pattern onto the target portion at one time, and so-called scanners, in which each target portion is irradiated by scanning the pattern through a radiation beam in a given direction (the “scanning”- direction) while synchronously scanning the target portions parallel or anti-parallel to this scanning direction. It is also possible to transfer the pattern from the patterning device to the substrate by imprinting the pattern onto the substrate.
  • a lithographic apparatus typically includes an illumination system that conditions radiation generated by a radiation source before the radiation is incident upon a patterning device.
  • a patterned beam of deep ultraviolet (DUV) light can be used to produce extremely small features on a substrate.
  • DUV light generally refers to radiation having wavelengths ranging from 126 nm to 428 nm.
  • a pulsed-discharge light source such as a gas discharge laser, can be used to generate DUV light.
  • a patterned beam of EUV light can be used to produce extremely small features on a substrate.
  • EUV light (also sometimes referred to as soft x-rays) is generally defined as electromagnetic radiation having wavelengths in the range of about 5-100 nm.
  • One particular wavelength of interest for EUV photolithography is 13.5 nm.
  • a series of electrical pulses i.e., discharge events
  • the discharge events can ignite a plasma of gas in a discharge region between the electrodes.
  • the generated plasma can release radiation, thereby operating as a radiation source.
  • a model capable of simulating future behavior of a light source may be trained and utilized.
  • a simulation may utilize a machine learning model to predict a series of future light pulses of a light source under varying operating conditions.
  • An optical system may include a processor and a memory.
  • the memory may have instructions stored thereon that, when executed, cause the processor to obtain time series data of light pulse signals from a light source.
  • the processor may encode the time series data to obtain encoded data, which includes a time series of vectors. Each vector in the time series of vectors may include one or more values corresponding to features of a light pulse signal in the time series data of light pulse signals.
  • the processor may use a machine learning model to determine one or more values corresponding to features of a next light pulse signal and advance an optical system simulation based on the features of the next light pulse signal.
  • a method of training a machine learning model to predict a next pulse of a light source may include an obtaining step, an encoding step, and a training step.
  • the obtaining step may include obtaining time series data of a plurality of light source signals collected from one or more light sources.
  • the encoding step may include encoding the time series data to generate encoded data.
  • the encoded data may contain a series of vectors where each vector contains numerical values corresponding to setting and measurement parameters of the light source.
  • the training step may include training a machine learning model using the encoded data.
  • FIG. 1A shows a schematic of a reflective lithographic apparatus, according to some aspects.
  • FIG. IB shows a schematic of a transmissive lithographic apparatus, according to some aspects.
  • FIGS. 2A, 2B, and 3 show more details of a reflective lithographic apparatus, according to some aspects.
  • FIG. 4 shows more details of a transmissive lithographic apparatus, according to some aspects.
  • FIG. 5 shows a lithographic cell, according to some aspects.
  • FIG. 6 shows a transformer architecture, according to some aspects.
  • FIG. 7 shows a flowchart of a method of training a machine learning model to predict a next light pulse signal for a light source, according to some aspects.
  • FIG. 8 shows a sample time series of encoded vectors, according to some aspects.
  • FIG. 9 shows a flowchart of a method of predicting a next light pulse signal for a light source given a prior light pulse signal, according to some aspects.
  • FIG. 10 shows a computer system, according to some aspects.
  • spatially relative terms such as “beneath,” “below,” “lower,” “above,” “on,” “upper” and the like, can be used herein for ease of description to describe one element or feature’s relationship to another element(s) or feature(s) as illustrated in the figures.
  • the spatially relative terms are intended to encompass different orientations of the device in use or operation in addition to the orientation depicted in the figures.
  • the apparatus can be otherwise oriented (rotated 90 degrees or at other orientations) and the spatially relative descriptors used herein can likewise be interpreted accordingly.
  • the terms “about,” “approximately,” or the like can be used herein indicates the value of a given quantity that can vary based on a particular technology. Based on the particular technology, the terms “about,” “approximately,” or the like can indicate a value of a given quantity that varies within, for example, 10-30% of the value (e.g., ⁇ 10%, ⁇ 20%, or ⁇ 30% of the value).
  • a machine- readable medium can include any mechanism for storing or transmitting information in a form readable by a machine (e.g., a computing device).
  • a machine -readable medium can comprise read only memory (ROM); random access memory (RAM); magnetic disk storage media; optical storage media; flash memory devices; electrical, optical, acoustical or other forms of propagated signals (e.g., carrier waves, infrared signals, digital signals, etc.), and others.
  • firmware, software, routines, and/or instructions can be described herein as performing certain actions. However, it should be appreciated that such descriptions are merely for convenience and that such actions result from computing devices, processors, controllers, or other devices executing the firmware, software, routines, instructions, etc.
  • the term “machine -readable medium” can be interchangeable with similar terms, for example, “computer program product,” “computer-readable medium,” “non-transitory computer- readable medium,” or the like.
  • non-transitory can be used herein to characterize one or more forms of computer readable media except for a transitory, propagating signal.
  • Light sources are typically designed to operate within various tolerances. These tolerances, which are sometimes described as manufacturers’ specifications or customers’ requirements, may specify the quality of the light produced by a laser.
  • the quality of the light may be assessed according to a variety of metrics. For example, each individual pulse may be assessed according to the energy it carries. This energy can be characterized by metrics such as the pulse’s total energy, or its peak power, or other factors.
  • an input command to an energy control module in a laser may instruct the laser to produce a pulse with a total energy of 9 mJ.
  • the actual resulting pulse may have some variation from this commanded value, such as a total energy of 8.5 mJ or 9. 1 mJ; these outcomes can be understood as errors of -0.5 mJ or +0.1 mJ.
  • these errors may or may not be acceptable for the laser’s performance.
  • the quality of the light pulses may be assessed according to other factors such as its spectral properties or its timing properties.
  • Spectral properties can be characterized by metrics such as the pulse’s power spectral density distribution, or center wavelength, or peak wavelength, or bandwidth (e.g., full-width-half-max (FWHM) bandwidth, or 95% energy bandwidth (E95), or other bandwidth metrics), for example.
  • Timing properties can be characterized by metrics such as onset time, peak time, and time duration (e.g., start-to-end duration or “integral square pulse duration” (US)), for example.
  • an energy control module in the laser may adapt its operation to avoid the unacceptable performance.
  • the predictions of variations or errors may be made based on factors such as the current operating conditions of the laser, the age of the laser, the recent operating history of the laser, the long-term operating history of the laser. For example, a review of past performance may be used to determine that during periods of operation immediately following a pause of more than 30 seconds, a particular laser tends to provide a 5% error of reduced pulse energy.
  • an energy control module in the laser might accordingly compensate to avoid or mitigate the error.
  • the current operating conditions and other factors may be used to predict the variation or error in other characteristics of an upcoming laser pulse, such as one or more of its spectral properties or one or more of its timing properties.
  • the techniques presented herein provide tools for predicting variations or errors in characteristics of an upcoming light pulse from a pulsed light source.
  • the characteristics can include properties of the light pulse such as its wavelength, bandwidth, energy, and timing, among others.
  • performance predictions may be made according to aspects herein to estimate or otherwise calculate a likelihood or an estimated timeframe of an upcoming failure (or other reduced performance) of a light source or of a component of a light source. These estimates may be used as part of a predictive maintenance protocol to schedule or prioritize a maintenance event for the light source (e.g., replacement of an optical component or execution of a gas refdl). Such predictive maintenance be used for proactively scheduling a downtime of equipment for maintenance, rather than responding to an unexpected failure by enduring an unscheduled downtime for maintenance or repair.
  • a sensitivity analysis may be performed according to aspects herein to determine how a light source will behave under different operating conditions.
  • a sensitivity analysis may also ensure that a light source stays within predefined specifications during operation.
  • sensitivity analysis requires costly and time consuming chamber testing. Chamber testing may be reduced if a simulation is able to predict future behavior of a light source.
  • FIGS. 1A and IB show schematic illustrations of a lithographic apparatus 100 and lithographic apparatus 100’, respectively, in which embodiments of the present disclosure may be implemented.
  • Lithographic apparatus 100 and lithographic apparatus 100’ each include the following: an illumination system (illuminator) IL configured to condition a radiation beam B (for example, deep ultra violet or extreme ultra violet radiation); a support structure (for example, a mask table) MT configured to support a patterning device (for example, a mask, a reticle, or a dynamic patterning device) MA and connected to a first positioner PM configured to accurately position the patterning device MA; and, a substrate table (for example, a wafer table) WT configured to hold a substrate (for example, a resist coated wafer) W and connected to a second positioner PW configured to accurately position the substrate W.
  • an illumination system illumination system
  • IL for example, deep ultra violet or extreme ultra violet radiation
  • a support structure for example, a mask table
  • MT configured to support
  • Lithographic apparatus 100 and 100’ also have a projection system PS configured to project a pattern imparted to the radiation beam B by patterning device MA onto a target portion (for example, comprising one or more dies) C of the substrate W.
  • the patterning device MA and the projection system PS are reflective.
  • the patterning device MA and the projection system PS are transmissive.
  • the illumination system IL may include various types of optical components, such as refractive, reflective, catadioptric, magnetic, electromagnetic, electrostatic, or other types of optical components, or any combination thereof, for directing, shaping, or controlling the radiation beam B.
  • optical components such as refractive, reflective, catadioptric, magnetic, electromagnetic, electrostatic, or other types of optical components, or any combination thereof, for directing, shaping, or controlling the radiation beam B.
  • the support structure MT holds the patterning device MA in a manner that depends on the orientation of the patterning device MA with respect to a reference frame, the design of at least one of the lithographic apparatus 100 and 100’, and other conditions, such as whether or not the patterning device MA is held in a vacuum environment.
  • the support structure MT may use mechanical, vacuum, electrostatic, or other clamping techniques to hold the patterning device MA.
  • the support structure MT may be a frame or a table, for example, which may be fixed or movable, as required. By using sensors, the support structure MT may ensure that the patterning device MA is at a desired position, for example, with respect to the projection system PS.
  • patterning device should be broadly interpreted as referring to any device that may be used to impart a radiation beam B with a pattern in its cross-section, such as to create a pattern in the target portion C of the substrate W.
  • the pattern imparted to the radiation beam B may correspond to a particular functional layer in a device being created in the target portion C to form an integrated circuit.
  • the terms “inspection apparatus,” “metrology system,” or the like may be used herein to refer to, e.g., a device or system used for measuring a property of a structure (e.g., overlay error, critical dimension parameters) or used in a lithographic apparatus to inspect an alignment of a wafer (e.g., alignment apparatus).
  • a property of a structure e.g., overlay error, critical dimension parameters
  • a lithographic apparatus e.g., alignment apparatus
  • the patterning device MA may be transmissive (as in lithographic apparatus 100’ of FIG. IB) or reflective (as in lithographic apparatus 100 of FIG. 1A).
  • Examples of patterning devices MA include reticles, masks, programmable mirror arrays, or programmable LCD panels.
  • Masks are well known in lithography, and include mask types such as binary, alternating phase shift, or 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 may be individually tilted so as to reflect an incoming radiation beam in different directions. The tilted mirrors impart a pattern in the radiation beam B, which is reflected by a matrix of small mirrors.
  • projection system PS may encompass any type of projection system, including refractive, reflective, catadioptric, magnetic, electromagnetic and electrostatic optical systems, or any combination thereof, as appropriate for the exposure radiation being used, or for other factors, such as the use of an immersion liquid on the substrate W or the use of a vacuum.
  • a vacuum environment may be used for EUV or electron beam radiation since other gases may absorb too much radiation or electrons.
  • a vacuum environment may therefore be provided to the whole beam path with the aid of a vacuum wall and vacuum pumps.
  • Lithographic apparatus 100 and/or lithographic apparatus 100’ may be of a type having two (dual stage) or more substrate tables WT (and/or two or more mask tables).
  • the additional substrate tables WT may be used in parallel, or preparatory steps may be carried out on one or more tables while one or more other substrate tables WT are being used for exposure.
  • the additional table may not be a substrate table WT.
  • the lithographic apparatus may also be of a type wherein at least a portion of the substrate may be covered by a liquid having a relatively high refractive index, e.g., water, so as to fdl a space between the projection system and the substrate.
  • a liquid having a relatively high refractive index e.g., water
  • An immersion liquid may also be applied to other spaces in the lithographic apparatus, for example, between the mask and the projection system. Immersion techniques are well known in the art for increasing the numerical aperture of projection systems.
  • immersion as used herein does not mean that a structure, such as a substrate, must be submerged in liquid, but rather only means that liquid is located between the projection system and the substrate during exposure.
  • the illuminator IL receives a radiation beam from a radiation source SO.
  • the source SO and the lithographic apparatus 100, 100’ may be separate physical entities, for example, in arrangements where the source SO is an excimer laser. In such cases, the source SO is not considered to form part of the lithographic apparatus 100 or 100’, and the radiation beam B passes from the source SO to the illuminator IL with the aid of a beam delivery system BD (in FIG. IB) including, for example, suitable directing mirrors and/or a beam expander.
  • the source SO may be an integral part of the lithographic apparatus 100, 100’, for example, in arrangements where the source SO is a mercury lamp.
  • the source SO and the illuminator IL, together with the beam delivery system BD, if required, may be referred to as a radiation system.
  • the illuminator IL may include an adjuster AD (in FIG. IB) for adjusting the angular intensity distribution of the radiation beam.
  • AD adjuster
  • the illuminator IL may comprise various other components (in FIG. IB), such as an integrator IN and a condenser CO.
  • the illuminator IL may be used to condition the radiation beam B to have a desired uniformity and intensity distribution in its cross section.
  • the radiation beam B is incident on the patterning device (for example, mask) MA, which is held on the support structure (for example, mask table) MT, and is patterned by the patterning device MA.
  • the radiation beam B is reflected from the patterning device (for example, mask) MA.
  • the radiation beam B passes through the projection system PS, which focuses the radiation beam B onto a target portion C of the substrate W.
  • the substrate table WT may be moved accurately (for example, so as to position different target portions C in the path of the radiation beam B).
  • the first positioner PM and another position sensor IF1 may be used to accurately position the patterning device (for example, mask) MA with respect to the path of the radiation beam B.
  • Patterning device (for example, mask) MA and substrate W may be aligned using mask alignment marks Ml, M2 and substrate alignment marks Pl, P2.
  • the radiation beam B is incident on the patterning device (for example, mask MA), which is held on the support structure (for example, mask table MT), and is patterned by the patterning device. Having traversed the mask MA, the radiation beam B passes through the projection system PS, which focuses the beam onto a target portion C of the substrate W.
  • the projection system has a pupil conjugate PPU to an illumination system pupil IPU. Portions of radiation emanate from the intensity distribution at the illumination system pupil IPU and traverse a mask pattern without being affected by diffraction at the mask pattern and create an image of the intensity distribution at the illumination system pupil IPU.
  • the projection system PS projects an image of the mask pattern MP, where the image is formed by diffracted beams produced from the mark pattern MP by radiation from the intensity distribution, onto a photoresist layer coated on the substrate W.
  • the mask pattern MP may include an array of lines and spaces. A diffraction of radiation at the array and different from zeroth order diffraction generates diverted diffracted beams with a change of direction in a direction perpendicular to the lines. Undiffracted beams (i.e., so-called zeroth order diffracted beams) traverse the pattern without any change in propagation direction.
  • the zeroth order diffracted beams traverse an upper lens or upper lens group of the projection system PS, upstream of the pupil conjugate PPU of the projection system PS, to reach the pupil conjugate PPU.
  • the portion of the intensity distribution in the plane of the pupil conjugate PPU and associated with the zeroth order diffracted beams is an image of the intensity distribution in the illumination system pupil IPU of the illumination system IU.
  • the aperture device PD for example, is disposed at or substantially at a plane that includes the pupil conjugate PPU of the projection system PS.
  • the projection system PS is arranged to capture, by means of a lens or lens group U, not only the zeroth order diffracted beams, but also first-order or first- and higher-order diffracted beams (not shown).
  • dipole illumination for imaging line patterns extending in a direction perpendicular to a line may be used to utilize the resolution enhancement effect of dipole illumination.
  • first-order diffracted beams interfere with corresponding zeroth -order diffracted beams at the level of the wafer W to create an image of the line pattern MP at highest possible resolution and process window (i.e., usable depth of focus in combination with tolerable exposure dose deviations).
  • astigmatism aberration may be reduced by providing radiation poles (not shown) in opposite quadrants of the illumination system pupil IPU. Further, in some embodiments, astigmatism aberration may be reduced by blocking the zeroth order beams in the pupil conjugate PPU of the projection system associated with radiation poles in opposite quadrants. This is described in more detail in US 7,511,799 B2, issued Mar. 31, 2009, which is incorporated by reference herein in its entirety.
  • the substrate table WT may be moved accurately (for example, so as to position different target portions C in the path of the radiation beam B).
  • the first positioner PM and another position sensor may be used to accurately position the mask MA with respect to the path of the radiation beam B (for example, after mechanical retrieval from a mask library or during a scan).
  • movement of the mask table MT may be realized 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.
  • movement of the substrate table WT may be realized using a long-stroke module and a short-stroke module, which form part of the second positioner PW.
  • the mask table MT may be connected to a short-stroke actuator only or may be fixed.
  • Mask MA and substrate W may be aligned using mask alignment marks Ml, M2, and substrate alignment marks Pl , P2.
  • the substrate alignment marks (as illustrated) occupy dedicated target portions, they may be located in spaces between target portions (known as scribe-lane alignment marks). Similarly, in situations in which more than one die is provided on the mask MA, the mask alignment marks may be located between the dies.
  • Mask table MT and patterning device MA may be in a vacuum chamber V, where an in-vacuum robot IVR may be used to move patterning devices such as a mask in and out of vacuum chamber.
  • an out-of-vacuum robot may be used for various transportation operations, similar to the in-vacuum robot IVR. Both the in-vacuum and out-of-vacuum robots need to be calibrated for a smooth transfer of any payload (e.g., mask) to a fixed kinematic mount of a transfer station.
  • the lithographic apparatus 100 and 100’ may be used in at least one of the following modes: [0053] 1.
  • step mode the support structure (for example, mask table) MT and the substrate table WT are kept essentially stationary, while an entire pattern imparted to the radiation beam B is projected onto a target portion C at one time (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 may be exposed.
  • the support structure (for example, mask table) MT and the substrate table WT are scanned synchronously while a pattern imparted to the radiation beam B is projected onto a target portion C (i.e., a single dynamic exposure).
  • the velocity and direction of the substrate table WT relative to the support structure (for example, mask table) MT may be determined by the (de- )magnification and image reversal characteristics of the projection system PS.
  • the support structure (for example, mask table) MT is kept substantially stationary holding a programmable patterning device, and the substrate table WT is moved or scanned while a pattern imparted to the radiation beam B is projected onto a target portion C.
  • a pulsed radiation source SO may be employed and the programmable patterning device is updated as required after each movement of the substrate table WT or in between successive radiation pulses during a scan.
  • This mode of operation may be readily applied to maskless lithography that utilizes a programmable patterning device, such as a programmable mirror array.
  • lithographic apparatus 100 includes an extreme ultraviolet (EUV) source, which is configured to generate a beam of EUV radiation for EUV lithography.
  • EUV extreme ultraviolet
  • the EUV source is configured in a radiation system, and a corresponding illumination system is configured to condition the EUV radiation beam of the EUV source.
  • FIG. 2A shows different view of lithographic apparatus 100, including source SO (e.g., source collector apparatus), illumination system IL, and projection system PS, according to some aspects.
  • Source SO e.g., source collector apparatus
  • illumination system IL illumination system
  • projection system PS projection system PS
  • Source SO is constructed and arranged such that a vacuum environment can be maintained in an enclosing structure 220 of source SO.
  • An EUV radiation emitting plasma 210 can be formed by a discharge-generated plasma source.
  • a plasma of excited tin (Sn) e.g., excited via a laser is used to produce EUV radiation.
  • the radiation emitted by the EUV radiation emitting plasma 210 can be passed from a source chamber 211 into a collector chamber 212 via an optional gas barrier or contaminant trap 230 (in some cases also referred to as contaminant barrier or foil trap), which is positioned in or behind an opening in source chamber 211.
  • Contaminant trap 230 can comprise a channel structure. Contamination trap 230 can also comprise a gas barrier and/or a channel structure.
  • collector chamber 212 can comprise a radiation collector CO.
  • Radiation collector CO can be a so-called grazing incidence collector.
  • Radiation collector CO can comprise an upstream radiation collector side 251 and a downstream radiation collector side 252.
  • Radiation that traverses radiation collector CO can be reflected off a grating spectral fdter 240 to be focused in a virtual source point INTF.
  • Virtual source point INTF can be referred to as the intermediate focus.
  • Source collector apparatus can be arranged such that the intermediate focus INTF is located at or near an opening 219 of enclosing structure 220.
  • the virtual source point INTF can be an image of the EUV radiation emitting plasma 210.
  • Grating spectral fdter 240 can be used for suppressing infrared (IR) radiation.
  • Illumination system IL can include a faceted field mirror device 222 and a faceted pupil mirror device 224 arranged to provide a desired angular distribution of radiation beam 221, at patterning device MA, as well as a desired uniformity of radiation intensity at patterning device MA.
  • a patterned beam 226 is formed and the patterned beam 226 is imaged by projection system PS via reflective elements 228, 229 onto substrate W held by the wafer stage or substrate table WT.
  • other configurations of mirrors and/or optical devices can be used to direct radiation beam 221 to patterning device MA.
  • Grating spectral filter 240 can optionally be present, depending upon the type of lithographic apparatus. Further, there can be more mirrors present than those shown in the FIG. 2A, for example there can be one to six additional reflective elements present in the projection system PS than shown in FIG. 2A.
  • uniformity compensator UC, sensor ES, and/or measurement sensor MS shown in FIGS. 2A and 2B can be as described above in reference to FIG. 1A.
  • Collector CO is depicted as an example of a nested collector with grazing incidence reflectors 253, 254, and 255 (or collector mirror).
  • Grazing incidence reflectors 253, 254, and 255 can be disposed axially symmetric around an optical axis O.
  • a collector optic of this type can be used in combination with a discharge -generated plasma source, often called a DPP source.
  • FIG. 2B shows a portion of lithographic apparatus 100 (e.g., FIG. 1A), but with alternative collection optics in source SO, according to some aspects. It should be appreciated that structures shown in FIG. 2A that do not appear in FIG. 2B (for drawing clarity) can still be included in aspects referring to FIG. 2B. Elements in FIG. 2B having the same reference numbers as those in FIG. 2A have the same or substantially similar structures and functions as described in reference to FIG. 2A.
  • the lithographic apparatus 100 can be used, for example, to expose a substrate W such as a resist-coated wafer with a patterned beam of EUV illumination.
  • a substrate W such as a resist-coated wafer with a patterned beam of EUV illumination.
  • illumination system IL and projection system PS are represented combined as an exposure device 256 (e.g., an integrated circuit lithography tool such as a stepper, scanner, step and scan system, direct write system, device using a contact and/or proximity mask, etc.) that uses EUV light from source SO.
  • Lithographic apparatus 100 can also comprise collector 258 that reflects EUV light from the EUV radiation emitting plasma 210 along a path into the exposure device 256 to irradiate substrate W.
  • Collector 258 can comprise a near-normal incidence collector mirror having a reflective surface in the form of a prolate spheroid (e.g., an ellipse rotated about its major axis).
  • the prolate spheroid structure can have a graded multi-layer coating with alternating layers of Molybdenum and Silicon, and in some cases, one or more high temperature diffusion barrier layers, smoothing layers, capping layers and/or etch stop layers.
  • FIG. 3 shows a detailed view of a portion of lithographic apparatus 100 (e.g., FIGS. 1A, 2A, and 2B), according to one or more aspects. Elements in FIG. 3 having the same reference numbers as those in FIGS. 1, 2A, and 2B have the same or substantially similar structures and functions as described in reference to FIGS. 1, 2A, and 2B.
  • source SO can be a LPP EUV source.
  • Source SO can comprise a laser system 302 for generating a train of light pulses and delivering the light pulses into a light source chamber 212.
  • the light pulses can travel along one or more beam paths from the laser system 302 and into the chamber 212 to illuminate a source material at an irradiation region 304 to generate a plasma (e.g., plasma region located at EUV radiation emitting plasma 210 in FIG. 2B) that produces EUV light for substrate exposure in the exposure device 256.
  • laser system 302 can comprise a pulsed laser device, e.g., a pulsed gas discharge CO2 laser device producing radiation at 9.3 pm or 10.6 pm, e.g., with DC or RF excitation, operating at relatively high power, e.g., 10 kW or higher and high pulse repetition rate, e.g., 50 kHz or more.
  • the laser can be an axial-flow RF-pumped CO2 laser having an oscillator amplifier configuration (e.g., master oscillator/power amplifier (MOPA) or power oscillator/power amplifier (POPA)) with multiple stages of amplification and having a seed pulse that is initiated by a Q-switched oscillator with relatively low energy and high repetition rate, e.g., capable of 100 kHz operation. From the oscillator, the laser pulse can then be amplified, shaped and/or focused before reaching the irradiation region 304. Continuously pumped CO2 amplifiers can be used for the laser system 302. Alternatively, the laser can be configured as a so-called “self-targeting” laser system in which the droplet serves as one mirror of the optical cavity of the laser.
  • an oscillator amplifier configuration e.g., master oscillator/power amplifier (MOPA) or power oscillator/power amplifier (POPA)
  • MOPA master oscillator/power amplifier
  • POPA power oscillator/power amplifier
  • lasers can also be suitable, e.g., an excimer or molecular fluorine laser operating at high power and high pulse repetition rate.
  • a solid state laser e.g., having a fiber, rod, slab, or disk-shaped active media
  • other laser architectures having one or more chambers, e.g., an oscillator chamber and one or more amplifying chambers (with the amplifying chambers in parallel or in series)
  • a master oscillator/power oscillator (MOPO) arrangement e.g., a master oscillator/power ring amplifier (MOPRA) arrangement
  • MOPRA master oscillator/power ring amplifier
  • solid state laser that seeds one or more excimer, molecular fluorine or CO2 amplifier or oscillator chambers, can be suitable.
  • Other suitable designs are envisaged.
  • a source material can first be irradiated by a pre-pulse and thereafter irradiated by a main pulse.
  • Pre-pulse and main pulse seeds can be generated by a single oscillator or two separate oscillators.
  • One or more common amplifiers can be used to amplify both the pre -pulse seed and main pulse seed.
  • separate amplifiers can be used to amplify the pre-pulse and main pulse seeds.
  • source SO can also comprise a beam conditioning unit 306 having one or more optics for beam conditioning, such as expanding, steering, and/or focusing the beam between the laser system 302 and irradiation region 304.
  • a steering system which can comprise one or more mirrors, prisms, lenses, etc., can be provided and arranged to steer the laser focal spot to different locations in the chamber 212.
  • the steering system can comprise a first flat mirror mounted on a tip-tilt actuator, which can move the first mirror independently in two dimensions, and a second flat mirror mounted on a tip-tilt actuator which can move the second mirror independently in two dimensions.
  • the steering system can controllably move the focal spot in directions substantially orthogonal to the direction of beam propagation (beam axis or optical axis).
  • Beam conditioning unit 306 can comprise a focusing assembly to focus the beam to irradiation region 304 and adjust the position of the focal spot along the beam axis.
  • a focusing assembly an optic, such as a focusing lens or mirror, can be used that is coupled to an actuator for movement in a direction along the beam axis to move the focal spot along the beam axis.
  • the source SO can also comprise a source material delivery system 308 for delivering source material, such as tin droplets, to irradiation region 304, where the droplets can interact with light pulses from the laser system 302 to produce plasma and generate an EUV emission.
  • the EUV emission is used to expose a substrate such as a resist-coated wafer at exposure device 256. More details regarding various droplet dispenser configurations can be found in, e.g., U.S. Pat. No. 7,872,245, issued on January 18, 2011, titled “Systems and Methods for Target Material Delivery in a Laser Produced Plasma EUV Light Source”, U.S. Pat. No.
  • the source material for producing an EUV light output for substrate exposure can include, but is not necessarily limited to, a material that includes tin, lithium, xenon or combinations thereof.
  • the source material can be in the form of liquid droplets and/or solid particles contained within liquid droplets.
  • the element tin can be used as pure tin, as a tin compound, e.g., SnBn. SnBr2, SnFL, as a tin alloy, e.g., tin-gallium alloys, tin-indium alloys, tin-indium-gallium alloys, or a combination thereof.
  • the source material when sent to irradiation region 304, can be at various temperatures, for example, room temperature or near room temperature (e.g., tin alloys, SnB ), at an elevated temperature (e.g., pure tin), or at temperatures below room temperature (e.g., SnFL).
  • room temperature or near room temperature e.g., tin alloys, SnB
  • elevated temperature e.g., pure tin
  • SnFL room temperature below room temperature
  • the source SO can also comprise a controller 310 and/or a drive laser control system 312 for controlling devices in laser system 302 to generate light pulses for delivery into the chamber 212 and/or for controlling movement of optics in beam conditioning unit 306.
  • Source SO can also comprise a droplet position detection system which can comprise one or more droplet imagers 314 that provide an output signal indicative of the position of one or more droplets (e.g., to ensure that droplets arrive on target at irradiation region 304).
  • the droplet imager(s) 314 can provide measurement output to a droplet position detection feedback system 316.
  • Droplet position detection feedback system 316 can compute a droplet position and trajectory, from which a droplet error can be computed (e.g., on a droplet-by-droplet basis, or on average).
  • the droplet error can then be provided as an input to controller 310, which can, for example, provide a position, direction and/or timing correction signal to laser system 302 to control laser trigger timing and/or to control movement of optics in beam conditioning unit 306, e.g., to change the location and/or focal power of the light pulses being delivered to irradiation region 304 in chamber 212.
  • source material delivery system 308 can comprise a control system operable in response to a signal from controller 310 (which in some implementations can include the droplet error described above, or some quantity derived therefrom) to modify the release point, initial droplet stream direction, droplet release timing and/or droplet modulation to correct for errors in the droplets arriving at irradiation region 304.
  • controller 310 which in some implementations can include the droplet error described above, or some quantity derived therefrom
  • the lithographic apparatus 100 can also comprise a collector 258 and a gas dispenser device 320.
  • Gas dispenser device 320 can dispense gas in the path of the source material from source material delivery system 308 (e.g., irradiation region 304).
  • Gas dispenser device 320 can comprise a nozzle through which dispensed gas can exit.
  • Gas dispenser device 320 can be structured (e.g., having an aperture) such that, when placed near the optical path of laser system 302, light from laser system 302 is not blocked by gas dispenser device 320 and is allowed to reach irradiation region 304.
  • a buffer gas such as hydrogen, helium, argon or combinations thereof, can be introduced into chamber 212.
  • the buffer gas can be present in the chamber 212 during plasma discharge and can act to slow plasma-created ions, reduce degradation of optics, and/or increase plasma efficiency.
  • a magnetic field and/or electric field (not shown) can be used alone, or in combination with a buffer gas, to reduce damage caused by fast-moving ions.
  • collector 258 can be a near-normal incidence collector mirror having a reflective surface in the form of a prolate spheroid as described above.
  • Collector 258 can be formed with an aperture to allow the light pulses generated by laser system 302 to pass through and reach irradiation region 304. The same, or another aperture, can be used to allow gas from the gas dispenser device 320 to flow into chamber 212.
  • the collector 258 can be, e.g., a prolate spheroid mirror that has a first focus within or near the irradiation region 304 and a second focus at an intermediate region 318, where the EUV light can be transmitted to exposure device 256.
  • collectors other than collector 258 (e.g., collector CO (FIG. 2A)).
  • FIG. 4 shows a radiation source 400, according to some aspects.
  • radiation source 400 is a pulsed-discharge radiation source.
  • a gas discharge laser such as an excimer laser is an example of a pulsed-discharge radiation source.
  • Source SO of lithographic apparatus 100’ can implement radiation source 400.
  • Radiation source 400 can comprise a gas chamber 402, a window 404, conduit system 406, and one or more electrodes 410 (also “electrical connection”).
  • Conduit system 406 can comprise a network of valves, conduits, and contaminant filters (not shown).
  • gas chamber 402 can confine a gas 408.
  • Gas 408 includes at least one of nitrogen, halogens, and noble gases, such as fluorine, neon, krypton, argon, xenon and the like. Gas 408 can be rarified or pressurized via a pressure control system that controls a pressure within gas chamber 402.
  • Conduit system 406 is connected to gas chamber 402.
  • Conduit system 406 can allow management of gas 408 in gas chamber 402.
  • conduit system 406 can direct a flow (e.g., circulation) of gas 408 to a filter within conduit system 406 to purify gas 408.
  • a voltage/current can be supplied to gas 408 (e.g., via one or more electrodes 410) to generate radiation 412.
  • the voltage/current can be in the form of pulse with sufficient power to strike a plasma of gas 408.
  • the plasma can generate radiation with a set of wavelengths that depend on energy states of the plasma.
  • the type of gas 408 e.g., a mixture of krypton and fluorine or a mixture of argon and fluorine
  • Window 404 can allow radiation 412 to exit gas chamber 402.
  • FIG. 5 shows a lithographic cell 500, also sometimes referred to a lithocell or cluster, according to some aspects.
  • Lithographic apparatus 100 or 100’ can form part of lithographic cell 400.
  • Lithographic cell 400 can also comprise one or more apparatuses to perform pre -exposure and post-exposure processes on a substrate. These can include spin coaters SC to deposit resist layers, developers DE to develop exposed resist, chill plates CH, and bake plates BK.
  • a substrate handler, or robot, RO picks up substrates from input/output ports I/O I . I/O2, moves them between the different process apparatuses and delivers them to the loading bay LB of the lithographic apparatus 100.
  • Light sources are typically designed to operate within various tolerances. These tolerances, which are sometimes described as manufacturers’ specifications or customers’ requirements, may specify the quality of the light produced by a laser.
  • the quality of the light may be assessed according to a variety of metrics. For example, each individual pulse may be assessed according to the energy it carries. This energy can be characterized by metrics such as the pulse’s total energy, its peak power, or other factors.
  • the quality of the light pulses may be assessed according to other factors such as a light pulses spectral properties or its timing properties.
  • Spectral properties can be characterized by metrics such as the pulse’s power spectral density distribution, or center wavelength, or peak wavelength, or bandwidth (e.g., full-width-half-max (FWHM) bandwidth, or 95% energy bandwidth (E95), or other bandwidth metrics), for example.
  • Timing properties can be characterized by metrics such as onset time, peak time, and time duration (e.g., start-to-end duration or “integral square pulse duration” (7is)). for example.
  • the ability to predict future light pulses can improve testing and operations of laser systems. For example, if a prediction can be made that the next light pulse is likely to have an energy that is unacceptable (e.g., outside of specification or outside of tolerances) then an energy control module in the laser may adapt its operation to avoid the unacceptable performance.
  • an energy control module in the laser may adapt its operation to avoid the unacceptable performance.
  • sensitivity analysis the behavior of a light source is tested under different operating conditions. For example, sensitivity analysis may determine whether or not a light source is stable when parameters, such as pressure, are changed. A sensitivity analysis may also ensure that a light source stays within predetermined specifications during operation. Currently, sensitivity analysis requires costly and time consuming chamber testing. Chamber testing may be reduced if a simulation is able to predict future behavior of a light source.
  • a machine learning model may be able to “learn” enough laser physics or enough about the historical performance of a particular version of a light source to accurately predict the behavior of a light source.
  • a machine learning model may be trained to predict future light pulses in a light source using time series data collected from one or more light sources. Examples of machine learning modes include, but are not limited to, recurrent neural networks, long-short term memory models, and transformer architectures.
  • a transformer architecture is a type of attention-based neural network that learns context by tracking relationships in sequential data.
  • the transformer architecture used by some aspects described herein may be a generative pre-trained transformer (“GPT”), such as GPT-4 available from OpenAI, Inc. More information on the transformer architecture may be found in Vashwani et al. “Attention is all you need.” arXiv: 1706.03762v7, 2023.
  • GPT generative pre-trained transformer
  • FIG. 6 shows an example of a transformer architecture 600.
  • Transformer architecture 600 may comprise an encoder 602 and a decoder 604.
  • Encoder 602 may comprise a stack of identical layers 606.
  • encoder 602 comprises 6 or more layers.
  • Each layer 606 may comprise sub-layers 608 and 610.
  • sub-layer 608 may comprise a multi-head self-attention mechanism.
  • an input vector 605 may be multiplied by weighted matrices to generate query, key, and value vectors.
  • the query and key vectors may have dimension dk, while the value vector may have output d v .
  • Query, key, and value vectors calculated for each input vector 605 may be combined into query, key, and value matrices.
  • the query and key matrices can be multiplied together, scaled (e.g., divided by Vd fe ). and normalized into a vectorized probability distribution using, for example, a softmax function.
  • the output of the softmax function may be multiplied by the value matrix to obtain an attention vector.
  • sub-layer 610 may comprise a feed-forward network.
  • a feed-forward network may transform the output of the attention mechanism into a form that is usable by the next encoder or decoder layer.
  • the feed-forward network may comprise one or more hidden layers.
  • the feed-forward network may have two linear transformations with a rectified linear unit (“ReLU”) activation function in between.
  • ReLU rectified linear unit
  • the keys, values, and queries of an encoder layer come from the output of the previous decoder layer. Each position in the encoder can attend to all positions in the previous layer of the encoder.
  • Decoder 604 may comprise a stack of identical layers 612.
  • decoder 604 comprises 6 or more layers.
  • Each layer 612 may comprise sub-layers 614, 616, and 618.
  • Sub-layer 616 may comprise a multi -head attention layer that performs attention over the output 620 of encoder 602.
  • Sub-layer 614 may comprise a multi-head self-attention mechanism that operates similarly to sub-layer 608.
  • Sub-layer 618 may comprise a feed-forward network.
  • sub-layer 616 of decoder 604 may receive hidden states and attention values from the encoder. Hidden states with high scores may be amplified and hidden states with low scores may be downsized.
  • FIG. 7 shows a flowchart of a method 700 of training a machine learning model to predict the next light pulse of a light source after a given light pulse, according to some aspects.
  • Method 700 may be used to train a transformer architecture, as described with respect to FIG. 6.
  • Method 700 may comprise steps 702, 704, and 706.
  • Step 702 comprises obtaining time series data.
  • Time series data may include a plurality of light source signals collected from one or more light sources.
  • the plurality of light source signals are collected from a database containing recorded observations from a fleet of light sources.
  • a light source signal may contain values corresponding to light source settings and light pulse measurements.
  • light source settings may include pressure and temperature of a master oscillator and/or power ring amplifier of a source laser, timing and operation modes of a source laser, voltage set points for the source laser, modes and statuses of light source components, or the like.
  • Light pulse measurements may include, as non-limiting examples, light pulse energy, bandwidth, wavelength, energy duty cycle, pulse shot number, energy stability metrics, or the like. Light pulse measurements may also include averages, standard deviations, integral values, and errors related to light pulse energy, bandwidth and wavelength. Measurements may be captured by one or more sensors in a lithography system and may include estimated/calculated values.
  • a light source signal may also include metadata such as a laser serial number, file data, file name, file size, etc. In some aspects, a light source signal may contain up to 80 setting and measurement values. Table 1 illustrates an example light source signal.
  • Time series data may be collected as raw shot data files during operation of a light source.
  • one or more sensors may collect 500 data points or more per second, 1000 data points or more per second, 5000 data points or more per second, 6000 data points or more per second, or 6750 data points or more per second.
  • time series data may be continuously streamed by a light source and stored in an external database. Time series data may also be stored locally on a processor in the light source and written to an output file. Output files may be transferred to an external database, such as a data lake and/or a cloud storage medium.
  • certain triggers may cause the processor to write the time series data to an output file. For example, if an issue occurs in the light source, the processor may output the 2,500 light pulse signals preceding the issue. In some aspects, a user may prompt a light source to output the last 10,000 light pulse signals automatically stored on a local processor in the light source.
  • time series data sets may be stored in a centralized location.
  • time series data may be stored in a centralized data lake and/or a cloud storage medium.
  • time series data may be filtered and/or grouped according to parameters of interest. For example, multiple time series datasets corresponding to high power operations of a laser may be grouped together into a larger data set. Portions of datasets that contain parameters not of interest, such as software events and null values, may be filtered from the time series data.
  • Step 704 comprises encoding the time series data to generate encoded data.
  • Encoding can preserve sequential ordering information of time series data that may be otherwise lost in a multi -head attention process.
  • encoding comprises transforming input data (e.g., time series data) into a series of vectors comprising numerical values. The numerical values may correspond to light source settings and light pulse measurements.
  • encoding may further comprise relating settings from a vector in the series of vectors to pulse measurements from the preceding vector in the series of vectors.
  • An example of encoded data is shown in FIG. 8 below.
  • Step 706 comprises training a machine learning model, such as atransformer architecture, using the encoded data as inputs to the model.
  • a series of vectors comprising numerical values as encoded in step 704 may constitute input vectors 605 used to train transformer architecture 600 as described above.
  • Parameters of the machine learning model may be initialized at the start of training.
  • loss may be calculated by comparing the output of the machine learning model to an expected output.
  • an expected output may be an actual next light pulse signal in atime series of light pulse signals.
  • Parameters of the machine learning model may be adjusted until the loss is minimized.
  • parameters of the machine learning model include weighted matrices in a multi-head attention mechanism and/or weights of a fully connected feed-forward network sub-layer.
  • Training may additionally comprise setting a learning rate for the machine learning model.
  • the learning rate may be varied over the course of training. For example, the learning rate may be increased during warmup training steps and decreased thereafter.
  • the method steps of FIG. 7 are repeated to retrain the machine learning model after a period of deployment. For example, if model performance starts to lose accuracy, data collected during operation of a light source may be used adjust parameters of the machine learning model (i.e., retrain).
  • the machine learning model can be trained (or retrained/tuned) to make predictions for a specific combinations and/or subsets of light sources.
  • a machine learning model is configured to make predictions for a particular model of laser, or for a particular configuration of that model (e.g., a configuration with a particular version of LAM or with a particular software upgrade).
  • the machine learning model can be generalized to make predictions for a larger combination of light source models and configurations.
  • the machine learning models can also be tuned to make predictions for light sources used by a particular customer or a particular fabrication facility using only training data from the customer or fabrication facility.
  • the machine learning model can be trained to make predictions for a light source that is being used for a particular type of wafer (e.g., memory or logic).
  • the training (or retraining) data may be slightly different.
  • a model configured to make predictions for multiple types of lasers may using training data from of multiple types of lasers, while a model configured to only make predictions for a specific model of laser may only train on data from the specific model of laser.
  • FIG. 8 shows an example of a subset of a time series of encoded vectors 800, according to some aspects.
  • Time series of encoded vectors 800 may include vectorized light pulse signals 802a, 802b, and 802c.
  • numerical values of vectorized light pulse signals 802a- c may be grouped into settings 804a-c and measurements 806a-c.
  • encoding includes relating measurements of a vectorized light pulse to settings in the next vectorized light pulse in the time series. For example, settings 804b may correspond to measurements 806a, and settings 804c may correspond to measurements 806b.
  • FIG. 9 shows a flowchart of a method 900 of predicting a next light pulse signal in a light source using a trained machine learning model, given a prior light source signal, according to some aspects.
  • Predicting a light pulse signal may include predicting values such as pulse energy, wavelength, bandwidth, timing error, and other settings and measurements as described in Table 1.
  • Method 900 may comprise steps 902, 904, 906, and 908.
  • Step 902 comprises obtaining time series data.
  • the time series data may comprise one or more light pulse signals of a light source.
  • a light pulse signal may comprise values corresponding to light source settings (e.g., temperature, pressure, time since last pulse) and light pulse measurements (e.g., energy, bandwidth, wavelength).
  • Light pulse measurements may be captured by one or more sensors in a system.
  • the one or more light pulse signals are collected by a light source during operation and sent to a processor.
  • the processor may be external to the light source.
  • Step 904 comprises encoding the time series data to generate encoded data. Encoding can preserve sequential ordering information of time series data that may be otherwise lost in a multi -head attention process.
  • encoding comprises transforming input data (e.g., time series data) into a series of vectors comprising numerical values.
  • the numerical values may correspond to light source settings and light pulse measurements.
  • encoding may further comprise relating settings from a vector in the series of vectors to pulse measurements from the preceding vector in the series of vectors, as illustrated in FIG. 8.
  • Step 906 comprises determining the next light pulse signal of a light source using a trained machine learning model, such as described with respect to FIG. 6.
  • the machine learning model may comprise a transformer architecture, as further described with respect to FIG. 6.
  • the output of the machine learning model may comprise a vector containing predicted values corresponding to settings and measurements of the next light pulse.
  • the output of the machine learning model may be a probability distribution comprising a plurality of possible values for each setting and/or measurement and each possible value’s relative probability.
  • a prediction of a value can include an estimate or other calculation of a single value such as an integer number or real number with or without units (e.g., 73, 5.6, 5.2 mJ), or a binary value (“out of range,” or “in range,” or “TRUE,” or “FALSE”) or a trinary value (“low,” or “medium,” or “high”), or other numeric or textual value (“pulse energy fault,” “LNM near tuning limit”), among others.
  • a prediction of an upcoming value can include an estimate or other calculation of a statistical description of the value, such as a range of values (7-12, 3. 1-4.3, 9.1 mJ — 10.3 mJ), or a confidence interval, or a standard deviation, or a peak, mean, or median value of a statistical distribution, among others.
  • Step 908 comprises advancing a light source simulation based on the output of the machine learning model.
  • the machine learning model may use a predicted next pulse to generate another next pulse. This cycle can repeat, using each predicted next pulse as a new prior pulse, to predict a series of several subsequent pulses of a light source.
  • a module may be configured to implement the steps of method 900.
  • the module can include hardware, software, or a combination thereof.
  • the module can also be a computer readable medium with instructions that can be read and executed by one or more processors.
  • the module may comprise a software program loaded onto a computer system in a lithography apparatus.
  • the module may receive light pulse signals from a light source.
  • the module may also be configured to send instructions, such as setting modifications, to a light source based on results of the light source simulation.
  • the steps of method 900 may be used to perform a sensitivity analysis of a light source.
  • data collected by one or more sensors during operation of a light source may be sent to a processor.
  • the processor can encode the data, and determine a sequence of several subsequent light pulse signals using the trained machine learning model.
  • the sequence of subsequent light pulse signals may be analyzed to determine whether a light source is likely to stay within a specification.
  • the light source may then be conditioned as needed based on the likelihood of the light source to stay within the specification. For example, if the light source is likely to drift outside the specification, the light source may be conditioned in such a way that actual future pulses are altered so as to stay within the specification.
  • a light source simulation may act as a virtual laboratory for a light source.
  • Time series data collected by one more sensors during light source operation can be modified before the data is encoded. For example, setting values, such as temperature, pressure, and time since last shot can be altered. Subsequent light pulses can be simulated using the altered data to determine how a light source will behave under different operating conditions.
  • the machine learning models can be tuned to make predictions for various combinations and subsets of light sources.
  • a machine learning model is configured to make predictions for a particular model of laser, or for a particular configuration of that model (e.g., a configuration with a particular version of LAM or with a particular software upgrade).
  • the machine learning model can be customized to make predictions for a combination of light source models and configurations.
  • the machine learning models can also be tuned to make predictions for light sources used by a particular customer or a particular fabrication facility using only training data from the customer or fabrication facility.
  • the training data may be slightly different.
  • a machine learning model configured to make predictions for multiple types and/or models of lasers may using training data from of multiple models of lasers, while a model configured to make predictions on a single type or model of laser may only use training data from a single model of laser.
  • a processor can use a series of predicted light pulse signals to determine ideal operating parameters of a light source. The processor may send modification instructions to a controller in the light source.
  • the method steps of FIG. 9 are performed in real-time, such that a next light pulse can be predicted after emission of the prior light pulse but prior to the actual emission of that next light pulse. Such real-time performance allows feedback to be sent to the controller in real time to modify the next pulse if needed.
  • time series data collected during operation of the light source may be used to retrain the machine learning model. For example, steps 704 and 706 of method 700 may be repeated using data collected during step 902.
  • FIG. 10 illustrates an example computer system useful for implementing various embodiments in Figures 1-9.
  • FIG. 10 Various embodiments may be implemented, for example, using one or more well-known computer systems, such as computer system 1000 shown in FIG. 10.
  • One or more computer systems 1000 may be used, for example, to implement any of the embodiments discussed herein, as well as combinations and sub-combinations thereof.
  • Computer system 1000 may include one or more processors (also called central processing units, or CPUs), such as a processor 1004.
  • processors also called central processing units, or CPUs
  • Processor 1004 may be connected to a communication infrastructure or bus 1006.
  • Computer system 1000 may also include user input/output device(s) 1003, such as monitors, keyboards, pointing devices, cameras, other imaging devices etc., which may communicate with communication infrastructure 1006 through user input/output interface(s) 1002.
  • user input/output device(s) 1003 such as monitors, keyboards, pointing devices, cameras, other imaging devices etc.
  • communication infrastructure 1006 may communicate with user input/output interface(s) 1002.
  • processors 1004 may be a graphics processing unit (GPU).
  • a GPU may be a processor that is a specialized electronic circuit designed to process mathematically intensive applications.
  • the GPU may have a parallel structure that is efficient for parallel processing of large blocks of data, such as mathematically intensive data common to computer graphics applications, images, videos, etc.
  • Computer system 1000 may also include a main or primary memory 1008, such as random access memory (RAM).
  • Main memory 1008 may include one or more levels of cache.
  • Main memory 1008 may have stored therein control logic (i.e., computer software) and/or data.
  • Computer system 1000 may also include one or more secondary storage devices or memory 1010.
  • Secondary memory 1010 may include, for example, a hard disk drive 1012 and/or a removable storage device or drive 1014.
  • Removable storage drive 1014 may be a floppy disk drive, a magnetic tape drive, a compact disk drive, an optical storage device, tape backup device, and/or any other storage device/drive.
  • Removable storage drive 1014 may interact with a removable storage unit 1018.
  • Removable storage unit 1018 may include a computer usable or readable storage device having stored thereon computer software (control logic) and/or data.
  • Removable storage unit 1018 may be a floppy disk, magnetic tape, compact disk, DVD, optical storage disk, and/ any other computer data storage device.
  • Removable storage drive 1014 may read from and/or write to removable storage unit 1018.
  • Secondary memory 1010 may include other means, devices, components, instrumentalities or other approaches for allowing computer programs and/or other instructions and/or data to be accessed by computer system 1000. Such means, devices, components, instrumentalities or other approaches may include, for example, a removable storage unit 1022 and an interface 1020.
  • Examples of the removable storage unit 1022 and the interface 1020 may include a program cartridge and cartridge interface (such as that found in video game devices), a removable memory chip (such as an EPROM or PROM) and associated socket, a memory stick and USB port, a memory card and associated memory card slot, and/or any other removable storage unit and associated interface.
  • a program cartridge and cartridge interface such as that found in video game devices
  • a removable memory chip such as an EPROM or PROM
  • associated socket such as an EPROM or PROM
  • a memory stick and USB port such as an EPROM or PROM
  • Computer system 1000 may further include a communication or network interface 1024.
  • Communication interface 1024 may enable computer system 1000 to communicate and interact with any combination of external devices, external networks, external entities, etc. (individually and collectively referenced by reference number 1028).
  • communication interface 1024 may allow computer system 1000 to communicate with external or remote devices 1028 over communications path 1026, which may be wired and/or wireless (or a combination thereof), and which may include any combination of LANs, WANs, the Internet, etc.
  • Control logic and/or data may be transmitted to and from computer system 1000 via communication path 1026.
  • Computer system 1000 may also be any of a personal digital assistant (PDA), desktop workstation, laptop or notebook computer, netbook, tablet, smart phone, smart watch or other wearable, appliance, part of the Intemet-of-Things, and/or embedded system, to name a few nonlimiting examples, or any combination thereof.
  • PDA personal digital assistant
  • Computer system 1000 may be a client or server, accessing or hosting any applications and/or data through any delivery paradigm, including but not limited to remote or distributed cloud computing solutions; local or on-premises software (“on-premise” cloud-based solutions); “as a service” models (e.g., content as a service (CaaS), digital content as a service (DCaaS), software as a service (SaaS), managed software as a service (MSaaS), platform as a service (PaaS), desktop as a service (DaaS), framework as a service (FaaS), backend as a service (BaaS), mobile backend as a service (MBaaS), infrastructure as a service (laaS), etc.); and/or a hybrid model including any combination of the foregoing examples or other services or delivery paradigms.
  • “as a service” models e.g., content as a service (CaaS), digital content as a service (DCaaS), software as a service
  • Any applicable data structures, file formats, and schemas in computer system 1000 may be derived from standards including but not limited to JavaScript Object Notation (JSON), Extensible Markup Language (XML), Yet Another Markup Language (YAML), Extensible Hypertext Markup Language (XHTML), Wireless Markup Language (WML), MessagePack, XML User Interface Language (XUL), or any other functionally similar representations alone or in combination.
  • JSON JavaScript Object Notation
  • XML Extensible Markup Language
  • YAML Yet Another Markup Language
  • XHTML Extensible Hypertext Markup Language
  • WML Wireless Markup Language
  • MessagePack XML User Interface Language
  • XUL XML User Interface Language
  • a tangible, non-transitory apparatus or article of manufacture comprising a tangible, non-transitory computer useable or readable medium having control logic (software) stored thereon may also be referred to herein as a computer program product or program storage device.
  • control logic software stored thereon
  • control logic when executed by one or more data processing devices (such as computer system 1000), may cause such data processing devices to operate as described herein.
  • UV radiation for example, having a wavelength X of 365, 248, 193, 157 or 126 nm
  • extreme ultraviolet (EUV or soft X-ray) radiation for example, having a wavelength in the range of 5-100 nm such as, for example, 13.5 nm
  • hard X-ray working at less than 5 nm as well as particle beams, such as ion beams or electron beams.
  • UV refers to radiation with wavelengths of approximately 100-400 nm.
  • Vacuum UV, or VUV refers to radiation having a wavelength of approximately 100-200 nm.
  • Deep UV generally refers to radiation having wavelengths ranging from 126 nm to 428 nm, and in some aspects, an excimer laser can generate DUV radiation used within a lithographic apparatus. It should be appreciated that radiation having a wavelength in the range of, for example, 5-20 nm relates to radiation with a certain wavelength band, of which at least part is in the range of 5-20 nm.
  • An optical system comprising: a processor; and a memory having instructions stored thereon that, when executed, cause the processor to: obtain time series data of light pulse signals from a light source; encode the time series data to obtain encoded data, wherein the encoded data comprises a time series of vectors, wherein each vector contains one or more values corresponding to features of a light pulse signal in the time series data of light pulse signals; determine, using a machine learning model, one or more values corresponding to features of a next light pulse signal; and advance an optical system simulation based on the one or more values corresponding to features of the next light pulse signal.
  • a method comprising: obtaining time series data of a plurality of light pulse signals collected from one or more light sources; encoding the time series data to generate encoded data, wherein the encoded data comprises a time series of vectors, wherein each vector in the time series of vectors contains one or more values corresponding to one or more features of a light pulse signal in the plurality of light pulse signals; and training a machine learning model using the encoded data, wherein the trained machine learning model is configured to determine one or more subsequent light pulse signals of a light source given a prior light pulse signal.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Plasma & Fusion (AREA)
  • Exposure And Positioning Against Photoresist Photosensitive Materials (AREA)

Abstract

Un système optique comprend un processeur et une mémoire. La mémoire contient des instructions qui, lorsqu'elles sont exécutées, amènent le processeur à obtenir des données de série chronologique de signaux d'impulsion lumineuse provenant d'une source de lumière. Le processeur peut coder les données de série chronologique pour obtenir des données codées qui comprennent une série chronologique de vecteurs. Chaque vecteur de la série chronologique de vecteurs comprend une ou plusieurs valeurs correspondant à des caractéristiques d'un signal d'impulsion lumineuse dans les données chronologiques de signaux d'impulsion lumineuse. Le processeur peut utiliser un modèle d'apprentissage machine pour déterminer une ou plusieurs valeurs correspondant à des caractéristiques d'un signal d'impulsion lumineuse subséquent et faire avancer une simulation de système optique sur la base des caractéristiques du signal d'impulsion lumineuse subséquent.
PCT/IB2025/053611 2024-05-01 2025-04-04 Systèmes et procédés de prédiction d'impulsion subséquente dans une source de lumière Pending WO2025229431A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US202463641367P 2024-05-01 2024-05-01
US63/641,367 2024-05-01

Publications (1)

Publication Number Publication Date
WO2025229431A1 true WO2025229431A1 (fr) 2025-11-06

Family

ID=95474637

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/IB2025/053611 Pending WO2025229431A1 (fr) 2024-05-01 2025-04-04 Systèmes et procédés de prédiction d'impulsion subséquente dans une source de lumière

Country Status (1)

Country Link
WO (1) WO2025229431A1 (fr)

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7372056B2 (en) 2005-06-29 2008-05-13 Cymer, Inc. LPP EUV plasma source material target delivery system
US7405416B2 (en) 2005-02-25 2008-07-29 Cymer, Inc. Method and apparatus for EUV plasma source target delivery
US7511799B2 (en) 2006-01-27 2009-03-31 Asml Netherlands B.V. Lithographic projection apparatus and a device manufacturing method
US7872245B2 (en) 2008-03-17 2011-01-18 Cymer, Inc. Systems and methods for target material delivery in a laser produced plasma EUV light source
WO2019137846A1 (fr) 2018-01-12 2019-07-18 Asml Netherlands B.V. Appareil et procédé de commande de coalescence de gouttelettes dans un flux de gouttelettes
US20210349396A1 (en) * 2020-05-07 2021-11-11 Taiwan Semiconductor Manufacturing Co., Ltd. System and method for performing extreme ultraviolet photolithography processes
WO2022225647A1 (fr) * 2021-04-19 2022-10-27 Cymer, Llc Formation de multiples images aériennes en un seul passage d'exposition lithographique
US20230223734A1 (en) * 2020-06-09 2023-07-13 Cymer, Llc Systems and methods for controlling a center wavelength

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7405416B2 (en) 2005-02-25 2008-07-29 Cymer, Inc. Method and apparatus for EUV plasma source target delivery
US7372056B2 (en) 2005-06-29 2008-05-13 Cymer, Inc. LPP EUV plasma source material target delivery system
US7511799B2 (en) 2006-01-27 2009-03-31 Asml Netherlands B.V. Lithographic projection apparatus and a device manufacturing method
US7872245B2 (en) 2008-03-17 2011-01-18 Cymer, Inc. Systems and methods for target material delivery in a laser produced plasma EUV light source
WO2019137846A1 (fr) 2018-01-12 2019-07-18 Asml Netherlands B.V. Appareil et procédé de commande de coalescence de gouttelettes dans un flux de gouttelettes
US20210349396A1 (en) * 2020-05-07 2021-11-11 Taiwan Semiconductor Manufacturing Co., Ltd. System and method for performing extreme ultraviolet photolithography processes
US20230223734A1 (en) * 2020-06-09 2023-07-13 Cymer, Llc Systems and methods for controlling a center wavelength
WO2022225647A1 (fr) * 2021-04-19 2022-10-27 Cymer, Llc Formation de multiples images aériennes en un seul passage d'exposition lithographique

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
VASHWANI ET AL.: "Attention is all you need", ARXIV:1706.03762V7, 2023

Similar Documents

Publication Publication Date Title
US9052605B2 (en) Illumination system for lithographic apparatus with control system to effect an adjustment of an imaging parameter
TW202113500A (zh) 用於改善圖案化製程之訓練機器學習模型的方法
US10859926B2 (en) Methods for defect validation
US11183434B2 (en) Methods of guiding process models and inspection in a manufacturing process
US20230010985A1 (en) Source material delivery system, euv radiation system, lithographic apparatus, and methods thereof
US11474436B2 (en) Tuning patterning apparatus based on optical characteristic
CN113728277B (zh) 光刻装置和照射均匀性校正系统
WO2025229431A1 (fr) Systèmes et procédés de prédiction d'impulsion subséquente dans une source de lumière
US20240142959A1 (en) Method for decoupling sources of variation related to semiconductor manufacturing
KR102890150B1 (ko) 버스트 통계 데이터 통합 필터
US12444901B2 (en) Seed laser system for radiation source
KR20240021180A (ko) 레티클용 냉각 후드
US20240353756A1 (en) Lithographic method to enhance illuminator transmission
CN114930656B (zh) 具有独立电压和定时控制并降低功耗的双脉冲功率系统
WO2025141350A1 (fr) Ouverture de faisceau laser
WO2025256809A1 (fr) Systèmes et procédés d'augmentation de sortie optique laser
WO2025186638A1 (fr) Procédé et système de mesure de longueur d'onde pour appareil lithographique ou appareil d'inspection
WO2025133738A1 (fr) Systèmes et procédés de commande en temps réel de carré sommable dans le temps (tis) et de contraste de granularité (sc) dans les lasers duv
WO2024193946A1 (fr) Procédé et système pour la génération d'un profil de pupille sensible à un processus lithographique
WO2025233713A1 (fr) Électrodes dans une chambre de décharge pour source de rayonnement