View source for Extreme ultraviolet lithography

{{Short description|Lithography using 13.5 nm UV light}}
'''Extreme ultraviolet lithography''' ('''EUVL''', also known simply as '''EUV''') is a technology used in the [[semiconductor industry]] for manufacturing [[integrated circuit]]s (ICs). It is a type of [[photolithography]] that uses 13.5 nm [[extreme ultraviolet]] (EUV) light from a laser-pulsed tin (Sn) plasma to create intricate patterns on semiconductor substrates. 

{{As of|2023}}, [[ASML Holding]] is the only company that produces and sells EUV systems for chip production, targeting [[5 nm|5 nanometer]] (nm) and [[3 nm]] process nodes.

The EUV wavelengths that are used in EUVL are near 13.5&nbsp;[[nanometer]]s (nm), using a laser-pulsed tin (Sn) droplet plasma to produce a pattern by using a reflective photomask to expose a substrate covered by [[photoresist]].  Tin ions in the ionic states from Sn&nbsp;IX to Sn&nbsp;XIV give photon emission spectral peaks around 13.5&nbsp;nm from 4p<sup>6</sup>4d<sup>''n''</sup> – 4p<sup>5</sup>4d<sup>''n''+1</sup> + 4d<sup>''n''−1</sup>4f ionic state transitions.<ref>{{Cite journal |last1=O'Sullivan |first1=Gerry |last2=Li |first2=Bowen |last3=D'Arcy |first3=Rebekah |last4=Dunne |first4=Padraig |last5=Hayden |first5=Paddy |last6=Kilbane |first6=Deirdre |last7=McCormack |first7=Tom |last8=Ohashi |first8=Hayato |last9=O'Reilly |first9=Fergal |last10=Sheridan |first10=Paul |last11=Sokell |first11=Emma |last12=Suzuki |first12=Chihiro |last13=Higashiguchi |first13=Takeshi |date=2015 |title=Spectroscopy of highly charged ions and its relevance to EUV and soft x-ray source development |url=https://iopscience.iop.org/article/10.1088/0953-4075/48/14/144025 |journal=Journal of Physics B: Atomic, Molecular and Optical Physics |volume=48 |issue=144025 |page=144025 |doi=10.1088/0953-4075/48/14/144025|bibcode=2015JPhB...48n4025O |s2cid=124221931 }}</ref>

[[File:EUV photoelectrons and secondaries (vector).svg|right|thumb|350px|Image formation mechanism in EUV lithography.
{{Legend|grey|EUV multilayer of [[silicon]]-based glass spacer and [[molybdenum]] reflectors}}{{Legend|purple|Absorber}}
{{Legend|red|EUV radiation}}
{{Legend|orange|Resist}}
{{Legend|brown|Substrate}}
{{Legend|blue|Secondary electrons}}
EUV multilayer and absorber constituting mask pattern for imaging a line. EUV radiation reflected from the mask pattern is absorbed in the resist and substrate, producing photoelectrons and secondary electrons. These electrons increase the extent of chemical reactions in the resist. A secondary-electron pattern that is random in nature is superimposed on the optical image. The unwanted secondary-electron exposure results in loss of resolution, observable line edge roughness and linewidth variation.]]

{{toclimit|3}}

==History and economic impact==
In the 1960s, visible light was used for the production of integrated circuits, with wavelengths as small as 435&nbsp;[[Nanometer|nm]] ([[Mercury-vapor lamp#Emission line spectrum|mercury "g&nbsp;line"]]).

Later, [[ultraviolet]] (UV) light was used, at first with a wavelength of 365&nbsp;nm (mercury "i&nbsp;line"), then with excimer wavelengths, first of 248&nbsp;nm ([[krypton fluoride laser]]), then 193&nbsp;nm ([[argon fluoride laser]]), which was called deep UV.

The next step, going even smaller, was called extreme UV, or EUV. The EUV technology was considered impossible by many.

EUV light is absorbed by glass and air, so instead of using lenses to focus the beams of light as done previously, mirrors in vacuum would be needed. A reliable production of EUV was also problematic. Then, leading producers of steppers [[Canon Inc.|Canon]] and [[Nikon]] stopped development, and some predicted the end of [[Moore's law]].{{Citation needed|date=May 2023}}

In 1991, scientists at [[Bell Labs]] published a paper demonstrating the possibility of using a wavelength of 13.8&nbsp;nm for the so-called soft X-ray projection lithography.<ref>{{Cite journal |last1=Bjorkholm |first1=J. |last2=Bokor |first2=J. |last3=Eichner |first3=L. |last4=Freeman |first4=R. |last5=Mansfield |first5=W. |last6=Szeto |first6=L. |last7=Taylor |first7=D. |last8=Tennant |first8=D. |last9=Wood II |first9=O. |last10=Jewell |first10=T. |last11=White |first11=D. |last12=Waskiewicz |first12=W. |last13=Windt |first13=D. |last14=MacDowell |first14=A. |date=1991 |title=Soft x-ray projection lithography |journal=Optics and Photonics News |volume=2 |issue=5 |page=27 |doi=10.1364/OPN.2.5.000027}}</ref>

To address the challenge of EUV lithography, researchers at [[Lawrence Livermore National Laboratory]], [[Lawrence Berkeley National Laboratory]], and [[Sandia National Laboratories]] were funded in the 1990s to perform basic research into the technical obstacles. The results of this successful effort were disseminated via a public/private partnership Cooperative R&D Agreement (CRADA) with the invention and rights wholly owned by the US government, but licensed and distributed under approval by [[United States Department of Energy|DOE]] and Congress.<ref>{{cite web | url=https://www.asml.com/en/news/stories/2022/making-euv-lab-to-fab | title=Making EUV: From lab to fab | date=30 March 2022 }}</ref> The CRADA consisted of a consortium of private companies and the Labs, manifested as an entity called the Extreme Ultraviolet Limited Liability Company (EUV LLC).<ref>{{Cite book |title=EUV lithography |date=2018 |publisher=SPIE Press |isbn=978-1-5106-1679-0 |editor-last=Bakshi |editor-first=Vivek |edition=Second |series=SPIE PM |location=Bellingham, Washington, USA}}</ref>

[[Intel]], Canon, and Nikon (leaders in the field at the time), as well as the [[Netherlands|Dutch]] company [[ASML Holding|ASML]] and Silicon Valley Group (SVG) all sought licensing. Congress denied{{Citation needed|date=June 2024}} the Japanese companies the necessary permission, as they were perceived as strong technical competitors at the time and should not benefit from taxpayer-funded research at the expense of American companies.<ref>{{cite news| title=U.S. gives ok to ASML on EUV effort | website=EE Times | date=24 February 1999 | url=https://www.eetimes.com/u-s-gives-ok-to-asml-on-euv-effort/ | access-date=17 May 2023}}</ref> In 2001 SVG was acquired by ASML, leaving ASML as the sole benefactor of the critical technology.<ref>{{cite web | url=https://www.asml.com/en/company/about-asml/history | title=History |website= ASML}}</ref>

By 2018, ASML succeeded in deploying the intellectual property from the EUV-LLC after several decades of developmental research, with incorporation of European-funded EUCLIDES (Extreme UV Concept Lithography Development System) and long-standing partner German optics manufacturer [[Carl Zeiss AG|ZEISS]] and synchrotron light source supplier Oxford Instruments. This led [[MIT Technology Review]] to name it "the machine that saved Moore's law".<ref>{{cite web | url=https://www.technologyreview.com/2021/10/27/1037118/moores-law-computer-chips/ | title=Inside the machine that saved Moore's Law }}</ref> The first prototype in 2006 produced one [[Wafer (electronics)|wafer]] in 23&nbsp;hours. As of 2022, a scanner produces up to 200 wafers per hour. The scanner uses Zeiss optics, which that company calls "the most precise mirrors in the world", produced by locating imperfections and then knocking off individual molecules with techniques such as ion beam figuring.<ref>{{cite web |author=Sascha Migura |title=Optics for EUV Lithography |url=https://www.euvlitho.com/2018/P22.pdf |website=euvlitho.com |access-date=17 May 2023 |year=2018}}</ref>

This made the once small company ASML the world leader in the production of scanners and monopolist in this cutting-edge technology and resulted in a record turnover of 27.4 billion euros in 2021, dwarfing their competitors Canon and Nikon, who were denied IP access. Because it is such a key technology for development in many fields, the [[United States]] licenser pressured Dutch authorities to not sell these machines to [[China]]. ASML has followed the guidelines of Dutch export controls and until further notice will have no authority to ship the machines to China.<ref>{{cite web | url=https://www.thewirechina.com/2021/02/07/the-chip-chokepoint/ | title=The Chip Choke Point TheWire China | date=8 February 2021 }}</ref> China, at the same time, also has invested heavily into their domestic EUV project, and Chinese leading companies such as [[Huawei]] and [[Shanghai Micro Electronics Equipment|SMEE]] also filled patents for their alternative proposals relevant to EUV technologies.<ref>  Stefano Lovati. February 11, 2025. [https://www.powerelectronicsnews.com/china-invests-e37-billion-to-develop-domestic-euv-lithography-systems/ China Invests €37 Billion to Develop Domestic EUV Lithography Systems] Power Electronics News</ref>

Along with [[multiple patterning]], EUV has paved the way for higher transistor densities, allowing the production of higher-performance processors. Smaller transistors also require less power to operate, resulting in more energy-efficient electronics.

===Market growth projection===

According to a report by Pragma Market Research,<ref>{{Cite web |title=Extreme Ultraviolet (EUV) Lithography Market Size and Forecast to 2032 |url=https://www.pragmamarketresearch.com/reports/121334/extreme-ultraviolet-euv-lithography-market-size |access-date=2024-07-04 |website=PragmaMarketResearch |language=en}}</ref> the global extreme ultraviolet (EUV) lithography market is projected to grow from US$8,957.8&nbsp;million in 2024 to US$17,350&nbsp;million by 2030, at a [[compound annual growth rate]] (CAGR) of 11.7%. This significant growth reflects the rising demand for miniaturized electronics in various sectors, including [[smartphone]]s, [[artificial intelligence]], and [[high-performance computing]].

==Fab tool output==
Requirements for EUV steppers, given the number of layers in the design that require EUV, the number of machines, and the desired throughput of the fab, assuming 24 hours per day operation.<ref>{{cite web |url=https://www.anandtech.com/show/13904/asml-to-ship-30-euv-scanners-in-2019 |title=ASML to Ship 30 EUV Scanners in 2019: Faster EUV Tools Coming |author=Anton Shilov |website=AnandTech |date=2019-01-28}}</ref>
{| class="wikitable" style="text-align:center"
! Number of layers<br /> requiring EUV
! Avg. stepper speed<br /> in wafers per hour
! Number of<br /> EUV machines
! Wafer<br /> per month
|-
| {{0}}5
| 62.5
| {{0}}5
| 45000
|-
| 10
| 62.5
| 10
| 45000
|-
| 15
| 62.5
| 15
| 45000
|-
| 15
| 62.5
| 30
| 90000
|-
| 20
| 62.5
| 40
| 90000
|-
| 25
| 62.5
| 50
| 90000
|}

==Masks==
EUV [[photomask]]s work by reflecting light,<ref>{{Cite web |url=https://www.nist.gov/sites/default/files/documents/pml/div683/conference/Rice_2011.pdf |title=Overview of EUV Mask Metrology |access-date=2019-06-23 |archive-date=2017-06-02 |archive-url=https://web.archive.org/web/20170602020548/https://www.nist.gov/sites/default/files/documents/pml/div683/conference/Rice_2011.pdf }}</ref> which is achieved by using multiple alternating layers of [[molybdenum]] and [[silicon]]. This is in contrast to conventional photomasks which work by blocking light using a single chromium layer on a quartz substrate. An EUV mask consists of 40–50<ref name="auto4">https://www.semiconductor-digest.com/wp-content/uploads/2020/02/SST-Oct-2013.pdf {{Bare URL PDF|date=August 2024}}</ref> alternating silicon and molybdenum layers;<ref>{{Cite conference |title=Multilayer reflective coatings for extreme-ultraviolet lithography |conference=23. SPIE annual international symposium on microlithography conference, Santa Clara, CA (United States), 22-27 Feb 1998 |first=C. |last=Montcalm |date=March 10, 1998 |osti=310916}}</ref> this is a multilayer which acts to reflect the extreme ultraviolet light through [[Bragg diffraction]]; the reflectance is a strong function of incident angle and wavelength, with longer wavelengths reflecting more near normal incidence and shorter wavelengths reflecting more away from normal incidence. The multilayer may be protected by a thin [[ruthenium]] layer, called a capping layer.<ref name="auto4"/><ref>{{Cite book |series=Proceedings of SPIE |volume=10807 |doi = 10.1117/12.2324670 | chapter=EUV capping layer integrity | title=Photomask Japan 2018: XXV Symposium on Photomask and Next-Generation Lithography Mask Technology | date=2018 | editor-last1=Takehisa | editor-first1=Kiwamu | last1=Krome | first1=Thorsten | last2=Schmidt | first2=Jonas | last3=Nesládek | first3=Pavel | page=108070E | isbn=978-1-5106-2201-2 }}</ref><ref>{{cite patent |country=US |number=20030008180A1 |invent1=Sasa Bajt |invent2=James Folta |invent3=Eberhard Spiller |title=Optimized capping layers for EUV multilayers |pridate=2001-07-03 |fdate=2002-02-01 |pubdate=2003-01-09 |url=https://patents.google.com/patent/US20030008180}}.</ref> The pattern is defined in a [[tantalum]]-based absorbing layer over the capping layer.<ref>{{Cite web |url=https://semiengineering.com/next-euv-issue-mask-3d-effects/ |title=Next EUV Issue: Mask 3D Effects |date=April 25, 2018 |website=Semiconductor Engineering}}</ref> 

Blank photomasks are mainly made by two companies: [[AGC Inc.]] and [[Hoya Corporation]].<ref>{{Cite web |url=https://semiengineering.com/euv-mask-blank-biz-heats-up/ |title=EUV Mask Blank Battle Brewing |website=Semiconductor Engineering |date=November 15, 2018}}</ref> Ion-beam deposition equipment mainly made by [[Veeco]] is often used to deposit the multilayer.<ref name="auto4"/> A blank photomask is covered with [[photoresist]], which is then baked (solidified) in an oven, and later the pattern is defined on the photoresist using [[maskless lithography]] with an electron beam. This step is called exposure.<ref>{{Cite book |series=Proceedings of SPIE |volume=11908  |doi=10.1117/12.2604378 |chapter=Multi-beam mask writer, MBM-2000 |title=Photomask Japan 2021: XXVII Symposium on Photomask and Next-Generation Lithography Mask Technology |date=2021 |last1=Matsumoto |first1=Hiroshi |last2=Yamaguchi |first2=Keisuke |last3=Kimura |first3=Hayato |last4=Nakayamada |first4=Noriaki |page=119080L |isbn=978-1-5106-4685-8 |editor-first1=Akihiko |editor-last1=Ando }}</ref> The exposed photoresist is developed (removed), and the unprotected areas are etched. The remaining photoresist is then removed. Masks are then inspected and later repaired using an [[electron beam]].<ref>{{Cite book |series=Proceedings of SPIE |doi=10.1117/12.863542 |chapter=E-beam induced EUV photomask repair: A perfect match |title=26th European Mask and Lithography Conference |date=2010 |editor-last1=Behringer |editor-first1=Uwe F.W. |last1=Waiblinger |first1=M. |last2=Kornilov |first2=K. |last3=Hofmann |first3=T. |last4=Edinger |first4=K. |volume=7545 |pages=75450P |editor-first2=Wilhelm |editor-last2=Maurer }}</ref> Etching must be done only in the absorbing layer<ref name="auto4"/> and thus there is a need to distinguish between the capping and the absorbing layer, which is known as etch selectivity<ref>{{cite book | url=https://books.google.com/books?id=91XeKLC9MUEC&dq=euv+capping+layer+etching&pg=PA360 | isbn=978-0-8194-6964-9 | title=EUV Lithography | date=2009 | publisher=SPIE Press }}</ref> and is unlike etching in conventional photomasks, which only have one layer critical to their function.<ref>{{cite book | url=https://books.google.com/books?id=I-bLBQAAQBAJ&dq=photomask+layers&pg=PA560 | isbn=978-1-4200-2878-2 | title=Handbook of Photomask Manufacturing Technology | date=3 October 2018 | publisher=CRC Press }}</ref>

==Tool==
[[File:Extreme ultraviolet lithography tool.jpg|right|thumb|300 px|An EUVL tool, [[Lawrence Livermore National Laboratory]]]]
An EUV tool (EUV photolithography machine) has a laser-driven tin (Sn) plasma light source, reflective optics comprising multilayer mirrors, contained within a [[hydrogen]] gas ambient.<ref>{{Cite journal |last=Tomie |first=Toshihisa |date=2012-05-21 |title=Tin laser-produced plasma as the light source for extreme ultraviolet lithography high-volume manufacturing: history, ideal plasma, present status, and prospects |journal=Journal of Micro/Nanolithography, MEMS, and MOEMS |language=en |volume=11 |issue=2 |pages=021109–1 |doi=10.1117/1.JMM.11.2.021109 |issn=1932-5150|doi-access=free }}</ref> The hydrogen is used to keep the EUV collector mirror, as the first mirror collecting EUV emitted over a large range in angle (~2π&nbsp;[[Solid angle|sr]]) from the Sn plasma, in the source free of Sn deposition.<ref>{{Cite journal |last1=Elg |first1=Daniel T. |last2=Sporre |first2=John R. |last3=Panici |first3=Gianluca A. |last4=Srivastava |first4=Shailendra N. |last5=Ruzic |first5=David N. |date=2016 |title=In situ collector cleaning and extreme ultraviolet reflectivity restoration by hydrogen plasma for extreme ultraviolet sources |url=https://cpmi.illinois.edu/files/2016/03/In-Situ-Collector-Cleaning-and-EUV-Reflectivity-Restoration-by-Hydrogen-Plasma-for-EUV-Sources.pdf |journal=Journal of Vacuum Science & Technology A |volume=34 |issue=21305 |page=021305 |doi=10.1116/1.4942456|bibcode=2016JVSTA..34b1305E }}</ref> Specifically, the hydrogen buffer gas in the EUV source chamber or vessel decelerates or possibly pushes back Sn ions and Sn debris traveling toward the EUV collector (collector protection) and enable a chemical reaction of <chem>Sn(s) + 4H(g) -> SnH4(g)</chem> to remove Sn deposition on the collector in the form of <chem>SnH4</chem> gas (collector reflectivity restoration).

EUVL is a significant departure from the deep-ultraviolet lithography standard. All matter absorbs [[Extreme ultraviolet|EUV]] radiation. Hence, EUV lithography requires vacuum. All optical elements, including the [[photomask]], must use defect-free [[molybdenum]]/[[silicon]] (Mo/Si) multilayers (consisting of 50 Mo/Si bilayers, which theoretical reflectivity limit at 13.5&nbsp;nm is ~75%<ref>{{Cite journal |last1=Bosgra |first1=Jeroen |last2=Zoethout |first2=Erwin |last3=van der Eerden |first3=Ad M. J. |last4=Verhoeven |first4=Jan |last5=van de Kruijs |first5=Robbert W. E. |last6=Yakshin |first6=Andrey E. |last7=Bijkerk |first7=Fred |date=2012 |title=Structural properties of subnanometer thick Y layers in extreme ultraviolet multilayer mirrors |url=https://opg.optica.org/ao/viewmedia.cfm?uri=ao-51-36-8541&seq=0 |journal=Applied Optics |volume=51 |issue=36 |pages=8541–8548 |doi=10.1364/AO.51.008541|pmid=23262592 |bibcode=2012ApOpt..51.8541B }}</ref>) that act to reflect light by means of interlayer wave interference; any one of these mirrors absorb around 30% of the incident light, so the mirror temperature control is important.

EUVL systems, as of 2002-2009, contain at least two [[Condenser (optics)|condenser]] multilayer mirrors, six projection multilayer mirrors and a multilayer object (mask). Since the mirrors absorb 96% of the EUV light, the ideal EUV source needs to be much brighter than its predecessors. EUV source development has focused on [[Plasma (physics)|plasmas]] generated by laser or discharge pulses. The mirror responsible for collecting the light is directly exposed to the plasma and is vulnerable to damage from high-energy [[ions]]<ref>H. Komori ''et al.'', ''Proc. SPIE'' 5374, pp. 839–846 (2004).</ref><ref>B. A. M. Hansson ''et al.'', ''Proc. SPIE'' 4688, pp. 102–109 (2002).</ref> and other debris<ref>S. N. Srivastava ''et al.'', ''J. Appl. Phys.'' 102, 023301 (2007).</ref> such as tin droplets, which require the costly collector mirror to be replaced every year.<ref>{{Cite web |url=https://www.laserfocusworld.com/optics/article/16566714/optics-for-scanning-multilayer-mirrors-enable-nextgeneration-euv-lithography |title=OPTICS FOR SCANNING: Multilayer mirrors enable next-generation EUV lithography |website=Laser Focus World |author1=Nigel Farrar |author2=David Brandt |author3=Norbert Böwering |date=26 February 2009 }}</ref>

===Resource requirements===
{| class="wikitable"
! Utility
! 200 W output EUV
! 90 W output ArF<br />immersion double-patterning
|-
| Electrical power (kW)
| 532
| 49
|-
| Cooling water flow (L/min)
| 1600
| 75
|-
| Gas lines
| 6
| 3
|}

The required utility resources are significantly larger for EUV compared to [[immersion lithography|193&nbsp;nm immersion]], even with two exposures using the latter. At the 2009 EUV Symposium, [[Hynix]] reported that the wall plug efficiency was ~0.02% for EUV, i.e., to get 200&nbsp;watts at intermediate focus for 100 wafers per hour, one would require 1&nbsp;megawatt of input power, compared to 165&nbsp;kilowatts for an ArF immersion scanner, and that even at the same throughput, the footprint of the EUV scanner was ~3× the footprint of an ArF immersion scanner, resulting in productivity loss.<ref>{{cite web |url=http://www.sematech.org/meetings/archives/litho/8653/pres/Keynote3_Kim_Hynix.pdf |author=H. S. Kim |title=Future of Memory Devices and EUV Lithography |work=2009 EUV Symposium |access-date=2012-10-25 |archive-url=https://web.archive.org/web/20150710163014/http://www.sematech.org/meetings/archives/litho/8653/pres/Keynote3_Kim_Hynix.pdf |archive-date=2015-07-10}}</ref> Additionally, to confine ion debris, a superconducting magnet may be required.<ref>H. Mizoguchi, "Laser Produced Plasma EUV Light Source Gigaphoton Update", EUVL Source Workshop, May 12, 2008.</ref>

A typical EUV tool weighs nearly 200 tons<ref>{{Cite web | url=https://youtube.com/Jv40Viz-KTc?si=aSZbO5T5RVTwP0zq&t=59/ | title=Behind this Door: Learn about EUV, Intel's Most Precise, Complex Machine | website=[[YouTube]]}}</ref> and costs around 180 million USD.<ref name="auto5">{{Cite web |url=https://www.digitimes.com/news/a20240628PD205/euv-technology-2030-adoption-ic-manufacturing-equipment.html |title=Hyper-NA EUV to debut in 2030, primes foundry market for transformation |date=June 28, 2024 |website=DIGITIMES}}</ref>

EUV tools consume at least 10× more energy than immersion tools.<ref>ASML 2020 Annual Report, p. 68.</ref>
{| class="wikitable"
|+ DUV vs. EUV tool energy consumption (measured 2020)
|-
!colspan="2" {{Diagonal split header 2|Parameter|Platform}}
! DUV immersion<br /> NXT:2050i
! EUV NXE:3400C<br /> (30 mJ/cm<sup>2</sup>)
|-
|colspan="2"| Energy consumption
| 0.13 MW
| 1.31 MW
|-
|colspan="2"| Energy efficiency per wafer pass
| 0.45 kWh
| 9.64 kWh
|-
|rowspan="2"| Throughput,<br /> wafers
| per hour
| 296
| 136
|-
| per year
| 2,584,200
| 1,191,360
|}

===Summary of key features===
The following table summarizes key differences between EUV systems in development and [[immersion lithography|ArF immersion]] systems which are widely used in production today:
{| class="wikitable"
!
! EUV
! ArF immersion
|-
| Wavelength
| 2% FWHM bandwidth about 13.5&nbsp;nm
| 193&nbsp;nm
|-
| Photon energy
| 91–93 eV
| 6.4 eV
|-
| Light source
| Sn plasma produced by CO<sub>2</sub> laser hitting Sn droplet<ref>This EUV source has been chosen due to its high CO<sub>2</sub> laser to EUV conversion efficiency (~ 5 % or more). See "Igor Fomenkov, [https://euvlitho.com/2019/S1.pdf EUV Source for Lithography in HVM - performance and prospects], ASML Fellow, Source workshop, Amsterdam, 2019-11-05".</ref><ref>{{Cite journal |last=Yang |first=De-Kun |date=2022-07-13 |title=The development of laser-produced plasma EUV light source |journal=Chip |volume=1 |issue=3 |at=100019 |doi=10.1016/j.chip.2022.100019 |doi-access=free }}</ref>
| ArF excimer laser
|-
| Wavelength bandwidth
| 5.9%<ref>{{Cite web|url=http://www.gigaphoton.com/wp-content/uploads/pdf/6517-23_LPP40W_Manuscript.pdf|title=Gigaphoton|access-date=17 May 2023|archive-date=9 July 2020|archive-url=https://web.archive.org/web/20200709233755/https://www.gigaphoton.com/wp-content/uploads/pdf/6517-23_LPP40W_Manuscript.pdf}}</ref>
| <0.16%<ref>{{Cite web|url=https://www.cymer.com/wp-content/uploads/2018/01/Cymer_SPIE_AdvancedLithography_2017v2-1.pdf|title=Cymer SPIE 2018}}</ref>
|-
| Secondary electrons produced by absorption
| Yes
| No
|-
| Optics
| Reflective multilayers (~40% absorbing per mirror)
| Transmissive lenses
|-
| Numerical aperture (NA)
| 0.25: NXE:3100<br />0.33: NXE:33x0 and NXE:3400B<br />High NA (0.55): Installed in Intel in 2024
| 1.20,<br />1.35
|-
| Resolution spec {{nowrap|''k''<sub>1</sub> {{=}} resolution / (wavelength / numerical aperture)}}
| NXE:3100:{{0|B}} 27&nbsp;nm (''k''<sub>1</sub> = 0.50)<br />NXE:3300B: 22&nbsp;nm (''k''<sub>1</sub> = 0.54),<br />{{0|NXEI3100BJ}} 18&nbsp;nm (''k''<sub>1</sub> = 0.44) with off-axis illumination<br />NXE:3350B: 16&nbsp;nm (''k''<sub>1</sub> = 0.39)<br />NXE:3400B/C, NXE:3600D: 13&nbsp;nm (''k''<sub>1</sub> = 0.32)
| 38&nbsp;nm (''k''<sub>1</sub> = 0.27)
|-
| Flare
| 4%<ref>{{Cite web|url=https://www.euvlitho.com/2018/P22.pdf|title=Zeiss 2018 EUVL Workshop update}}</ref>
| <1%<ref>{{cite web |url=https://staticwww.asml.com/doclib/immersion/en/SPIE2007_XT1900_paper.pdf |title=SPIE 2007 paper |access-date=2018-07-28 |archive-url=https://web.archive.org/web/20170812064448/http://staticwww.asml.com/doclib/immersion/en/SPIE2007_XT1900_paper.pdf |archive-date=2017-08-12}}</ref>
|-
| Illumination
| Central angle 6° off-axis onto reticle
| On axis
|-
| Field size
| 0.25 and 0.33 [[Numerical aperture|NA]]: 26&nbsp;mm × 33&nbsp;mm<br /> High NA: 26&nbsp;mm × 16.5&nbsp;mm<ref>{{Cite web|url=https://www.euvlitho.com/2016/P61.pdf|title=ASML, 2016 EUVL Workshop, p. 14}}</ref>
| 26&nbsp;mm × 33&nbsp;mm
|-
| Magnification
| 0.25 and 0.33 NA: 4× isomorphic <br /> High NA: 4×/8× anamorphic
| 4×
|-
| Ambient
| Vacuum, hydrogen
| Air (exposed wafer area underwater)
|-
| Aberration control (including thermal)
| None
| Yes, e.g., FlexWave<ref>Y. Wang and Y. Liu, Proc. SPIE 9283, 928314 (2014).</ref>
|-
| Illumination slit
| Arc-shaped<ref name=arc186>{{Cite web|url=https://www.zeiss.com/content/dam/smt/downloads/products_and_solutions/photomask-systems/Publications/9231_24.pdf|title=R. Capelli et al., Proc. SPIE 9231, 923109 (2014).|access-date=17 May 2023|archive-date=10 August 2017|archive-url=https://web.archive.org/web/20170810121443/https://www.zeiss.com/content/dam/smt/downloads/products_and_solutions/photomask-systems/Publications/9231_24.pdf}}</ref>
| Rectangular<ref>{{cite web |url=https://staticwww.asml.com/doclib/productandservices/94081.pdf |title=M. van den Brink et al., Proc. SPIE 2726 (1996) |access-date=2018-07-17 |archive-url=https://web.archive.org/web/20170809032333/http://staticwww.asml.com/doclib/productandservices/94081.pdf |archive-date=2017-08-09}}</ref>
|-
| Reticle
| Pattern on reflective multilayer
| Pattern on transmissive substrate
|-
| Wafer pattern shift with reticle vertical position
| Yes (due to reflection); ~1:40<ref>{{cite book |chapter-url=https://www.researchgate.net/publication/241466754 |title=Emerging Lithographic Technologies XII |first1=Thomas |last1=Schmoeller |first2=T. |last2=Klimpel |first3=I. |last3=Kim |first4=G. |last4=Lorusso |first5=A. F. |last5=Myers |first6=Rik |last6=Jonckheere |first7=Anne-Marie |last7=Goethals |first8=K. |last8=Ronse |editor-first1=Frank M. |editor-last1=Schellenberg |chapter=EUV pattern shift compensation strategies |date=14 March 2008 |series=Proceedings of SPIE |volume=6921 |pages=69211B |via=ResearchGate |doi=10.1117/12.772640 |s2cid=121926142}}</ref>
| No
|-
| Pellicle
| Available, but has issues
| Yes
|-
| Wafers per day (depends on tool and dose)
| 1500
| 6000
|-
| Number of tools in field
| >90 (all 0.33 NA tool models)
| >400
|}

The different degrees of resolution among the 0.33 NA tools are due to the different illumination options. Despite the potential of the optics to reach sub-20&nbsp;nm resolution, secondary electrons in resist practically limit the resolution to around 20&nbsp;nm (more on this below).<ref>A. N. Broers, IEEE Trans. Elec. Dev. 28, 1268 (1981).</ref>

==Light source power, throughput, and uptime==
[[File:NXE3400 TPT vs dose.png|thumb|300px|EUV throughput as a function of dose. The [[wafer (electronics)|wafer]] throughput of an EUV tool is actually a function of exposure dose, for a fixed source power.]]
Neutral atoms or condensed matter cannot emit EUV radiation. [[Ionization]] must precede EUV emission in matter. The thermal production of multicharged positive ions is only possible in a hot dense [[Plasma (physics)|plasma]], which itself strongly absorbs EUV.<ref>{{cite journal |first=Y. |last=Tao |title=Characterization of density profile of laser-produced Sn plasma for 13.5 nm extreme ultraviolet source |journal=Appl. Phys. Lett. |volume=86 |issue=20 |page=201501 |year=2005 |doi=10.1063/1.1931825 |display-authors=etal |bibcode=2005ApPhL..86t1501T }}</ref> As of 2025, the established EUV light source is a laser-pulsed tin plasma.<ref>{{cite web |url=http://www.nifs.ac.jp/itc/itc14/abstract/P1-30.html |title=Abstract 107 Last Page |website=www.nifs.ac.jp}}</ref> The ions absorb the EUV light they emit and are easily neutralized by electrons in the plasma to lower charge states, which produce light mainly at other, unusable wavelengths, resulting in a much reduced efficiency of light generation for lithography at higher plasma power density.

The throughput is tied to the source power, divided by the dose.<ref>I. Fomenkov et al., Adv. Opt. Tech. 6, 173 (2017).</ref> A higher dose requires a slower stage motion (lower throughput) if pulse power cannot be increased.

EUV collector reflectivity degrades ~0.1–0.3% per billion 50&nbsp;kHz pulses (~10% in ~2 weeks), leading to loss of uptime and throughput, while even for the first few billion pulses (within one day), there is still 20% (±10%) fluctuation.<ref>I. V. Fomenkov, Proc. SPIE 10957, 1095719 (2019).</ref> This could be due to the accumulating Sn residue mentioned above which is not completely cleaned off.<ref name=snh4/><ref name=sn_contam/> On the other hand, conventional immersion lithography tools for double-patterning provide consistent output for up to a year.<ref>R. Rokitski ''et al.'', Proc. SPIE 7640, 76401Q (2010).</ref>

Recently, the NXE:3400B illuminator features a smaller pupil fill ratio (PFR) down to 20% without transmission loss.<ref name="auto1">M. van de Kerkhof et al., Proc. SPIE 10143, 101430D (2017).</ref> PFR is maximized and greater than 0.2 around a metal pitch of 45&nbsp;nm.<ref name=7nmt2t>Y. Chen et al., J.Vac. Sci. Tech. B35, 06G601 (2017).</ref>

Due to the use of EUV mirrors which also absorb EUV light, only a small fraction of the source light is finally available at the wafer. There are 4 mirrors used for the illumination optics and 6 mirrors for the projection optics. The EUV mask or reticle is itself an additional mirror. With 11 reflections, only ~2% of the EUV source light is available at the wafer.<ref name=mizoguchi>{{Cite web|url=http://pfwww.kek.jp/PEARL/EUV-FEL_Workshop2/Proceedings/07_Mizoguchi.pdf|title=H. Mizoguchi et al., 2017 EUV-FEL Workshop, p. 4.}}</ref>

The throughput is determined by the EUV resist dose, which in turn depends on the required resolution.<ref>{{Cite web |url=https://www.linkedin.com/pulse/pinning-down-euv-resists-resolution-vs-throughput-frederick-chen-aehxf |title=Pinning Down an EUV Resist's Resolution vs. Throughput |website=www.linkedin.com}}</ref> A dose of 40&nbsp;mJ/cm<sup>2</sup> is expected to be maintained for adequate throughput.<ref>I. Seshadri et al., IEDM 2023.</ref>

===Tool uptime===
The EUV light source limits tool uptime besides throughput. In a two-week period, for example, over seven hours downtime may be scheduled, while total actual downtime including unscheduled issues could easily exceed a day.<ref name=mizoguchi/> A dose error over 2% warrants tool downtime.<ref name=mizoguchi/>

The wafer exposure throughput steadily expanded up to around 1000 wafers per day (per system) over the 2019–2022 period,<ref>{{Cite web |url=https://www.linkedin.com/pulse/assessing-euv-wafer-output-2019-2022-frederick-chen |title=Assessing EUV Wafer Output: 2019–2022 |website=www.linkedin.com}}</ref><ref>C. Smeets et al., Proc. SPIE 12494, 1249406 (2023).</ref> indicating substantial idle time, while at the same time running >120 wafers per day on a number of multipatterned EUV layers, for an EUV wafer on average.

===Comparison to other lithography light sources===
[[Extreme ultraviolet|EUV]] (10–121&nbsp;nm) is the band longer than [[X-ray]]s (0.1–10&nbsp;nm) and shorter than the hydrogen [[Lyman-alpha line]].

While state-of-the-art 193&nbsp;nm [[Argon fluoride laser|ArF]] [[excimer laser]]s offer intensities of 200&nbsp;W/cm<sup>2</sup>,<ref>{{cite journal |first=R. |last=Paetzel |editor1-first=Anthony |editor1-last=Yen |title=Excimer lasers for superhigh NA 193-nm lithography |journal=Proc. SPIE |volume=5040 |page=1665 |year=2003 |doi=10.1117/12.485344 |display-authors=etal|series=Optical Microlithography XVI |bibcode=2003SPIE.5040.1665P |s2cid=18953813 }}</ref> lasers for producing EUV-generating plasmas need to be much more intense, on the order of 10<sup>11</sup>&nbsp;W/cm<sup>2</sup>.<ref>{{cite journal |first=S. S. |last=Harilal |s2cid=34621555 |title=Spectral control of emissions from tin doped targets for extreme ultraviolet lithography |journal=J. Phys. D |volume=39 |issue=3 |pages=484–487 |year=2006 |doi=10.1088/0022-3727/39/3/010 |display-authors=etal|bibcode=2006JPhD...39..484H }}</ref> A state-of-the-art ArF immersion lithography 120&nbsp;W light source requires no more than 40&nbsp;kW electrical power,<ref>T. Asayama ''et al.'', Proc. SPIE vol. 8683, 86831G (2013).</ref> while EUV sources are targeted to exceed 40&nbsp;kW.<ref>{{Cite web|url=http://www.euvlitho.com/2013/S2.pdf|title=ASML update Nov. 2013, Dublin}}</ref>

The optical power target for EUV lithography is at least 250&nbsp;W, while for other conventional lithography sources, it is much less.<ref name=mizoguchi/> For example, immersion lithography light sources target 90&nbsp;W, dry ArF sources 45&nbsp;W, and KrF sources 40&nbsp;W. High-NA EUV sources are expected to require at least 500&nbsp;W.<ref name=mizoguchi/>

==EUV-specific optical issues==
===Reflective optics===
[[File:EUV_H-V_Best_Focus.png|thumb|left|EUV H-V difference of focus. Horizontal (H) and vertical (V) mask (reticle) pattern features are focused differently in EUV optical systems. The [[numerical aperture]] (NA) also makes a difference.]]
A fundamental aspect of EUVL tools, resulting from the use of reflective optics, is the [[off-axis illumination]] (at an angle of 6°, in different direction at different positions within the illumination slit)<ref>L. Peters, "Double Patterning Leads Race for 32&nbsp;nm", ''Semiconductor International'', October 18, 2007.</ref> on a multilayer mask (reticle). This leads to shadowing effects resulting in asymmetry in the diffraction pattern that degrade pattern fidelity in various ways as described below.<ref>M. Sugawara ''et al.'', ''J. Vac. Sci. Tech. B'' 21, 2701 (2003).</ref><ref>{{Cite web |url=https://www.youtube.com/watch?v=4B3_Lp4H5fc |title=What is Shadowing in EUV Lithography? |date=28 January 2022 |via=www.youtube.com}}</ref> For example, one side (behind the shadow) would appear brighter than the other (within the shadow).<ref name=euvlmask>{{cite conference |author=Yunfei Deng |author2=Bruno M. La Fontaine |author3=Harry J. Levinson |author4=Andrew R. Neureuther |date=2003 |title=Rigorous EM simulation of the influence of the structure of mask patterns on EUVL imaging |editor=Roxann L. Engelstad |book-title=Emerging Lithographic Technologies VII |volume=5037 |doi=10.1117/12.484986 |s2cid=137035695}}</ref>

The behavior of light rays within the plane of reflection (affecting horizontal lines) is different from the behavior of light rays out of the plane of reflection (affecting vertical lines).<ref>G. McIntyre ''et al.'', Proc. SPIE vol. 7271, 72711C (2009).</ref> Most conspicuously, identically sized horizontal and vertical lines on the EUV mask are printed at different sizes on the wafer.
[[File:36_nm_pitch_2-bar_CD_delta_vs_focus.png|thumb|right|2-bar CD difference vs. focus. The difference between the widths of two adjacent horizontal lines varies as a function of focus.]]

The combination of the off-axis asymmetry and the mask shadowing effect leads to a fundamental inability of two identical features even in close proximity to be in focus simultaneously.<ref>T. Last ''et al.'', Proc. SPIE 9985, 99850W (2016).</ref> One of EUVL's key issues is the asymmetry between the top and bottom line of a pair of horizontal lines (the so-called "two-bar"). Some ways to partly compensate are the use of assist features as well as asymmetric illumination.<ref name=2bar>T. Last ''et al.'', Proc. SPIE vol. 10143, 1014311 (2017).</ref>

An extension of the two-bar case to a grating consisting of many horizontal lines shows similar sensitivity to defocus.<ref>W. Gao ''et al.'', Proc. SPIE vol. 10143, 101430I (2017).</ref> It is manifest in the critical dimension (CD) difference between the top and bottom edge lines of the set of 11 horizontal lines.

Polarization by reflection also leads to partial polarization of EUV light, which favors imaging of lines perpendicular to the plane of the reflections.<ref>{{Cite web |url=https://www.youtube.com/watch?v=agMx-nuL_Qg |title=Polarization by Reflection in EUV Lithography Systems |date=21 August 2022 |via=www.youtube.com}}</ref><ref>{{Cite web |url=https://www.linkedin.com/pulse/growing-significance-polarization-euv-lithography-frederick-chen |title=The Growing Significance of Polarization in EUV Lithography |website=www.linkedin.com}}</ref>

====Pattern shift from defocus (non-telecentricity)====
[[File:P28 dipole leaf image fading.png|thumb|300px|right|Due to different [[phase shift]]s from reflection from the EUV mask, different illumination angles result in different shifts. This results in reduced [[image contrast]], also known as fading.]]
The EUV mask absorber, due to partial transmission, generates a phase difference between the 0th and 1st diffraction orders of a line-space pattern, resulting in image shifts (at a given illumination angle) as well as changes in peak intensity (leading to linewidth changes) which are further enhanced due to defocus.<ref>M. Burkhardt et al., Proc. SPIE 10957, 1095710 (2019).</ref><ref>A. Erdmann, P. Evanschitzky, T. Fuhrer, Proc. SPIE
7271, 72711E (2009).</ref> Ultimately, this results in different positions of best focus for different pitches and different illumination angles. Generally, the image shift is balanced out due to illumination source points being paired (each on opposite sides of the optical axis). However, the separate images are superposed and the resulting image contrast is degraded when the individual source image shifts are large enough. The phase difference ultimately also determines the best focus position.

The multilayer is also responsible for image shifting due to phase shifts from diffracted light within the multilayer itself.<ref>{{Cite web |url=https://www.youtube.com/watch?v=pvLTvoXmqmk |title=Double Diffraction Model of EUV Masks |date=26 September 2021 |via=www.youtube.com}}</ref> This is inevitable due to light passing twice through the mask pattern.<ref>{{Cite web |url=https://www.linkedin.com/pulse/double-diffraction-euv-masks-seeing-through-illusion-symmetry-chen |title=Double Diffraction in EUV Masks: Seeing Through The Illusion of Symmetry |website=www.linkedin.com}}</ref>

The use of reflection causes wafer exposure position to be extremely sensitive to the reticle flatness and the reticle clamp. Reticle clamp cleanliness is therefore required to be maintained. Small ([[milliradian]]-scale) deviations in mask flatness in the local slope, coupled with wafer defocus.<ref>{{cite web |url=http://www.sematech.org/meetings/archives/litho/7853/Flatness/S.%20Yoshitake_NuFlare.pdf |title=EUV Mask Flatness Requirements |access-date=2015-06-26 |archive-url=https://web.archive.org/web/20150626163721/http://www.sematech.org/meetings/archives/litho/7853/Flatness/S.%20Yoshitake_NuFlare.pdf |archive-date=2015-06-26}}</ref> More significantly, mask defocus has been found to result in large overlay errors.<ref>T. Schmoeller ''et al.'', Proc. SPIE vol. 6921, 69211B (2008).</ref><ref>P. Liu '' et al.'', Proc. SPIE vol. 8679, 86790W (2013).</ref> In particular, for a 10&nbsp;nm node metal 1 layer (including 48&nbsp;nm, 64&nbsp;nm, 70&nbsp;nm pitches, isolated, and power lines), the uncorrectable pattern placement error was 1&nbsp;nm for 40&nbsp;nm mask z-position shift.<ref name=10nmppe>M. Sugawara ''et al.'', Proc. SPIE 9048, 90480V (2014).</ref> This is a global pattern shift of the layer with respect to previously defined layers. However, features at different locations will also shift differently due to different local deviations from mask flatness, e.g., from defects buried under the multilayer. It can be estimated that the contribution of mask non-flatness to overlay error is roughly 1/40 times the peak-to-valley thickness variation.<ref name="xchen">X. Chen ''et al.'', Proc. SPIE 10143, 101431F (2017).</ref> With the blank peak-to-valley spec of 50&nbsp;nm, ~1.25&nbsp;nm image placement error is possible. Blank thickness variations up to 80&nbsp;nm also contribute, which lead to up to 2&nbsp;nm image shift.<ref name=xchen/>

The off-axis illumination of the reticle is also the cause of non-telecentricity in wafer defocus, which consumes most of the 1.4&nbsp;nm overlay budget of the NXE:3400 EUV scanner<ref>{{cite web |url=https://www.asml.com/products/systems/twinscan-nxe/twinscan-nxe3400b/en/s46772?dfp_product_id=10850 |title=TWINSCAN NXE:3400B |website=ASML |access-date=2017-07-02 |archive-date=2018-12-15 |archive-url=https://web.archive.org/web/20181215222052/https://www.asml.com/products/systems/twinscan-nxe/twinscan-nxe3400b/en/s46772?dfp_product_id=10850}}</ref> even for design rules as loose as 100&nbsp;nm pitch.<ref name="auto">X. Liu ''et al.'', Proc. SPIE vol. 9048, 90480Q (2014).</ref> The worst uncorrectable pattern placement error for a 24&nbsp;nm line was about 1.1&nbsp;nm, relative to an adjacent 72&nbsp;nm power line, per 80&nbsp;nm wafer focus position shift at a single slit position; when across-slit performance is included, the worst error is over 1.5&nbsp;nm in the wafer defocus window<ref name=10nmppe/> In 2017, an actinic microscope mimicking a 0.33 NA EUV lithography system with 0.2/0.9 quasar 45 illumination showed that an 80&nbsp;nm pitch contact array shifted −0.6 to 1.0&nbsp;nm while a 56&nbsp;nm pitch contact array shifted −1.7 to 1.0&nbsp;nm relative to a horizontal reference line, within a ±50&nbsp;nm defocus window.<ref>O. Wood et al., Proc. SPIE 10450, 1045008 (2017).</ref>

Wafer defocus also leads to image placement errors due to deviations from local mask flatness. If the local slope is indicated by an angle α, the image is projected to be shifted in a 4× projection tool by {{nowrap|1=8α × (DOF/2) = 4α DOF}}, where DOF is the depth of focus.<ref>S. Yoshitake et al., EUV Mask Flatness Requirements: E-beam Mask Writer Supplier Perspective.</ref> For a depth of focus of 100&nbsp;nm, a small local deviation from flatness of 2.5&nbsp;mrad (0.14°) can lead to a pattern shift of 1&nbsp;nm.

Simulations as well as experiments have shown that pupil imbalances in EUV lithography can result in pitch-dependent pattern placement errors.<ref>J.-H. Franke et al., Proc. SPIE 11147, 111470E (2019).</ref><ref>{{Cite web |url=https://www.youtube.com/watch?v=OXJwxQK4S8o |title=Defocus Induced Image Shift in EUV Lithography |date=24 January 2023 |via=www.youtube.com}}</ref> Since the pupil imbalance changes with EUV collector mirror aging or contamination, such placement errors may not be stable over time. The situation is specifically challenging for logic devices, where multiple pitches have critical requirements at the same time.<ref>A. Shchegrov et al., Proc. SPIE 11325, 113251P (2020).</ref> The issue is ideally addressed by multiple exposures with tailored illuminations.<ref>J-H. Franke et al., J. Micro/Nanopatterning, Materials,  and Metrology 21, 030501 (2022).</ref>

====Slit position dependence====
[[File:Rotated_EUV_illumination_through_slit.png|thumb|left|350px|Illumination rotation across ring-field slit. Light [[specular reflection|reflected]] from curved optical surfaces will generate [[arc (geometry)|arc]] segments.<ref>H. N. Chapman and K. A. Nugent,
Proc. SPIE 3767, 225 (1999).</ref> The illumination angles are rotated azimuthally across the arc-shaped slit (right), due to the reflection of an arc-shaped image from each pupil position as a point source (left).<ref>H. Komatsuda, Proc. SPIE 3997, 765 (2000).</ref><ref>Q. Mei et al., Proc. SPIE 8679, 867923 (2013).</ref> The angle-dependent and wavelength-dependent multilayer reflectance distribution pattern is rotated accordingly.]]
The direction of illumination is also highly dependent on slit position, essentially rotated azimuthally.<ref>D. Hellweg et al., Proc. SPIE 7969, 79690H (2011).</ref><ref>K. Hooker et al., Proc. SPIE 10446, 1044604 (2017).</ref><ref name=arc186/><ref name=garetto>A. Garetto et al., J. Micro/Nanolith. MEMS MOEMS 13, 043006 (2014).</ref><ref name=eom>T-S. Eom et al., Proc. SPIE 8679, 86791J (2013).</ref><ref>R. Capelli et al., Proc. SPIE 10957, 109570X (2019).</ref> Nanya Technology and Synopsys found that horizontal vs. vertical bias changed across slit with dipole illumination.<ref name="Proc 2020">J. Fu et al., Proc. SPIE 11323, 113232H (2020).</ref> The rotating plane of incidence (azimuthal range within −25° to 25°) is confirmed in the SHARP actinic review microscope at CXRO which mimics the optics for EUV projection lithography systems.<ref>{{Cite journal |url=https://escholarship.org/uc/item/07h5f8vn |title=Preparing for the Next Generation of EUV Lithography at the Center for X-ray Optics |first1=R. |last1=Miyakawa |first2=P. |last2=Naulleau |date=May 13, 2019 |journal=Synchrotron Radiation News |volume=32 |issue=4 |pages=15–21 |via=escholarship.org |doi=10.1080/08940886.2019.1634432 |bibcode=2019SRNew..32...15M |osti=1582044 |s2cid=202145457 }}</ref> The reason for this is a mirror is used to transform straight rectangular fields into arc-shaped fields.<ref>S. Koo et al., Proc. SPIE 7969, 79691N (2011).</ref><ref>US Patent Application 20070030948.</ref> In order to preserve a fixed plane of incidence, the reflection from the previous mirror would be from a different angle with the surface for a different slit position; this causes non-uniformity of reflectivity. To preserve uniformity, rotational symmetry with a rotating plane of incidence is used.<ref>{{Cite web |url=http://homepage.tudelft.nl/q1d90/FBweb/paraxial%20predesign.pdf |title=M. F. Bal et al., Appl. Opt. 42, 2301 (2003) |access-date=2020-08-02 |archive-date=2019-12-22 |archive-url=https://web.archive.org/web/20191222020938/http://homepage.tudelft.nl/q1d90/FBweb/paraxial%20predesign.pdf }}</ref> More generally, so-called "ring-field" systems reduce aberrations by relying on the rotational symmetry of an arc-shaped field derived from an off-axis annulus.<ref>D. M. Williamson, Proc. SPIE 3482, 369 (1998).</ref> This is preferred, as reflective systems must use off-axis paths, which aggravate aberrations. Hence identical die patterns within different halves of the arc-shaped slit would require different [[Optical proximity correction|OPC]]. This renders them uninspectable by die-to-die comparison, as they are no longer truly identical dies. For pitches requiring dipole, quadrupole, or hexapole illumination, the rotation also causes mismatch with the same pattern layout at a different slit position, i.e., edge vs. center. Even with annular or circular illumination, the rotational symmetry is destroyed by the angle-dependent multilayer reflectance described above. Although the azimuthal angle range is about ±20°<ref name=zeiss>{{Cite web|url=http://16025079.s21d-16.faiusrd.com/0/ABUIABA9GAAgz-HZ3gUo3cmplgU?f=1-P6_CarlZeiss_2018-1018+IWAPS+2018+ZEISS+handout+.pdf|title=Carl Zeiss 2018}}</ref> (field data indicated over 18°<ref>A. V. Pret et al., Proc. SPIE 10809, 108090A (2018).</ref>) on 0.33 NA scanners, at 7&nbsp;nm design rules (36–40&nbsp;nm pitch), the tolerance for illumination can be ±15°,<ref>L. van Look et al., Proc. SPIE 10809, 108090M (2018)</ref><ref>R-H. Kim et al., Proc. SPIE 9776, 97761R (2016).</ref> or even less.<ref name=setten9661>E. van Setten et al., Proc. SPIE 9661, 96610G (2015).</ref><ref>T. E. Brist and G. E. Bailey, Proc. SPIE 5042, 153 (2003).</ref><ref name=lfhex>M. Lim et al., Proc. SPIE 10583, 105830X (2018).</ref> Annular illumination nonuniformity and asymmetry also significantly impact the imaging.<ref>G. Zhang et al., Proc. SPIE 5040, 45 (2003).</ref> Newer systems have azimuthal angle ranges going up to ±30°.<ref> M. van den Kerkhof et al., Proc. SPIE 12051, 120510B (2022).</ref> On 0.33 NA systems, 30 nm pitch and lower already suffer sufficient reduction of pupil fill to significantly affect throughput.<ref>{{Cite web|url=https://frederickchen.substack.com/p/euv-pupil-rotation-impact-on-resolution|title=EUV Pupil Rotation Impact on Resolution|first=Frederick|last=Chen|date=November 4, 2024}}</ref>

The larger incident angle for pitch-dependent dipole illumination trend across slit does not affect horizontal line shadowing so much, but vertical line shadowing does increase going from center to edge.<ref name=cross_slit>{{Cite web |url=https://www.linkedin.com/pulse/horizontal-vertical-line-shadowing-across-slit-low-na-frederick-chen |title=Horizontal, Vertical, and Slanted Line Shadowing Across Slit in Low-NA and High-NA EUV Lithography Systems |website=www.linkedin.com}}</ref> In addition, higher-NA systems may offer limited relief from shadowing, as they target tighet pitches.<ref name=cross_slit/>
[[File:H_and_V_shadowing_across_EUV_slit.png|thumb|300px|right|Horizontal and vertical lines exhibit different shadowing across the slit.]]
The slit position dependence is particularly difficult for the tilted patterns encountered in DRAM.<ref name=eom/> Besides the more complicated effects due to shadowing and pupil rotation, tilted edges are converted to stair shape, which may be distorted by OPC. In fact, the 32&nbsp;nm pitch DRAM by EUV will lengthen up to at least 9''F''<sup>2</sup> cell area, where ''F'' is the active area half-pitch (traditionally, it had been 6''F''<sup>2</sup>).<ref name="Proc 2020"/> With a 2-D self-aligned double-patterning active area cut, the cell area is still lower at 8.9''F''<sup>2</sup>.<ref>K. Lee et al., J. Microlith/Nanolith. MEMS MOEMS 18, 040501 (2019).</ref>

[[Optical aberrations|Aberrations]], originating from deviations of optical surfaces from subatomic (<0.1&nbsp;nm) specifications<ref>K. A. Goldberg et al., Proc. SPIE 5900, 59000G (2005).</ref> as well as thermal deformations<ref>Y. Liu and Y. Li, Opt. Eng. 55, 095108 (2016).</ref><ref name=adaptive>{{cite thesis |degree=Doctoral |title=Adaptive Optics to Counteract Thermal Aberrations: System Design for EUV-Lithography with Sub-nm Precision |first=Saathof |last=R. |date=1 December 2018 |publisher=Technische Universiteit Delft |doi=10.4233/uuid:1d71e3e8-88ce-4260-aeda-af0ee7675445 |doi-access=free}}</ref> and possibly including polarized reflectance effects,<ref>T. S. Jota and R. A. Chipman, Proc. SPIE 9776, 977617 (2016).</ref> are also dependent on slit position,<ref name=litho2017>{{cite web |url=https://nikonereview.com/2017/mentor-graphics-director-details-challenges-for-edge-placement-control-in-2020/ |title=Mentor Graphics Director Details Challenges for Edge Placement Control in 2020 |website=nikonereview.com |access-date=2017-10-24 |archive-date=2018-12-01 |archive-url=https://web.archive.org/web/20181201095015/https://nikonereview.com/2017/mentor-graphics-director-details-challenges-for-edge-placement-control-in-2020/ }}</ref><ref name=adaptive/> as will be further discussed below, with regard to source-mask optimization (SMO). The thermally induced aberrations are expected to exhibit differences among different positions across the slit, corresponding to different field positions, as each position encounters different parts of the deformed mirrors.<ref>M. Habets et al., Proc. SPIE 9776, 97762D (2016).</ref> Ironically, the use of substrate materials with high thermal and mechanical stability make it more difficult to compensate wavefront errors.<ref>M. Bayraktar et al., Opt. Exp. 22, 30623 (2014).</ref>

In combination with the range of wavelengths, the rotated plane of incidence aggravates the already severe stochastic impact on EUV imaging.<ref>{{Cite web |url=https://www.youtube.com/watch?v=xzYHovuzMx8 |title=Nonideal Imaging in EUV Lithography Systems |date=11 September 2021 |via=www.youtube.com}}</ref>

===Wavelength bandwidth (chromatic aberration)===
[[File:Defocus_pattern_shift_vs_wavelength.png|thumb|left|300px|'''Image shift due to defocus depends on wavelength.''' The angular dependence of multilayer reflectance of the object (mask) is different for different wavelengths, leading to different shifts when defocused.]]
Unlike deep ultraviolet (DUV) lithography sources, based on excimer lasers, EUV plasma sources produce light across a broad range of wavelengths<ref>{{Cite journal|title=Out-of-band exposure characterization with the SEMATECH Berkeley 0.3-NA microfield exposure tool|first1=Simi A.|last1=George|first2=Patrick|last2=Nauleau|first3=Senajith|last3=Rekawa|first4=Eric|last4=Gullikson|first5=Charles D.|last5=Kemp|editor2-first=Bruno M|editor2-last=La Fontaine|editor1-first=Frank M|editor1-last=Schellenberg|date=February 23, 2009|journal=Journal of Micro/Nanolithography, MEMS, and MOEMS|volume = 7271|pages = 72710X|doi=10.1117/12.814429|osti = 960237|series = Alternative Lithographic Technologies|bibcode = 2009SPIE.7271E..0XG|s2cid=55241073|url = https://digital.library.unt.edu/ark:/67531/metadc929666/}}</ref> roughly spanning a 2% FWHM bandwidth near 13.5 nm (13.36nm – 13.65nm at 50% power). [[Extreme ultraviolet|EUV]] (10–121nm) is the band longer than [[x-ray|X-Rays]] (0.1–10nm) and shorter than the hydrogen [[Lyman-alpha line]].

Though the EUV spectrum is not completely monochromatic, nor even as spectrally pure as DUV laser sources, the working wavelength has generally been taken to be 13.5&nbsp;nm. In actuality, the reflected power is distributed mostly in the 13.3-13.7&nbsp;nm range.<ref name=chao>{{Cite web|url=https://www2.eecs.berkeley.edu/Pubs/TechRpts/2013/EECS-2013-56.html|title=Measurement and characterization of EUV mask performance at high-NA {{pipe}} EECS at UC Berkeley|website=www2.eecs.berkeley.edu}}</ref> The bandwidth of EUV light reflected by a multilayer mirror used for EUV lithography is over +/-2% (>270 pm);<ref>{{Cite web|url=http://www1.semi.org/eu/sites/semi.org/files/events/presentations/01_Dirk%20Juergens_CarlZeiss.pdf|title=Carl Zeiss SMT GMbH, Semicon Europa, Nov. 16 2018.|access-date=17 May 2023|archive-date=19 June 2020|archive-url=https://web.archive.org/web/20200619053056/http://www1.semi.org/eu/sites/semi.org/files/events/presentations/01_Dirk%20Juergens_CarlZeiss.pdf}}</ref> the phase changes due to wavelength changes at a given illumination angle may be calculated<ref name=lbl>{{Cite web|url=http://henke.lbl.gov/optical_constants/multi2.html|title=Multilayer Reflectivity|website=henke.lbl.gov}}</ref>
and compared to the aberration budget.<ref>Y. Nakajima et al., Proc. SPIE 7379, 73790P (2009).</ref> Wavelength dependence of [[reflectance]]<ref name=lbl/><ref name=chao/> also affects the apodization, or illumination distribution across the pupil (for different angles); different wavelengths effectively 'see' different illuminations as they are reflected differently by the multilayer of the mask.<ref name=davydova>N. Davydova et al., Proc. SPIE 8166, 816624 (2011).</ref><ref name=chao/> This effective source illumination tilt can lead to large image shifts due to defocus.<ref>G. J. Stagaman et al., Proc. SPIE 2726, 146 (1996).</ref> Conversely, the peak reflected wavelength varies across the pupil due to different incident angles.<ref name=chao/><ref>M.F. Ravet et al., Proc. SPIE 5250, 99 (2004).</ref> This is aggravated when the angles span a wide radius, e.g., annular illumination. The peak reflectance wavelength increases for smaller incident angles.<ref>F. Scholze et al., Proc. SPIE 6151, 615137 (2006).</ref> Aperiodic multilayers have been proposed to reduce the sensitivity at the cost of lower reflectivity but are too sensitive to random fluctuations of layer thicknesses, such as from thickness control imprecision or interdiffusion.<ref>{{Cite journal|title=[PDF] Properties of broadband depth-graded multilayer mirrors for EUV optical systems. {{pipe}} Semantic Scholar|year=2010|doi=10.1364/OE.18.006957|pmid=20389715|s2cid=16163302|last1=Yakshin|first1=A. E.|last2=Kozhevnikov|first2=I. V.|last3=Zoethout|first3=E.|last4=Louis|first4=E.|last5=Bijkerk|first5=F.|journal=Optics Express|volume=18|issue=7|pages=6957–71|doi-access=free}}</ref> 

A narrower bandwidth would increase sensitivity to mask absorber and buffer thickness on the 1&nbsp;nm scale.<ref>M. Sugawara et al., J. Micro/Nanolith. MEMS MOEMS 2, 27–33 (2003).</ref>

===Flare===
Flare is the presence of background light originating from scattering off of surface features which are not resolved by the light. In EUV systems, this light can be EUV or out-of-band (OoB) light that is also produced by the EUV source. The OoB light adds the complication of affecting the resist exposure in ways other than accounted for by the EUV exposure. OoB light exposure may be alleviated by a layer coated above the resist, as well as 'black border' features on the EUV mask.{{cn|date=February 2025}} However, the layer coating inevitably absorbs EUV light, and the black border adds EUV mask processing cost.

===Line tip effects===
A key challenge for EUV is the counter-scaling behavior of the line tip-to-tip (T2T) distance as half-pitch (hp) is scaled down.<ref name=setten9661/> This is in part due to lower image contrast for the binary masks used in EUV lithography, which is not encountered with the use of phase shift masks in immersion lithography.<ref>C. S. Choi ''et al.'', Proc. SPIE 9235, 92351R (2014).</ref><ref>[http://www.lithoguru.com/textbook/Chapter10_Figures.ppt Fundamental Principles of Optical Lithography] Chris A. Mack, p. 37.</ref> The rounding of the corners of the line end leads to line end shortening,<ref>C. A. Mack, Microlith. World, 9–4, 25 (2000)</ref> and this is worse for binary masks.<ref>J. S. Petersen et al., Proc. SPIE 3546, 288 (1998).</ref> The use of phase-shift masks in EUV lithography has been studied but encounters difficulties from phase control in thin layers<ref>{{Cite web |url=http://www.sematech.org/meetings/archives/litho/8059/poster/MA-P17-Kim.pdf |title=Optical and Physical Characteristics of EUV Phase Shift Masks |access-date=2017-02-05 |archive-url=https://web.archive.org/web/20170205182557/http://www.sematech.org/meetings/archives/litho/8059/poster/MA-P17-Kim.pdf |archive-date=2017-02-05 }}</ref> as well as the bandwidth of the EUV light itself.<ref>[https://www.euvlitho.com/2011/P19.pdf Thin Half-tone Phase Shift Mask Stack for Extreme Ultraviolet Lithography] Inhwan Lee, Sangsul Lee, Jae Uk Lee, Chang Young Jeong2, Sunyoung Koo, Changmoon Lim, and Jinho Ahn</ref> More conventionally, [[optical proximity correction]] (OPC) is used to address the corner rounding and line-end shortening. In spite of this, it has been shown that the tip-to-tip resolution and the line tip printability are traded off against each other, being effectively CDs of opposite polarity.<ref>L. Yuan et al., Proc. SPIE 8322, 832229 (2012).</ref>

In unidirectional metal layers, tip-to-tip spacing is one of the more severe issues for single exposure patterning. For the 40&nbsp;nm pitch vertical lines, an 18&nbsp;nm nominal tip-to-tip drawn gap resulted in an actual tip-to-tip distance of 29&nbsp;nm with OPC,<ref name=setten9661/> while for 32&nbsp;nm pitch horizontal lines, the tip-to-tip distance with a 14&nbsp;nm nominal gap went to 31&nbsp;nm with OPC.<ref name=2014euvl>E. van Setten et al., Intl. Symp. on EUV Lithography, 2014.</ref> These actual tip-to-tip distances define a lower limit of the half-pitch of the metal running in the direction perpendicular to the tip. In this case, the lower limit is around 30&nbsp;nm. With further optimization of the illumination (discussed in the section on source-mask optimization), the lower limit can be further reduced to around 25&nbsp;nm.<ref>V. M. Blanco Carballo et al., Proc. SPIE 10143, 1014318 (2017).</ref>

For larger pitches, where conventional illumination can be used, the line tip-to-tip distance is generally larger. For the 24&nbsp;nm half-pitch lines, with a 20&nbsp;nm nominally drawn gap, the distance was actually 45&nbsp;nm, while for 32&nbsp;nm half-pitch lines, the same nominal gap resulted in a tip-to-tip distance of 34&nbsp;nm.<ref name=2014euvl/> With OPC, these become 39&nbsp;nm and 28&nbsp;nm for 24&nbsp;nm half-pitch and 32&nbsp;nm half-pitch, respectively.<ref name=setten9231>E. van Setten et al., Proc. SPIE 9231, 923108 (2014).</ref>

==Enhancement opportunities for EUV patterning==
===Assist features===
[[File:Assist_feature_OPC.png|thumb|left|300px|'''Assist feature OPC.''' Assist features help improve the image of isolated features (blue) to be more like dense features (gray). However, the more effective they are, the greater the risk that the assist feature will print (orange).]]
Assist features are often used to help balance asymmetry from non-telecentricity at different slit positions, due to different illumination angles, starting at the 7&nbsp;nm node,<ref>F. Jiang ''et al.'', Proc. SPIE vol. 9422, 94220U (2015).</ref><ref name=af>I. Mochi ''et al.'', Proc. SPIE 9776, 97761S (2015).</ref> where the pitch is ~ 41&nbsp;nm for a wavelength ~13.5&nbsp;nm and NA=0.33, corresponding to k1 ~ 0.5.<ref>J. G. Garofalo et al., Proc. SPIE 2440, 302 (1995).</ref> However, the asymmetry is reduced but not eliminated, since the assist features mainly enhance the highest spatial frequencies, whereas intermediate spatial frequencies, which also affect feature focus and position, are not much affected. The coupling between the primary image and the self images is too strong for the asymmetry to be eliminated by assist features; only asymmetric illumination can achieve this.<ref name=2bar/> Assist features may also get in the way of access to power/ground rails. Power rails are expected to be wider, which also limits the effectiveness of using assist features, by constraining the local pitch. Local pitches between 1× and 2× the minimum pitch forbid assist feature placement, as there is simply no room to preserve the local pitch symmetry. In fact, for the application to the two-bar asymmetry case, the optimum assist feature placement may be less than or exceed the two-bar pitch.<ref name=af/> Depending on the parameter to be optimized (process window area, depth of focus, exposure latitude), the optimum assist feature configuration can be very different, e.g., pitch between assist feature and bar being different from two-bar pitch, symmetric or asymmetric, etc..

At pitches smaller than 58&nbsp;nm, there is a tradeoff between depth of focus enhancement and contrast loss by assist feature placement.<ref name=af/> Generally, there is still a focus-exposure tradeoff as the dose window is constrained by the need to have the assist features not print accidentally.

An additional concern comes from shot noise;<ref>{{Cite web|url=https://www.jstage.jst.go.jp/article/photopolymer/26/5/26_617/_pdf|title=Understanding EUV Shot Noise}}</ref> sub-resolution assist features (SRAFs) cause the required dose to be lower, so as not to print the assist features accidentally.<ref>D. Civay et al., Proc. SPIE 9048, 90483D (2014).</ref> This results in fewer photons defining smaller features (see discussion in section on shot noise).

As SRAFs are smaller features than primary features and are not supposed to receive doses high enough to print, they are more susceptible to stochastic dose variations causing printing errors; this is particularly prohibitive for EUV, where phase-shift masks may need to be used.<ref>A.Burov et al., Proc. SPIE 11518, 115180Y (2020).</ref>

===Source-mask optimization===
[[File:EUV SMO effectiveness vs. pitch.png|thumb|right|250px|'''Pitch effect on SMO.''' SMO carried out targeted for one pitch may have varying performance for other pitches.]]
Due to the effects of non-telecentricity, standard illumination pupil shapes, such as disc or annular, are not sufficient to be used for feature sizes of ~20&nbsp;nm or below (10&nbsp;nm node and beyond).<ref name="auto"/> Instead certain parts of the pupil (often over 50%) must be asymmetrically excluded. The parts to be excluded depend on the pattern. In particular, the densest allowed lines need to be aligned along one direction and prefer a dipole shape. For this situation, double exposure lithography would be required for 2D patterns, due to the presence of both X- and Y-oriented patterns, each requiring its own 1D pattern mask and dipole orientation.<ref>A-Y. Je ''et al.'', Proc. SPIE 7823, 78230Z (2010).</ref><ref>T. Huynh-Bao ''et al.'', Proc. SPIE 9781, 978102 (2016).</ref> There may be 200–400 illuminating points, each contributing its weight of the dose to balance the overall image through focus. Thus the shot noise effect (to be discussed later) critically affects the image position through focus, in a large population of features.

Double- or multiple-patterning would also be required if a pattern consists of sub-patterns which require significantly different optimized illuminations, due to different pitches, orientations, shapes, and sizes.

====Impact of slit position and aberrations====
[[File:EUV_pupil_wavelength_dependence.png|thumb|left|300px|'''Impact of different wavelengths.''' Different wavelengths effectively have different pupils, resulting in different results of source-mask optimization.]]
Largely due to the slit shape,<ref name=zeiss/> and the presence of residual aberrations,<ref>V. Philipsen et al., Proc. SPIE 9235, 92350J (2014).</ref> the effectiveness of SMO varies across slit position.<ref name="auto2">W. Gillijns et al., Proc. SPIE 10143, 1014314 (2017).</ref> At each slit position, there are different aberrations<ref name=litho2017/> and different azimuthal angles of incidence leading to different shadowing.<ref name="arc186"/> Consequently, there could be uncorrected variations across slit for aberration-sensitive features, which may not be obviously seen with regular line-space patterns.<ref name=af/> At each slit position, although [[optical proximity correction|optical proximity correction (OPC)]], including the assist features mentioned above, may also be applied to address the aberrations,<ref name=pole_aberration>Y-G Wang et al., Proc. SPIE 10143, 1014320 (2017).</ref><ref>US Patent 9715170.</ref> they also feedback into the illumination specification,<ref>S. Nagahara et al., Proc. SPIE 7640, 76401H (2010).</ref><ref name="auto2"/><ref>L. Pang et al., Proc. SPIE 7520, 75200X (2009).</ref><ref>{{cite journal|title=Challenges of anamorphic high-NA lithography and mask making|first1=Stephen D.|last1=Hsu|first2=Jingjing|last2=Liu|date=1 January 2017|journal=Advanced Optical Technologies|volume=6|issue=3–4|page=293|doi=10.1515/aot-2017-0024|bibcode=2017AdOT....6..293H|s2cid=67056068}}</ref> since the benefits differ for different illumination conditions.<ref name=pole_aberration/> This would necessitate the use of different source-mask combinations at each slit position, i.e., multiple mask exposures per layer.<ref name=litho2017/><ref>{{Cite journal|doi=10.1364/OE.418242 |title=Source mask optimization for extreme-ultraviolet lithography based on thick mask model and social learning particle swarm optimization algorithm |year=2021 |last1=Zhang |first1=Zinan |last2=Li |first2=Sikun |last3=Wang |first3=Xiangzhao |last4=Cheng |first4=Wei |last5=Qi |first5=Yuejing |journal=Optics Express |volume=29 |issue=4 |pages=5448–5465 |pmid=33726081 |bibcode=2021OExpr..29.5448Z |s2cid=232263498 |doi-access=free }}</ref>

The above-mentioned chromatic aberrations, due to mask-induced apodization,<ref name=davydova/> also lead to inconsistent source-mask optimizations for different wavelengths.

====Pitch-dependent focus windows====
The best focus for a given feature size varies as a strong function of pitch, polarity, and orientation under a given illumination.<ref>{{Cite web|url=https://www.euvlitho.com/2018/P62.pdf|title=IMEC EUVL 2018 Workshop}}</ref> At 36&nbsp;nm pitch, horizontal and vertical darkfield features have more than 30&nbsp;nm difference of focus. The 34&nbsp;nm pitch and 48&nbsp;nm pitch features have the largest difference of best focus regardless of feature type. In the 48–64&nbsp;nm pitch range, the best focus position shifts roughly linearly as a function of pitch, by as much as 10–20&nbsp;nm.<ref>C. Krautschik et al., Proc. SPIE 4343, 392 (2001).</ref> For the 34–48&nbsp;nm pitch range, the best focus position shifts roughly linearly in the opposite direction as a function of pitch. This can be correlated with the phase difference between the zero and first diffraction orders.<ref>A. Erdmann, P. Evanschitzky, and T. Fuhner, Proc. SPIE 7271, 72711E (2009).</ref> Assist features, if they can fit within the pitch, were found not to reduce this tendency much, for a range of intermediate pitches,<ref>A. Erdmann et al., J. Micro/Nanolith. MEMS MOEMS 15, 021205 (2016).</ref> or even worsened it for the case of 18–27&nbsp;nm and quasar illumination.<ref>M. Burkhardt and A. Raghunathan, Proc. SPIE 9422, 94220X (2015).</ref> 50&nbsp;nm contact holes on 100&nbsp;nm and 150 pitches had best focus positions separated by roughly 25&nbsp;nm; smaller features are expected to be worse.<ref>Z. Zhu et al., Proc. SPIE 5037, 494 (2003)</ref> Contact holes in the 48–100&nbsp;nm pitch range showed a 37&nbsp;nm best focus range.<ref>V. Philipsen et al., Proc. SPIE 10143, 1014310 (2017).</ref> The best focus position vs. pitch is also dependent on resist.<ref>{{cite journal|title=Investigation of the Current Resolution Limits of Advanced Extreme Ultraviolet (EUV) Resists|bibcode = 2006SPIE.6151..289N|citeseerx = 10.1.1.215.7131|last1 = Naulleau|first1 = Patrick P.|last2 = Rammeloo|first2 = Clemens|last3 = Cain|first3 = Jason P.|last4 = Dean|first4 = Kim|last5 = Denham|first5 = Paul|last6 = Goldberg|first6 = Kenneth A.|last7 = Hoef|first7 = Brian|last8 = La Fontaine|first8 = Bruno|last9 = Pawloski|first9 = Adam R.|last10 = Larson|first10 = Carl|last11 = Wallraff|first11 = Greg|editor1-first = Michael J|editor1-last = Lercel|journal = Emerging Lithographic Technologies X|year = 2006|volume = 6151|page = 289|doi = 10.1117/12.657005|s2cid = 97250792}}</ref> Critical layers often contain lines at one minimum pitch of one polarity, e.g., darkfield trenches, in one orientation, e.g., vertical, mixed with spaces of the other polarity of the other orientation. This often magnifies the best focus differences, and challenges the tip-to-tip and tip-to-line imaging.<ref>A. Erdmann et al., J. Micro/Nanolith. MEMS MOEMS 15(2), 021205 (2016).</ref>

====Reduction of pupil fill====
[[File:Screenshot 20230216 130758 PowerPoint.jpg|right|thumb|300px|Pupil rotation across slit forces use of much lower pupil fill (within the trapezoids or rectangles) for dipole illumination.]]
A consequence of SMO and shifting focus windows has been the reduction of pupil fill. In other words, the optimum illumination is necessarily an optimized overlap of the preferred illuminations for the various patterns that need to be considered. This leads to lower pupil fill providing better results. However, throughput is affected below 20% pupil fill due to absorption.<ref>J. Finders et al., Proc. SPIE 9776, 97761P (2016).</ref><ref>D. Rio et al, Proc. SPIE 10809, 108090N (2018).</ref><ref name=auto1/>

===Phase shift masks===
[[File:EUV attPSM near field phase.png|thumb|right|250px|'''Phase profile of attenuated phase shift mask for EUV.''' The phase profile (red) for an attenuated phase shift mask used with a partially transmitting EUV absorber does not match the ideal profile design (dotted), due to oblique incidence illumination and absorber edge scattering.]]
A commonly touted advantage of EUV has been the relative ease of lithography, as indicated by the ratio of feature size to the wavelength multiplied by the numerical aperture, also known as the k1 ratio. An 18&nbsp;nm metal linewidth has a k1 of 0.44 for 13.5&nbsp;nm wavelength, 0.33 NA, for example. For the k1 approaching 0.5, some weak resolution enhancement including [[Phase-shift mask|attenuated phase shift masks]] has been used as essential to production with the ArF laser wavelength (193&nbsp;nm),<ref>C-H. Chang et al., Proc. SPIE 5377, 902 (2004).</ref><ref>T. Devoivre et al., MTDT 2002.</ref><ref>L. C. Choo et al., Proc. SPIE vol. 4000, 1193 (2000).</ref><ref>J. Word and K. Sakajiri, Proc. SPIE 6156, 61561I (2006).</ref><ref>T. Winkler et al., Prod. SPIE 5754, 1169 (2004).</ref><ref>Y. Borodovsky et al., Proc. SPIE 4754, 1 (2002).</ref> whereas this resolution enhancement is not available for EUV.<ref>S-S. Yu et al., Proc. SPIE 8679, 86791L (2013).</ref><ref name=3dattpsm>A. Erdmann et al., Proc. SPIE 10583, 1058312 (2018).</ref><ref>{{Cite web|url=https://www.linkedin.com/pulse/phase-shifting-masks-nils-improvement-handicap-euv-frederick-chen|title=Phase-Shifting Masks for NILS Improvement - A Handicap For EUV?|website=www.linkedin.com}}</ref> In particular, 3D mask effects including scattering at the absorber edges distort the desired phase profile.<ref name=3dattpsm/> Also, the phase profile is effectively derived from the plane wave spectrum reflected from the multilayer through the absorber rather than the incident plane wave.<ref>{{Cite web|url=https://www.euvlitho.com/2016/P51.pdf|title=Eigenmode analysis of EM fields in EUV masks}}</ref> Without absorbers, near-field distortion also occurs at an etched multilayer sidewall due to the oblique incidence illumination;<ref>{{Cite web|url=https://www.euvlitho.com/2017/P37.pdf|title=Ultra-high efficiency EUV etched phase-shift mask}}</ref> some light traverses only a limited number of bilayers near the sidewall.<ref name=euvlmask/> Additionally, the different polarizations (TE and TM) have different phase shifts.<ref name=euvlmask/>.Fundamentally, a chromeless phase shift mask enables pitch splitting by suppression of the zeroth diffracted order on the mask, but fabricating a high quality phase shift mask for EUV is certainly not a trivial task. One possible way to achieve this is through spatial filtering at the Fourier plane of the mask pattern. At Lawrence Berkeley National Lab, the light of the zeroth order is a centrally obscured system, and the +/-1 diffracted orders will be captured by the clear aperture, providing a functional equivalent to the chromeless phase shift mask while using a conventional binary amplitude mask.<ref>Naulleau, P., Anderson, C.N., Baclea-an, L.M., Chan, D., Denham, P., George, S., Goldberg, K.A., Hoef, B., Jones, G., Koh, C. and La Fontaine, B., 2010, March. The SEMATECH Berkeley MET pushing EUV development beyond 22nm half pitch. In Extreme Ultraviolet (EUV) Lithography (Vol. 7636, pp. 530-538). SPIE.</ref>

==EUV photoresist exposure: the role of electrons==
EUV light generates photoelectrons upon absorption by matter. These photoelectrons in turn generate secondary electrons, which slow down before engaging in chemical reactions.<ref>{{Cite web|url=https://www.jstage.jst.go.jp/article/photopolymer/26/5/26_625/_pdf|title=. Torok et al., "Secondary Electrons in EUV Lithography", J. Photopol. Sci. and Tech., 26, 625 (2013).}}</ref> At sufficient doses 40&nbsp;eV electrons are known to penetrate 180&nbsp;nm thick resist leading to development.<ref>K. Ishii and T. Matsuda, Jpn. J. Appl. Phys. 29, 2212 (1990).</ref> At a dose of 160&nbsp;μC/cm<sup>2</sup>, corresponding to 15&nbsp;mJ/cm<sup>2</sup> EUV dose assuming one electron/photon, 30&nbsp;eV electrons removed 7&nbsp;nm of PMMA resist after standard development.<ref>A. Thete et al., Proc. SPIE 9422, 94220A (2015).</ref> For a higher 30&nbsp;eV dose of 380&nbsp;μC/cm<sup>2</sup>, equivalent to 36&nbsp;mJ/cm<sup>2</sup> at one electron/photon, 10.4&nbsp;nm of PMMA resist are removed.<ref>{{Cite web|url=https://www.rug.nl/research/zernike/education/topmasternanoscience/programme-documents/ns200thesis/ns200_2014_sun_b.pdf|title=B. Sun thesis, p. 34}}</ref> These indicate the distances the electrons can travel in resist, regardless of direction.<ref>{{cite web|url=https://www2.eecs.berkeley.edu/Pubs/TechRpts/2017/EECS-2017-119.pdf|title=S. Bhattarai, Study of Line Edge Roughness and Interactions of Secondary Electrons in Photoresists for EUV Lithography, 2017, p. 100.|access-date=2018-09-16|archive-url=https://web.archive.org/web/20171021154706/https://www2.eecs.berkeley.edu/Pubs/TechRpts/2017/EECS-2017-119.pdf|archive-date=2017-10-21}}</ref>

The degree of photoelectron emission from the layer underlying the EUV photoresist has been shown to affect the depth of focus.<ref>D. D. Simone et al., Proc. SPIE 10143, 101430R (2017).</ref> Unfortunately, hardmask layers tend to increase photoelectron emission, degrading the depth of focus. Electrons from defocused images in the resist can also affect the best focus image.<ref>{{Cite web|url=https://www.youtube.com/watch?v=fe5F5MuIAhc|title=Defocus Impact on Electron Blur in EUV Lithography|date=4 March 2023 |via=www.youtube.com}}</ref>

The randomness of the number of secondary electrons is itself a source of stochastic behavior in EUV resist images. The scale length of electron blur itself has a distribution.<ref>M.I.Jacobs et al., Phys. Chem. Chem. Phys. 19(20) (2017).</ref> Intel demonstrated with a rigorous simulation that EUV-released electrons scatter distances larger than 15 nm in EUV resists.<ref>P. Theofanis et al., Proc. SPIE 11323, 113230I (2020).</ref>

The electron blur is also affected by total internal reflection from the top surface of the resist film.<ref>{{Cite web|url=https://www.linkedin.com/pulse/electron-blur-impact-euv-resist-films-from-interface-reflection-chen|title=Electron Blur Impact in EUV Resist Films from Interface Reflection|website=www.linkedin.com}}</ref><ref> O. Yu et al., J. Elec. Spec. and Rel. Phenom. 241, 146824 (2020).</ref>

A more realistic description of the electron blur uses the difference of two exponential functions.<ref>[https://frederickchen.substack.com/p/a-realistic-electron-blur-function A Realistic Electron Blur Function Shape for EUV Resist Modeling]</ref>

=== Effect of underlying layers ===
[[File:Secondary Electrons from Layers under EUV Resist.png|thumb|right|300px|Electrons from layers under the resist can affect the profile and onset of collapse.]]
Secondary electrons from layers underneath the resist can affect the resist profile as well as pattern collapse.<ref>N. Miyahara et al., Proc. SPIE 12498, 124981E (2023)</ref> Hence, selection of such both the underlayer and the layer under that layer are important considerations for EUV lithography. Moreover, the electrons from defocused images can aggravate the stochastic nature of the image.<ref>{{Cite web|url=https://www.youtube.com/watch?v=AJUKJdFI1So|title=Defocus Aggravates Stochastic EUV Images|date=December 30, 2023|via=YouTube}}</ref>

==Contamination effects==
===Resist outgassing===
[[File:Outgassing contamination vs dose.png|right|thumb|300px|'''Outgassing contamination vs. EUV dose:''' The increase of dose to size (E<sub>size</sub>) to reduce [[shot noise]] and roughness comes at price of increased contamination from [[outgassing]]. The contamination thickness shown here is relative to a reference resist.]]
Due to the high efficiency of absorption of EUV by photoresists, heating and outgassing become primary concerns. One well-known issue is contamination deposition on the resist from ambient or outgassed hydrocarbons, which results from EUV- or electron-driven reactions.<ref>J. Hollenshead and L. Klebanoff, ''J. Vac. Sci. & Tech. B'' 24, pp. 118–130 (2006).</ref> Organic photoresists outgas hydrocarbons<ref>G. Denbeaux ''et al.'', 2007 European Mask and Lithography Conference.</ref> while metal oxide photoresists outgas water and oxygen<ref>I. Pollentier ''et al.'', Proc. SPIE vol. 7972, 797208 (2011).</ref> and metal (in a hydrogen ambient); the last is uncleanable.<ref name=sn_contam/> The carbon contamination is known to affect multilayer reflectivity<ref>G. Denbeaux, 2009 Intl. Workshop on EUV Lithography.</ref> while the oxygen is particularly harmful for the ruthenium capping layers (relatively stable under EUV and hydrogen conditions) on the EUV multilayer optics.<ref>J. Y. Park ''et al.'', J. Vac. Sci. Tech. B29, 041602 (2011).</ref>

===Tin redeposition===
Atomic [[hydrogen]] in the tool chambers is used to clean [[tin]] and [[carbon]] which deposit on the EUV optical surfaces.<ref name="sn_clean">{{Cite web |last=Crijns |first=V. M. C. |year=2014 |title=Hydrogen atom based tin cleaning |url=https://pure.tue.nl/ws/portalfiles/portal/46963189/774678-1.pdf |website=Eindhoven University of Technology}}</ref> Atomic hydrogen is produced by EUV light directly photoionizing H<sub>2</sub>:<ref name=tvdeven>T. Van de Ven et al., J. Appl. Phys. 123, 063301 (2018).</ref>
: hν + H<sub>2</sub> → H<sup>+</sup> + H + e<sup>−</sup>.
Electrons generated in the above reaction may also dissociate H<sub>2</sub> to form atomic hydrogen:<ref name=tvdeven/>
: e<sup>−</sup> + H<sub>2</sub> → H<sup>+</sup> + H + 2e<sup>−</sup>.
The reaction with [[tin]] in the light source (e.g., tin on an optical surface in the source) to form volatile SnH<sub>4</sub> ([[stannane]]) that can be pumped out from the source proceeds via the reaction<ref name=sn_clean/>
: Sn(s) + 4 H(g) → SnH<sub>4</sub>(g).
The SnH<sub>4</sub> can reach the coatings of other EUV optical surfaces, where it redeposits Sn via the reaction<ref name=sn_clean/>
: SnH<sub>4</sub> → Sn(s) + 2 H<sub>2</sub>(g).
Redeposition may also occur by other intermediate reactions.<ref>[https://www.euvlitho.com/2018/S47.pdf Computer modeling of contamination and cleaning of EUV source optics]. RnD-ISAN/EUV Labs & ISTEQ BV.</ref>

The redeposited Sn<ref name=snh4>{{cite journal |url=http://www.physics.rutgers.edu/~faradjev/pdf/F46.pdf |author1=Nadir Faradzhev |author2=Vadim Sidorkin |title=Hydrogen mediated transport of Sn to Ru film surface |journal=J. Vac. Sci. Technol. A |volume=27 |issue=2 |year=2009 |pages=306–314 |doi=10.1116/1.3081968 |bibcode=2009JVSTA..27..306F |access-date=2016-12-14 |archive-url=https://web.archive.org/web/20161220125121/http://www.physics.rutgers.edu/~faradjev/pdf/F46.pdf |archive-date=2016-12-20}}</ref><ref name=sn_contam>{{cite web |url=http://ieuvi.org/TWG/Resist/2016/20160221Meeting/07_EIDEC_Shiobara.pdf |title=Update of Resist Outgas Testing at EIDEC |author=Eishi Shiobara |date=2016-02-16 |work=IEUVI Resist TWG, San Jose}}</ref> might be subsequently removed by atomic-hydrogen exposure. However, overall, the tin cleaning efficiency (the ratio of the removed tin flux from a tin sample to the atomic-hydrogen flux to the tin sample) is less than 0.01%, due to both redeposition and hydrogen desorption, leading to formation of hydrogen molecules at the expense of atomic hydrogen.<ref name=sn_clean/> The tin cleaning efficiency for tin oxide is found roughly twice higher than that of tin (with a native oxide layer of ~2&nbsp;nm on it).<ref name="sn_clean" /> Injecting a small amount of oxygen to the light source may improve the tin cleaning rate.

===Hydrogen blistering===
[[File:Atomic_hydrogen_in_multilayer.png|thumb|right|300px|Hydrogen-induced blistering defects. Atomic hydrogen (red dots) used for cleaning surfaces can penetrate underneath the surface. In the Mo/Si multilayers, H<sub>2</sub> (paired red dots) is formed and trapped, resulting in blister (white region) formation.]]
Hydrogen also reacts with metal-containing compounds to reduce them to metal,<ref>[http://eprints.gla.ac.uk/84097/2/84097.pdf The Denitridation of Nitrides Under Hydrogen].</ref> and diffuses through the silicon<ref>{{Cite web|url=http://mrlweb.mrl.ucsb.edu/~vandewalle/publications/MRS557,255(1999)-aSiH.pdf|title=C. G. van de Walle and B. Tuttle, THEORY OF HYDROGEN INTERACTIONS WITH AMORPHOUS SILICON in ''Amorphous and Heterogeneous Silicon Thin Films — Fundamentals to Devices'', edited by H. M. Branz, R. W. Collins, H. Okamoto, S. Guha, and B. Schropp, MRS Symposia Proceedings, Vol. 557 (MRS, Pittsburgh, Pennsylvania, 1999), p. 255.}}</ref> and molybdenum<ref>T. Tanabe, Y. Yamanishi, and S. Imoto, J. Nucl. Mat. 191–194, 439 (1992).</ref> in the multilayer, eventually causing blistering.<ref name=bliseuvm>{{Cite web|url=https://www.youtube.com/watch?v=FZxzwhBR5Bk|title=Hydrogen Blistering in EUV Multilayers|date=6 October 2022 |via=www.youtube.com}}</ref><ref name=capclean>{{Cite web|url=http://cpmi.illinois.edu/files/2016/03/In-Situ-Collector-Cleaning-and-EUV-Reflectivity-Restoration-by-Hydrogen-Plasma-for-EUV-Sources.pdf|title=D. T. Elg ''et al.'', J. Vac. Sci. Tech. A 34, 021305 (2016).}}</ref><ref>{{Cite web|url=https://www.differ.nl/sites/default/files/attachments/biblio/2014_56992.pdf|title=Hydrogen-induced blistering in thin film multilayers}}</ref> Capping layers that mitigate hydrogen-related damage often reduce reflectivity to well below 70%.<ref name=capclean/> Capping layers are known to be permeable to ambient gases including oxygen<ref>I-Y. Jang et al., Proc. SPIE 9256, 92560I (2014)</ref> and hydrogen,<ref>{{Cite web|url=http://www.msrjournal.com/article_13532_0f1457fd762850932cb8d044b60d3571.pdf|title=Hydrogen penetration of Ru and Pd/Ru}}</ref><ref>{{cite journal|url=https://www.researchgate.net/publication/253761381|title=Ruthenium gate electrodes on SiO2 and HfO2: Sensitivity to hydrogen and oxygen ambients|first1=L|last1=Pantisano|first2=Tom|last2=Schram|first3=Z|last3=Li|first4=Judit|last4=Lisoni|first5=Geoffrey|last5=Pourtois|first6=Stefan|last6=De Gendt|first7=D|last7=P. Brunco|first8=A|last8=Akheyar|first9=V.V.|last9=Afanas'ev|first10=Sheron|last10=Shamuilia|first11=A|last11=Stesmans|date=12 June 2006|journal=Applied Physics Letters|volume=88|issue=24|page=243514|via=ResearchGate|doi=10.1063/1.2212288|bibcode=2006ApPhL..88x3514P}}</ref><ref>{{Cite web|url=http://pubman.mpdl.mpg.de/pubman/item/escidoc:2143692:1/component/escidoc:2143691/alimov.doc|title=Hydrogen penetration of boron carbide}}</ref><ref>M. Mayer, M. Balden, and R. Behrisch, J. Nucl. Mat. 252, 55 (1998).</ref> as well as susceptible to the hydrogen-induced blistering defects.<ref name=bd>S-S. Kim et al., Proc. SPIE 10143, 1014306 (2017).</ref><ref name=bliseuvm/> Hydrogen may also react with the capping layer, resulting in its removal.<ref>{{Cite web|url=http://euvlsymposium.lbl.gov/pdf/2004/presentations/day1/Co07_Sasa_Bajt.pdf|title=Screening of oxidation-resistance capping layers}}</ref> TSMC proposed some means for mitigating hydrogen blistering defects on EUV masks, which may impact productivity.<ref name=tsmc>{{Cite web|url=https://frederickchen.substack.com/p/tsmc-confronts-mask-defects-from|title=TSMC Confronts Mask Defects from EUV Hydrogen Plasmas|first=Frederick|last=Chen|date=December 6, 2024}}</ref>

=== Tin spitting ===
Hydrogen can penetrate molten tin (Sn), creating hydrogen bubbles inside it. If the bubbles move at the molten tin surface, then it bursts with tin, resulting in tin spreading over a large angle range. This phenomenon is called tin spitting and is one of EUV Collector contamination sources.

===Resist erosion===
Hydrogen also reacts with resists to etch<ref>B. Thedjoisworo ''et al.'', J. Vac. Sci. Tech. A 30, 031303 (2012).</ref><ref>{{Cite web|url=http://www.jsts.org/html/journal/journal_files/2013/08/year2013volume13_04_13.pdf|title=Hydrogen plasma for photoresist stripping|access-date=2019-01-06|archive-date=2020-03-21|archive-url=https://web.archive.org/web/20200321034016/http://jsts.org/html/journal/journal_files/2013/08/Year2013Volume13_04_13.pdf}}</ref> or decompose<ref>{{Cite web|url=http://ieuvi.org/TWG/Resist/2016/20160221Meeting/07_EIDEC_Shiobara.pdf|title=ieuvi.org|website=ieuvi.org}}</ref> them. Besides photoresist, hydrogen plasmas can also etch silicon, albeit very slowly.<ref name="ThedjoisworoCheung2013">{{cite journal|last1=Thedjoisworo|first1=Bayu|last2=Cheung|first2=David|last3=Crist|first3=Vince|title=Comparison of the effects of downstream H2- and O2-based plasmas on the removal of photoresist, silicon, and silicon nitride|journal=Journal of Vacuum Science & Technology B, Nanotechnology and Microelectronics: Materials, Processing, Measurement, and Phenomena|volume=31|issue=2|year=2013|page=021206|issn=2166-2746|doi=10.1116/1.4792254|bibcode=2013JVSTB..31b1206T|doi-access=free}}</ref>{{Primary source inline|date=January 2019}}

===Membrane===
To help mitigate the above effects, the latest EUV tool introduced in 2017, the NXE:3400B, features a membrane that separates the wafer from the projection optics of the tool, protecting the latter from outgassing from the resist on the wafer.<ref name="auto1"/> The membrane contains layers which absorb DUV and IR radiation, and transmits 85–90% of the incident EUV radiation. There is of course, accumulated contamination from wafer outgassing as well as particles in general (although the latter are out of focus, they may still obstruct light).

==EUV-induced plasma==
[[File:Spread_of_electrons_outside_EUV_exposure_area.png|left|thumb|Electron charging from the EUV-induced plasma occurs even outside the EUV exposure area (purple borders)]]
EUV lithographic systems using EUV light operate in 1–10 Pa hydrogen background gas.<ref>J. Beckers et al., Appl. Sci. 9,2827 (2019).</ref> The plasma is a source of VUV radiation<ref>P. De Schepper et al., J. Micro/Nanolith. MEMS MOEMS 13, 023006 (2014).</ref> as well as electrons and hydrogen ions<ref name=plasmadamage>E-S. Choe et al., Adv. Mater. Interfaces 2023, 2300867.</ref> This plasma is known to etch exposed materials.<ref name=plasmadamage/><ref>P. De Schepper et al., Proc. SPIE 9428, 94280C (2015).</ref>

In 2023, a study supported at TSMC was published which indicated net charging by electrons from the plasma as well as from electron emission.<ref>{{cite journal | pmc=9950305 | date=2023 | last1=Huang | first1=Y. H. | last2=Lin | first2=C. J. | last3=King | first3=Y. C. | title=A study of hydrogen plasma-induced charging effect in EUV lithography systems | journal=Discover Nano | volume=18 | issue=1 | page=22 | doi=10.1186/s11671-023-03799-4 | doi-access=free | pmid=36823307 | bibcode=2023NRL....18...22H }}</ref> The charging was found to occur even outside the EUV exposure area, indicating that the surrounding area had been exposed to electrons.

Due to chemical sputtering of carbon by the hydrogen plasma,<ref>{{Cite journal|title=EUV-induced hydrogen plasma and particle release|first1=Mark|last1=van de Kerkhof|first2=Andrei M.|last2=Yakunin|first3=Vladimir|last3=Kvon|first4=Andrey|last4=Nikipelov|first5=Dmitry|last5=Astakhov|first6=Pavel|last6=Krainov|first7=Vadim|last7=Banine|date=June 3, 2022|journal=Radiation Effects and Defects in Solids|volume=177|issue=5–6|pages=486–512|doi=10.1080/10420150.2022.2048657|doi-access=free|bibcode=2022REDS..177..486V }}</ref> there can be generation of nanoparticles,<ref>K. Bystrov et al., J. Vac. Sci. Tech. A 31, 011303 (2013).</ref> which can obstruct the EUV resist exposure.<ref>{{Cite web|url=https://frederickchen.substack.com/p/nanoparticles-in-the-euv-induced|title=Nanoparticles in the EUV-Induced Plasma: Another Possible Origin for Stochastic Defects in EUV Lithography|first=Frederick|last=Chen|date=July 11, 2024}}</ref><ref>{{Cite web|url=https://www.linkedin.com/pulse/nanoparticles-euv-induced-plasma-another-possible-origin-chen-z4w3c/|title=Nanoparticles in the EUV-Induced Plasma: Another Possible Origin for Stochastic Defects|website=www.linkedin.com}}</ref>

==Mask defects==
[[File:EUVL_printable_defects.png|right|thumb|300px|EUV defect printability vs. pitch. The printability (here 10% CD) of a defect of a given height and width varies with pitch. Note that even the surface roughness on the multilayer here can have noticeable impact.]]
Reducing defects on extreme ultraviolet (EUV) masks is currently one of the most critical issues to be addressed for commercialization of EUV lithography.<ref>{{cite web|url=http://spie.org/x48080.xml?pf=true&ArticleID=x48080|title=Getting up to speed with roadmap requirements for extreme-UV lithography|work=spie.org}}</ref> Defects can be buried underneath or within the multilayer stack<ref>{{cite web|url=http://www.eecs.berkeley.edu/Pubs/TechRpts/2005/EECS-2005-28.html|title=Fast Simulation Methods for Non-Planar Phase and Multilayer Defects in DUV and EUV Photomasks for Lithography|work=berkeley.edu}}</ref> or be on top of the multilayer stack. Mesas or protrusions form on the sputtering targets used for multilayer deposition, which may fall off as particles during the multilayer deposition.<ref>H. Yu ''et al.'', J. Vac. Sci. Tech. A31, 021403 (2013).</ref> In fact, defects of atomic scale height (0.3–0.5&nbsp;nm) with 100&nbsp;nm [[Full width at half maximum|FWHM]] can still be printable by exhibiting 10% CD impact.<ref>S. Huh ''et al.'', ''Proc. SPIE'' 7271 (2009).</ref> IBM and Toppan reported at Photomask Japan 2015 that smaller defects, e.g., 50&nbsp;nm size, can have 10% CD impact even with 0.6&nbsp;nm height, yet remain undetectable.<ref>K. Seki ''et al.'', Proc. SPIE 9658, 96580G (2015).</ref>

Furthermore, the edge of a phase defect will further reduce [http://henke.lbl.gov/optical_constants/multi2.html reflectivity] by more than 10% if its deviation from flatness exceeds 3 degrees, due to the deviation from the target angle of incidence of 84 degrees with respect to the surface. Even if the defect height is shallow, the edge still deforms the overlying multilayer, producing an extended region where the multilayer is sloped. The more abrupt the deformation, the narrower the defect edge extension, the greater the loss in reflectivity.

EUV mask defect repair is also more complicated due to the across-slit illumination variation mentioned above. Due to the varying shadowing sensitivity across the slit, the repair deposition height must be controlled very carefully, being different at different positions across the EUV mask illumination slit.<ref>A. Garetto ''et al.'', J. Micro/Nanolith. MEMS MOEMS 13, 043006 (2014).</ref>

===Multilayer reflectivity random variations===
GlobalFoundries and Lawrence Berkeley Labs carried out a Monte Carlo study to simulate the effects of intermixing between the molybdenum (Mo) and silicon (Si) layers in the multilayer that is used to reflect EUV light from the EUV mask.<ref name=ychen2017>Y. Chen et al., Proc. SPIE 10143, 101431S (2017).</ref> The results indicated high sensitivity to the atomic-scale variations of layer thickness. Such variations could not be detected by wide-area reflectivity measurements but would be significant on the scale of the critical dimension (CD).<ref name=ychen2017/> The local variation of reflectivity could be on the order of 10% for a few nm standard deviation.<ref>R. Jonckheere and L. S. Melvin III, Proc. SPIE 11517, 1151710 (2020).</ref>

===Multilayer damage===
Multiple EUV pulses at less than 10&nbsp;mJ/cm<sup>2</sup> could accumulate damage to a Ru-capped Mo/Si multilayer mirror optic element.<ref>M. Muller et al., Appl. Phys. A vol. 108, 263 (2012).</ref> The angle of incidence was 16° or 0.28 rads, which is within the range of angles for a 0.33 NA optical system.

===Pellicles===
Production EUV tools need a pellicle to protect the mask from contamination. Pellicles are normally expected to protect the mask from particles during transport, entry into or exit from the exposure chamber, as well as the exposure itself. Without pellicles, particle adders would reduce yield, which has not been an issue for conventional optical lithography with 193&nbsp;nm light and pellicles. However, for EUV, the feasibility of pellicle use is severely challenged, due to the required thinness of the shielding films to prevent excessive EUV absorption. Particle contamination would be prohibitive if pellicles were not stable above 200&nbsp;W, i.e., the targeted power for manufacturing.<ref>{{Cite web|url=http://euvlitho.com/2015/P23.pdf|title=EUVL activities in South Korea (including Samsung and SKHynix)}}</ref>

Heating of the EUV mask [[pellicle mirror|pellicle]] (film temperature up to 750&nbsp;K for 80&nbsp;W incident power) is a significant concern, due to the resulting deformation and transmission decrease.<ref>I-S. Kim ''et al.'', Proc. SPIE vol. 8322, 83222X (2012).</ref> [[ASML Holding|ASML]] developed a 70&nbsp;nm thick polysilicon pellicle membrane, which allows EUV transmission of 82%; however, less than half of the membranes survived expected EUV power levels.<ref>C. Zoldeski ''et al.'', Proc. SPIE vol. 9048, 90481N (2014).</ref> SiNx pellicle membranes also failed at 82&nbsp;W equivalent EUV source power levels.<ref>D. L. Goldfarb, Dec. 2015 BACUS Newsletter.</ref> At target 250&nbsp;W levels, the pellicle is expected to reach 686 degrees Celsius,<ref>{{cite web|url=https://semiengineering.com/euv-pellicle-uptime-and-resist-issues-continue/|title=EUV Pellicle, Uptime And Resist Issues Continue|date=26 September 2018 }}</ref> well over the melting point of aluminum. Alternative materials need to allow sufficient transmission as well as maintain mechanical and thermal stability. However, graphite, graphene or other carbon nanomaterials (nanosheets, nanotubes) are damaged by EUV due to the release of electrons<ref>A. Gao ''et al.'', J. Appl. Phys. 114, 044313 (2013).</ref> and also too easily etched in the hydrogen cleaning plasma expected to be deployed in EUV scanners.<ref>E. Gallagher ''et al.'', Proc. SPIE vol. 9635, 96350X (2015).</ref> Hydrogen plasmas can also etch silicon as well.<ref>C. Ghica ''et al.'', Rom. Rep. in Phys., vol. 62, 329–340 (2010).</ref><ref>L. Juan ''et al.'', Chin. Phys. B, vol., 22, 105101 (2013).</ref> A coating helps improve hydrogen resistance, but this reduces transmission and/or emissivity, and may also affect mechanical stability (e.g., bulging).<ref>I. Pollentier ''et al.'', Proc. SPIE vol. 10143, 101430L (2017).</ref>

Wrinkles on pellicles can cause CD nonuniformity due to uneven absorption; this is worse for smaller wrinkles and more coherent illumination, i.e., lower pupil fill.<ref>I-S. Kim et al., Microel. Eng. 177, 35 (2017).</ref>

In the absence of pellicles, EUV mask cleanliness would have to be checked before actual product wafers are exposed, using wafers specially prepared for defect inspection.<ref name=challenges>H. J. Levinson and T. A. Brunner, Proc. SPIE 10809, 1080903 (2018).</ref> These wafers are inspected after printing for repeating defects indicating a dirty mask; if any are found, the mask must be cleaned and another set of inspection wafers are exposed, repeating the flow until the mask is clean. Any affected product wafers must be reworked.

TSMC reported starting limited use of its own pellicle in 2019 and continuing to expand afterwards,<ref>{{cite web| url = https://www.anandtech.com/show/16732/tsmc-manufacturing-update| title = TSMC Manufacturing Update: N6 to Match N7 Output by EOY, N5 Ramping Faster, Better Yields Than N7}}</ref> and Samsung is planning pellicle introduction in 2022.<ref>{{cite web| url = https://english.etnews.com/20211013200002| title = Samsung to develop 'Pellicle', an essential EUV process product - ETNews| date = 13 October 2021}}</ref>

===Hydrogen bulging defects===
As discussed above, with regard to contamination removal, hydrogen used in recent EUV systems can penetrate into the EUV mask layers. TSMC indicated in its patent that hydrogen would enter from the mask edge.<ref name=US12025922>{{Cite web|url=https://patents.google.com/patent/US12025922B2|title=Methods and apparatus for reducing hydrogen permeation from lithographic tool}}</ref> Once trapped, bulge defects or blisters were produced,<ref name=bd/> which could lead to film peeling.<ref name=US12025922/> These are essentially the blister defects which arise after a sufficient number of EUV mask exposures in the hydrogen environment. TSMC proposed some means for mitigating hydrogen blistering defects on EUV masks, which may impact productivity.<ref name=tsmc/>

==EUV stochastic issues==
[[File:64 nm pitch EUV shot noise.png|thumb|right|300px|[[Shot noise]] causing significant CD variations]]
EUV lithography is particularly sensitive to stochastic effects.<ref name=jm32017>P. De Bisschop, "Stochastic effects in EUV lithography: random, local CD variability, and printing failures", J. Micro/Nanolith. MEMS MOEMS 16(4), 041013 (2017).</ref><ref>{{Cite web|url=https://www.youtube.com/watch?v=sP3NaLXtxEY|title=Visualizing EUV Stochastics for a 14nm DRAM Example|date=February 6, 2022|via=YouTube}}</ref> In a large population of features printed by EUV, although the overwhelming majority are resolved, some suffer complete failure to print, e.g. missing holes or bridging lines. A known significant contribution to this effect is the dose used to print.<ref name=105831K>P. De Bisschop and E. Hendrickx, Proc. SPIE 10583, 105831K (2018).</ref> This is related to [[shot noise]], to be discussed further below. Due to the stochastic variations in arriving photon numbers, some areas designated to print actually fail to reach the threshold to print, leaving unexposed defect regions. Some areas may be overexposed, leading to excessive resist loss or crosslinking. The probability of stochastic failure increases exponentially as feature size decreases, and for the same feature size, increasing distance between features also significantly increases the probability.<ref name=105831K/>  Line cuts which are misshapen are a significant issue due to potential arcing and shorting.<ref>{{Cite web|url=https://www.youtube.com/watch?v=ZnSNgB5b8B4|title=EUV Stochastic Variability in Line Cuts|date=March 13, 2022|via=YouTube}}</ref> Yield requires detection of stochastic failures down to below 1e-12.<ref name=105831K/>

The tendency to stochastic defects is worse from defocus over a large pupil fill.<ref>{{Cite web|url=https://www.linkedin.com/pulse/stochastic-impact-defocus-euv-lithography-frederick-chen|title=The Stochastic Impact of Defocus in EUV Lithography|website=www.linkedin.com}}</ref><ref>{{Cite web|url=https://semiwiki.com/lithography/287526-the-stochastic-impact-of-defocus-in-euv-lithography/|title=The Stochastic Impact of Defocus in EUV Lithography|first=Fred|last=Chen|date=February 16, 2025|website=Semiwiki}}</ref>
[[File:EUV Stochastic Hot Spots.png|thumb|right|150px|Stochastic defects tend to occur where there is stochastic absorbed dose variation near the threshold dose. These are indicated by where the blue specks tend to collect.]]
Multiple failure modes may exist for the same population. For example, besides bridging of trenches, the lines separating the trenches may be broken.<ref name=105831K/> This can be attributed to stochastic resist loss,<ref name=jm32017/> from secondary electrons.<ref name=narasimhan>A. Narasimhan et al., Proc. SPIE 9422, 942208 (2015).</ref><ref>{{Cite journal|title=Localized and cascading secondary electron generation as causes of stochastic defects in extreme ultraviolet projection lithography|first=Hiroshi|last=Fukuda|date=February 23, 2019|journal=Journal of Micro/Nanolithography, MEMS, and MOEMS|volume=18|issue=1|page=013503|doi=10.1117/1.JMM.18.1.013503|bibcode = 2019JMM&M..18a3503F|doi-access=free}}</ref> The randomness of the number of secondary electrons is itself a source of stochastic behavior in EUV resist images.

The coexistence of stochastically underexposed and overexposed defect regions leads to a loss of dose window at a certain post-etch defect level between the low-dose and high-dose patterning cliffs.<ref>L. Meli et al., J. Micro/Nanolith. MEMS MOEMS 18, 011006 (2019).</ref> Hence, the resolution benefit from shorter wavelength is lost.

The resist underlayer also plays an important role.<ref name=105831K/> This could be due to the secondary electrons generated by the underlayer.<ref name=se>N. Felix et al., Proc. SPIE 9776, 97761O (2015).</ref> Secondary electrons may remove over 10&nbsp;nm of resist from the exposed edge.<ref name=narasimhan/><ref name=bhattarai>{{Cite web|url=https://cloudfront.escholarship.org/dist/prd/content/qt8q3089t2/qt8q3089t2.pdf|title=S. Bhattarai, PhD Thesis,"Study of Line Edge Roughness and Interactions of Secondary Electrons in Photoresists for EUV Lithography," U. Calif. Berkeley, 2017.}}</ref>

The defect level is on the order of 1K/mm<sup>2</sup>.<ref>S. Larivière et al., Proc. SPIE 10583, 105830U (2018).</ref> In 2020, Samsung reported that 5&nbsp;nm layouts had risks for process defects and had started implementing automated check and fixing.<ref>J. Kim et al., Proc. SPIE 11328, 113280I (2020).</ref>

Photon shot noise also leads to stochastic edge placement error.<ref>S. M. Kim et al., Proc. SPIE 9048, 90480A (2014).</ref> The photon shot noise is augmented to some degree by blurring factors such as secondary electrons or acids in chemically amplified resists; when significant the blur also reduces the image contrast at the edge. An edge placement error (EPE) as large as 8.8 nm was measured for a 48 nm pitch EUV-printed metal pattern.<ref>S. Das et al., Proc. SPIE 10959, 109590H (2019).</ref>


With the natural [[Poisson distribution]] due to the random arrival and absorption times of the photons,<ref name=sseuv>H-W Kim et al., Proc. SPIE 7636, 76360Q (2010).</ref><ref>S-M. Kim et al., Proc. SPIE 9422, 94220M (2015).</ref> there is an expected natural dose (photon number) variation of at least several percent 3 sigma, making the exposure process susceptible to stochastic variations. The dose variation leads to a variation of the feature edge position, effectively becoming a blur component. Unlike the hard resolution limit imposed by diffraction, shot noise imposes a softer limit, with the main guideline being the ITRS line width roughness (LWR) spec of 8% (3s) of linewidth.<ref>B. Baylav, [http://scholarworks.rit.edu/cgi/viewcontent.cgi?article=8633&context=theses "Reduction of Line Edge Roughness (LER) in Interference-Like Large Field Lithography"], PhD dissertation, p. 37, 2014.</ref> Increasing the dose will reduce the shot noise,<ref name=tsmcshot>Z-Y. Pan et al., Proc. SPIE 6924, 69241K (2008).</ref> but this also requires higher source power.

The two issues of shot noise and EUV-released electrons point out two constraining factors: 1) keeping dose high enough to reduce shot noise to tolerable levels, but also 2) avoiding too high a dose due to the increased contribution of EUV-released photoelectrons and secondary electrons to the resist exposure process, increasing the edge blur and thereby limiting the resolution. Aside from the resolution impact, higher dose also increases outgassing<ref>{{Cite web|url=http://www.euvlitho.com/2013/P42.pdf|title=2013 Nissan Chemical Industries, 2013 International Workshop on EUV Lithography}}</ref> and limits throughput, and crosslinking<ref>T. G. Oyama ''et al.'', Appl. Phys. Exp. 7, 036501 (2014).</ref> occurs at very high dose levels. For chemically amplified resists, higher dose exposure also increases line edge roughness due to acid generator decomposition.<ref>T. Kozawa, ''Jpn. J. Appl. Phys.'' 51, 06FC01 (2012).</ref>

Even with higher absorption at the same dose, EUV has a larger shot noise concern than the ArF (193&nbsp;nm) wavelength, mainly because it is applied to thinner resists.<ref name=abs>{{Cite web|url=https://www.linkedin.com/pulse/euv-resist-absorption-impact-stochastic-defects-frederick-chen|title=EUV Resist Absorption Impact on Stochastic Defects|website=www.linkedin.com}}</ref>

Due to stochastic considerations, the IRDS 2022 lithography roadmap now acknowledges increasing doses for smaller feature sizes.<ref>{{Cite web|url=https://irds.ieee.org/editions/2022/irds%E2%84%A2-2022-lithography|title=IRDS™ 2022: Lithography - IEEE IRDS™|website=irds.ieee.org}}</ref> However, an upper limit to how much dose can be increased is imposed by resist loss.<ref>{{Cite web|url=https://www.youtube.com/watch?v=-7OBnsz4dCY|title=EUV Dose Upper Limit From Resist Loss|date=January 13, 2025|via=YouTube}}</ref>

Due to resist thinning with increased dose, EUV stochastic defectivity limits will define a narrow CD or dose window.<ref>{{Cite web|url=https://www.linkedin.com/pulse/resist-loss-model-euv-stochastic-defectivity-cliffs-frederick-chen-9bv3c/|title=Resist Loss Model for the EUV Stochastic Defectivity Cliffs|website=www.linkedin.com}}</ref><ref>{{Cite web|url=https://frederickchen.substack.com/p/resist-loss-model-for-the-euv-stochastic|title=Resist Loss Model for the EUV Stochastic Defectivity Cliffs|first=Frederick|last=Chen|date=January 20, 2025}}</ref> The thinner resist at higher incident dose reduces absorption, and hence, absorbed dose.

EUV resolution will likely be compromised by stochastic effects. Stochastic defect densities have exceeded 1/cm<sup>2</sup>, at 36 nm pitch.<ref>Y. J. Choi et al., "Stochastic defect removal coating for high-performance extreme ultraviolet lithography," J. Vac. Sci. Tech. B 40, 042602 (2022).</ref><ref>Y. Li, Q. Wu, Y. Zhao, "A Simulation Study for Typical Design Rule Patterns and Stochastic Printing Failures in a 5 nm Logic Process with EUV Lithography," CSTIC 2020.</ref> In 2024, an EUV resist exposure by ASML revealed a missing+bridging 32 nm pitch contact hole defect density floor >0.25/cm<sup>2</sup> (177 defects per wafer), made worse with thinner resist.<ref>T-T. Wu et al., Proc. SPIE 12955, 129552V (2024).</ref> ASML indicated 30 nm pitch would not use direct exposure but double patterning.<ref name=rsocha>R. Socha, Proc. SPIE 11328, 113280V (2020).</ref> Intel did not use EUV for 30 nm pitch.<ref>B. Sell et al., VLSI Tech. 2022]</ref>

===Pupil fill ratio===
For pitches less than half-wavelength divided by numerical aperture, dipole illumination is necessary. This illumination fills at most a leaf-shaped area at the edge of the pupil. However, due to 3D effects in the EUV  mask,<ref>J-H. Franke et al., Proc. SPIE 11517, 1151716 (2020).</ref> smaller pitches require even smaller portions of this leaf shape. Below 20% of the pupil, the throughput and dose stability begin to suffer.<ref name=auto1/> Higher numerical aperture allows a higher pupil fill to be used for the same pitch, but depth of focus is significantly reduced.<ref>[https://www.youtube.com/watch?v=5bRyIGUD51I&t=3s The Tradeoff of EUV Numerical Aperture: Depth-of-Focus vs. Pupil Fill].</ref>

A larger pupil fill is more susceptible to stochastic fluctuations from point to point in the pupil.<ref>{{Cite web|url=https://www.linkedin.com/pulse/stochastic-pupil-fill-euv-lithography-frederick-chen-ogqpc/|title=Stochastic Pupil Fill in EUV Lithography|website=www.linkedin.com}}</ref><ref>{{Cite web|url=https://frederickchen.substack.com/p/stochastic-pupil-fill-in-euv-lithography|title=Stochastic Pupil Fill in EUV Lithography|first=Frederick|last=Chen|date=December 21, 2024}}</ref>

==Use with multiple-patterning==
EUV is anticipated to use double-patterning at around 34&nbsp;nm pitch with 0.33&nbsp;NA.<ref>W. Gap et al., Proc. SPIE 10583, 105830O (2018).</ref><ref>D. De Simone et al., Advanced Lithography 2019, 10957-21.</ref> This resolution is equivalent to '1Y' for DRAM.<ref>{{Cite web|url=https://techinsights.com/about-techinsights/overview/blog/samsung-18-nm-dram-cell-integration-qpt-and-higher-uniformed-capacitor-high-k-dielectrics/|title=Samsung 18 nm DRAM cell integration: QPT and higher uniformed capacitor high-k dielectrics|website=techinsights.com}}</ref><ref>{{Cite web|url=https://epsnews.com/2018/05/24/prices-for-drams-continue-to-climb-while-nand-flash-asps-drop/|title=Prices for DRAMs Continue to Climb while NAND Flash ASPs Drop|first=Gina|last=Roos|date=May 24, 2018}}</ref> In 2020, ASML reported that 5&nbsp;nm M0 layer (30&nbsp;nm minimum pitch) required double-patterning.<ref name=rsocha/>
In H2 2018, TSMC confirmed that its 5&nbsp;nm EUV scheme still used [[multiple patterning|multi-patterning]],<ref>{{cite web|url=https://www.semiwiki.com/forum/content/7759-top-10-highlights-tsmc-open-innovation-platform-ecosystem-forum.html|title=SemiWiki.com - Top 10 Highlights from the TSMC Open Innovation Platform Ecosystem Forum|website=www.semiwiki.com|date=7 August 2023 }}</ref> also indicating that mask count did not decrease from its 7&nbsp;nm node, which used extensive DUV multi-patterning, to its 5&nbsp;nm node, which used extensive EUV.<ref>{{cite web|url=https://www.synopsys.com/community/resources/videos/tsmc-videos/dac-2018-arm-tsmc-breakfast.html|title=DAC 2018 TSMC/Arm/Synopsys Breakfast|website=www.synopsys.com|access-date=2018-10-05|archive-date=2018-10-05|archive-url=https://web.archive.org/web/20181005194937/https://www.synopsys.com/community/resources/videos/tsmc-videos/dac-2018-arm-tsmc-breakfast.html}}</ref> EDA vendors also indicated the continued use of multi-patterning flows.<ref>{{cite press release|url=https://www.businesswire.com/news/home/20181001005260/en/Cadence-Achieves-EDA-Certification-TSMC-5nm-7nm|title=Cadence Achieves EDA Certification for TSMC 5nm and 7nm+ FinFET Process Technologies to Facilitate Mobile and HPC Design Creation|date=October 2018}}</ref><ref>{{cite web|url=https://www.design-reuse.com/news/44846/synopsys-digital-and-custom-design-platforms-tsmc-5-nm-euv-based-process.html|title=Synopsys Digital and Custom Design Platforms Certified on TSMC 5-nm EUV-based Process Technology|website=Design And Reuse}}</ref> While Samsung introduced its own 7&nbsp;nm process with EUV single-patterning,<ref>{{cite web|url=https://www.synopsys.com/community/resources/videos/samsung-videos/dac-2018-samsung-breakfast.html|title=DAC 2018 Samsung/Synopsys Breakfast|website=www.synopsys.com|access-date=2018-10-05|archive-date=2018-10-05|archive-url=https://web.archive.org/web/20181005194938/https://www.synopsys.com/community/resources/videos/samsung-videos/dac-2018-samsung-breakfast.html}}</ref> it encountered severe photon shot noise causing excessive line roughness, which required higher dose, resulting in lower throughput.<ref name=sseuv/> TSMC's 5&nbsp;nm node uses even tighter design rules.<ref>{{Cite web|url=https://www.eetimes.com/document.asp?doc_id=1333827|title=TSMC Goes Photon to Cloud|first=Rick|last=Merritt|website=EETimes}}</ref> Samsung indicated smaller dimensions would have more severe shot noise.<ref name=sseuv/>
[[File:Via Triple Patterning for EUV.png|thumb|left|100px|At 38&nbsp;nm center-to-center spacing or less, a 0.33 NA EUV tool would require double- or even triple-patterning for the contact or via layer.]]
In Intel's complementary lithography scheme at 20&nbsp;nm half-pitch, EUV would be used only in a second line-cutting exposure after a first 193&nbsp;nm line-printing exposure.<ref>{{Cite web|url=http://www.euvlitho.com/2012/P1.pdf|title=Intel presentation on Complementary Lithography at 2012 International Workshop on EUV Lithography}}</ref>

Multiple exposures would also be expected where two or more patterns in the same layer, e.g., different pitches or widths, must use different optimized source pupil shapes.<ref name="auto3">{{Cite web | url=https://semiwiki.com/forum/index.php?threads/euv-was-never-going-to-be-single-patterning.8935/ | title=EUV was never going to be single patterning| date=5 February 2017}}</ref><ref>S. Hsu et al., Proc. SPIE 4691, 476 (2002).</ref><ref>X. Liu et al., Proc. SPIE 9048, 90480Q (2014).</ref><ref>S-Y. Oh et al., Proc. SPIE 4691, 1537 (2002).</ref> For example, when considering a staggered bar array of 64&nbsp;nm vertical pitch, changing the horizontal pitch from 64&nbsp;nm to 90&nbsp;nm changes the optimized illumination significantly.<ref name=7nmt2t/> Source-mask optimization that is based on line-space gratings and tip-to-tip gratings only does not entail improvements for all parts of a logic pattern, e.g., a dense trench with a gap on one side.<ref name="auto3"/><ref>D. Rio et al., Proc. SPIE 10809, 108090N (2018).</ref>

In 2020, [[ASML Holding|ASML]] reported that for the 3&nbsp;nm node, center-to-center contact/via spacings of 40&nbsp;nm or less would require double- or triple-patterning for some contact/via arrangements.<ref>W. Gao et al., Proc. SPIE 11323, 113231L (2020).</ref>

For the 24–36&nbsp;nm metal pitch, it was found that using EUV as a (second) cutting exposure had a significantly wider process window than as a complete single exposure for the metal layer.<ref>R. K. Ali et al., Proc. SPIE 10583, 1058321 (2018).</ref><ref name="auto3"/> However, using a second exposure in the LELE approach for double patterning does not get around the vulnerability to stochastic defects.<ref>{{Cite web|url=https://frederickchen.substack.com/p/can-lele-multipatterning-help-against|title=Can LELE Multipatterning Help Against EUV Stochastics?|first=Frederick|last=Chen|date=January 1, 2025}}</ref><ref>{{Cite web|url=https://www.linkedin.com/pulse/can-lele-multipatterning-help-against-euv-stochastics-frederick-chen-abmoc/|title=Can LELE Multipatterning Help Against EUV Stochastics?|website=www.linkedin.com}}</ref>

Multiple exposures of the same mask are also expected for defect management without pellicles, limiting productivity similarly to multiple-patterning.<ref name=challenges/>

Self-aligned litho-etch-litho-etch (SALELE) is a hybrid SADP/LELE technique whose implementation has started in 7&nbsp;nm.<ref>Q. Lin, Proc. SPIE 11327, 113270X (2020).</ref> 
Self-aligned litho-etch-litho-etch (SALELE) has become an accepted form of double-patterning to be used with EUV.<ref>R. Venkatesan et al., Proc. SPIE 12292, 1229202 (2022).</ref>

==Single-patterning extension: anamorphic high-NA==
[[File:EUV High-NA stochastic sidelobes.png|thumb|right|300px|Sidelobes are a greater risk with stochastic variations, at a certain center-to-center distance. This risk is elevated due to the central obscuration in High-NA EUV systems.]]
A return to extended generations of single-patterning would be possible with higher [[numerical aperture]] (NA) tools. An NA of 0.45 could require retuning of a few percent.<ref>J. T. Neumann ''et al.'', Proc. SPIE 8522, 852211 (2012).</ref> Increasing demagnification could avoid this retuning, but the reduced field size severely affects large patterns (one die per 26&nbsp;mm × 33&nbsp;mm field) such as the many-core multi-billion transistor 14&nbsp;nm Xeon chips.<ref>[http://www.anandtech.com/show/9802/supercomputing-15-intels-knights-landing-xeon-phi-silicon-on-display Intel's Xeon E5-2600 V4 Chips Feature An Insane 7.2 Billion Transistors on a 456mm2 Die],</ref> by requiring field stitching of two mask exposures.

In 2015, [[ASML Holding|ASML]] disclosed details of its anamorphic next-generation EUV scanner, with an NA of 0.55. These machines cost around USD 360 million.<ref name="auto5"/> The demagnification is increased from 4× to 8× only in one direction (in the plane of incidence).<ref name=highNA>J. van Schoot ''et al.'', Proc. SPIE 9422, 94221F (2015).</ref> However, the 0.55 NA has a much smaller depth of focus than immersion lithography.<ref>B. J. Lin, JM3 1, 7–12 (2002).</ref> Also, an anamorphic 0.52 NA tool has been found to exhibit too much CD and placement variability for 5&nbsp;nm node single exposure and multi-patterning cutting.<ref>E. R. Hosler ''et al.'', Proc. SPIE vol. 9776, 977616 (2015).</ref>

Depth of focus<ref name=dof>B. J. Lin, J. Microlith., Microfab., Microsyst. 1, 7–12 (2002).</ref> being reduced by increasing NA is also a concern,<ref>B. J. Lin, Microelec. Eng. 143, 91–101 (2015).</ref> especially in comparison with multi-patterning exposures using 193&nbsp;nm immersion lithography:

{| class="wikitable" style="text-align:center;"
|-
! Wavelength
! Refractive index
! NA
! DOF (normalized)<ref name=dof />
|-
| 193&nbsp;nm
| 1.44
| 1.35
| 1
|-
| 13.3–13.7&nbsp;nm
| 1
| 0.33
| 1.17
|-
| 13.3–13.7&nbsp;nm
| 1
| 0.55
| 0.40
|}

High-NA EUV tools focus horizontal and vertical lines differently from low-NA systems, due to the different demagnfication for horizontal lines.<ref> I. Lee et al., J. Micro/Nanopattern. Mater. Metrol. 22, 043202 (2023).</ref><ref>{{Cite web|url=https://frederickchen.substack.com/p/high-na-euv-has-astigmatism|title=High-NA EUV Has Astigmatism|first=Frederick|last=Chen|date=November 19, 2024}}</ref>

High-NA EUV tools also suffer from obscuration, which can cause errors in the imaging of certain patterns.<ref>B. Bilski et al., Proc. SPIE 11177, 111770I (2019).</ref> 

The first high-NA tools are expected at Intel by 2025 at earliest.<ref>{{Cite web|url=https://www.anandtech.com/show/16823/intel-accelerated-offensive-process-roadmap-updates-to-10nm-7nm-4nm-3nm-20a-18a-packaging-foundry-emib-foveros|title=Intel's Process Roadmap to 2025: with 4nm, 3nm, 20A and 18A?!|first=Dr Ian|last=Cutress|website=www.anandtech.com}}</ref>

For sub-2nm nodes, high-NA EUV systems will be affected by a host of issues: throughput, new masks, polarization, thinner resists, and secondary electron blur and randomness.<ref>{{Cite web|url=https://www.linkedin.com/pulse/reality-checks-high-na-euv-1x-nm-nodes-frederick-chen|title=Reality Checks for High-NA EUV for 1.x nm Nodes|website=www.linkedin.com}}</ref> Reduced depth of focus requires resist thickness less than 30 nm, which in turn increases stochastic effects, due to reduced photon absorption.

Electron blur is estimated to be at least ~2 nm, which is enough to thwart the benefit of High-NA EUV lithography.<ref>L. F. Miguez et al., Proc. SPIE 12498, 124980E (2023).</ref><ref>R. Fallica et al., Proc. SPIE 12498, 124980J (2023).</ref>

Beyond high-NA, ASML in 2024 announced plans for the development of a hyper-NA EUV tool with an NA beyond 0.55, such as an NA of 0.75 or 0.85.<ref name="auto6">{{cite web | url=https://www.eetimes.com/asml-aims-for-hyper-na-euv-shrinking-chip-limits/ | title=ASML Aims for Hyper-NA EUV, Shrinking Chip Limits | date=12 June 2024 }}</ref><ref>biz.chosun.com/it-science/ict/2024/06/21/OTIF4YUEGZCSLNPMHY53HH34DQ/</ref> These machines could cost USD 720 million each and are expected to be available in 2030.<ref name="auto5"/> A problem with Hyper-NA is polarization of the EUV light causing a reduction in image contrast.<ref name="auto6"/><ref>{{cite book | chapter-url=https://www.spiedigitallibrary.org/conference-proceedings-of-spie/12494/1249405/Hyper-NA-EUV-lithography-an-imaging-perspective/10.1117/12.2659153.short#_=_ | doi=10.1117/12.2659153 | date=2023 | last1=Lee | first1=Inhwan | last2=Franke | first2=Joern-Holger | last3=Philipsen | first3=Vicky | last4=Ronse | first4=Kurt | last5=De Gendt | first5=Stefan | last6=Hendrickx | first6=Eric | chapter=Hyper-NA EUV lithography: An imaging perspective | editor-first1=Anna | editor-first2=Martin | editor-last1=Lio | editor-last2=Burkhardt | title=Optical and EUV Nanolithography XXXVI | volume=12494 | page=7 | bibcode=2023SPIE12494E..05L | isbn=978-1-5106-6095-3 }}</ref>

==Beyond EUV wavelength==
A much shorter wavelength (~6.7&nbsp;nm) would be beyond EUV, and is often referred to as BEUV (beyond extreme ultraviolet).<ref>{{Cite web|url=http://www.euvlitho.com/2010/P14.pdf|title=ASML presentation at 2010 International Workshop on Extreme Ultraviolet Sources}}</ref> With current technology, BEUV wavelengths would have worse shot noise effects without ensuring sufficient dose.<ref>{{cite journal|last1=Mojarad|first1=Nassir|last2=Gobrecht|first2=Jens|last3=Ekinci|first3=Yasin|title=Beyond EUV lithography: a comparative study of efficient photoresists' performance|journal=Scientific Reports|date=18 March 2015|volume=5|issue=1|page=9235|doi=10.1038/srep09235|pmid=25783209|pmc=4363827|bibcode=2015NatSR...5.9235M}}</ref> (The generally accepted 'border' of UV is 10nm below which the (soft) x-ray region begins.)

==References==
{{Reflist|30em}}

==Further reading==
* {{cite book|title=Extreme Ultraviolet Lithography |author=Banqiu Wu and Ajay Kumar |date= May 2009 |publisher=[[McGraw-Hill Professional]], Inc. |isbn=978-0-07-154918-9 |url=http://www.mhprofessional.com/product.php?isbn=0-07-154918-8}}
* {{cite journal | url=http://www.opfocus.org/index.php?topic=story&v=7&s=4 | author=Banqiu Wu and Ajay Kumar | title=Extreme Ultraviolet Lithography: Towards the Next Generation of Integrated Circuits | journal=Optics & Photonics Focus | volume=7 | issue=4 | year=2009}}
* Michael Purvis, [https://strobe.colorado.edu/wp-content/uploads/STROBE_ASML-EUV-Sources_Purvis_25-Sept-2020-1.pdf An Introduction to EUV Sources for Lithography], ASML, STROBE, 2020-09-25.
* Igor Fomenkov, [https://euvlitho.com/2019/S1.pdf EUV Source for Lithography in HVM - performance and prospects], ASML Fellow, Source workshop, Amsterdam, 2019-11-05.

==Related links==
* [https://web.archive.org/web/20110723211724/http://semimd.com/blog/2011/03/25/euv-mask-cleaning-presents-economic-challenges/ EUV presents economic challenges]
* [https://www.eetimes.com/industry-mulls-6-7-nm-wavelength-euv Industry mulls 6.7-nm wavelength EUV]
{{Authority control}}
{{Nanolith}}
{{Microtechnology}}
{{Levels of technological manipulation of matter}}

{{DEFAULTSORT:Extreme Ultraviolet Lithography}}
[[Category:Lithography (microfabrication)]]
[[Category:Extreme ultraviolet]]