EUV With Fewer Mirrors?

[@OrangeC7]

I swear, every time I watch your lithography videos it's as if I'm learning how magic works

[@clintcowan9424]

Sufficiently advanced technology is indistinguishable from magic -- said someone

[@loneIyboy15]

I mean, you literally are. We boiled air out of rocks, turned them into beautiful crystals, cut them perfectly, then shone light on them until they could think.

[@willhatch7721]

The epitome of alchemy

[@SpankyK]

Awesome, right?

[@jimurrata6785]

I love that Jon is literally steeped in the history of semi!

[@lkk12342001]

Field curvature is not the only concern. The reason why there are many EUV mirrors in ASML system is to remove high-order aberrations which stongly affect the contrast of the image. Compared to DUV systems with tens of lenses to correct aberrations, 6 mirror design in the EUV system is already a trade off between aberration correctability and illumination power.

[@samfedorka5629]

I read the Zeiss slideshow and the system they presented also corrects many higher-order aberrations, just the slide that was shown here is the first iteration of such. They go on to show other designs which correct more abberations with more elements. Personally I think it would be easier to bend the wafer slightly like disposable cameras do with the film.

[@romanowskis1at]

Light from tin plasma is not spatial and not temporal coherent and this is pita of EUV. It is laser induced light, but it is not laser light. Wavefront of this light is not flat. Rayleih criteria is also applicabe to source light properties and it can say a lot about it quality. kzbin.infoV6G8ZZmeJzc

[@lubricustheslippery5028]

With multi wavelength light you need several lenses/mirrors to correct color aberration. I don't understand why that is necessary for monochromatic light, don't say you are wrong it's probably limitations on my understanding. Normal mirrors don't work for EUV and the mirrors that do don't have that good of an reflectivity so the machine is loosing lot of light for each mirror. So with less mirrors less light is needed and that is a big deal when it's that hard to make.

[@AJMansfield1]

I suspect that using both the front and back surface of M1 is key to the design. Lithography systems have to be absurdly precise: the old adage, "if you can't make it perfect, make it adjustable," is hardly enough, the optics in these machines have to be both perfect _and_ adjustable, and sacrifice complexity to get there. But using shared optical elements twice to keep the temp/tolerance/etc variations between them tightly controlled? Could absolutely be a reason a system like this is able to get away with fewer elements -- assuming it's actually using that correlation in a way that helps attenuate rather than amplify the most problematic error terms.

[@journey8533]

Less mirrors statistically leads to less bad luck in the long run

[@stefanodadamo6809]

That's science !

[@frankstrawnation]

​@@sub-vibesStannis Baratheon, is that you?

[@2drealms196]

Fewer mirrors mean vampires could finally be employed on ASML's nightshift?

[@-danR]

Something about EUV chip machines always bothered me but it never occurred to me that I couldn't see the trees for the forest. Too many gadomb trees and they were all staring me in the face.

[@clintcowan9424]

Shorting Carl Zeiss. Thanks for the tip

[@jamesjun1038]

You’re not just some random dude on internet. You’re one of the best educators of our time, and I truly value everything you create.

[@horizon7011]

I second that! Your the reason I know anything about EUV and the mind breaking physics at play and the people who discovered them.

[@ulogy]

​@@horizon7011Yeah, if you ever want to out yourself as the resident turbonerd in an office setting, give a brief explainer of how absurdly tight the tolerances are for producing modern chips. We're literally vaporizing metal to yield UV light, only to then focus it perfectly.

[@bubaks2]

If he is some random dude on the internet, then what am i? Im nothing.

[@dalexis21]

​@@bubaks2 don't worry, you're a random dude on the internet too

[@today273]

I'm always happy to see these updates to how semiconductor technology is being developed.

[@Swit-NL]

I follow your channel for a long time. i want to compliment you on your skills and effort to take complex information and present is such a clear and structured manner. I learned a lot and those are the best channels for me.

[@rarbiart]

"more power, Igor!" (Dr Frankenstein)

[@jonahansen]

How did you get a comment on this video 10 days ago seeing as how it's only been up for 59 minutes?

[@rarbiart]

@@jonahansen that's fairly simple: by the power of being a patreon!

[@jonahansen]

@@rarbiart Right - thanks.

[@ADHJkvsNgsMBbTQe]

Another excellent piece. It’s really impressive that you are now getting the story from the experts themselves. Please keep up the great work!

[@AC-jk8wq]

Jon, you are so focused! 😃

[@aniksamiurrahman6365]

Jon is a ultra-high NA Ultra-EUV.

[@HighYield]

2:44 best moment! ;)

[@giantnanomachine]

An important factor that is easily overlooked is that less mirrors means less opportunities for image correction. Optimization of the projected image is a hugely important factor in high end nodes. So there's a balance between power (loss) ie throughput and achievable contrast+overlay quality.

[@MCasterAnd]

KZbin recommended this video to me 47 seconds after it was released. KZbin's algorithm understands my preferences thoroughly...

[@testboga5991]

Same

[@aniksamiurrahman6365]

It knows more about u than u or your nearby ones do.

[@Jeremy-Ai]

Tread cautiously. It is a good sign that you are being offered something meaningful. (It is a representation of your efforts to seek it out ) Make no mistake, Everything you do and say on anything digital has long been calculated and it is a distortion of your responsibility to act accordingly with proper agency. Dont assume your algorithm is doing you a favour by finally giving you what you seek in a timely manner. Assume you have been given the evidence that what you seek… matters. It won’t be hard to bend a brilliant person into a corner of their own interests and keep them there in ignorance of otherwise. I doubt this message will reach you. However… it isn’t intended for you.. its intended

[@der.Schtefan]

Thanks for working hard to pronounce Schwarzschild correctly. It is highly appreciated.

[@erloriel]

German is a harsh mistress.

[@klauszinser]

@@erloriel Even as German, I did not know about him: 'Karl Schwarzschild was born on 9 October 1873 in Frankfurt on Main, the eldest of six boys and one girl,[8][9] to Jewish parents. His father was active in the business community of the city, and the family had ancestors in Frankfurt from the sixteenth century onwards.[10] The family owned two fabric stores in Frankfurt. His brother Alfred became a painter.[11] The young Schwarzschild attended a Jewish primary school until 11 years of age[12] and then the Lessing-Gymnasium (secondary school). He received an all-encompassing education, including subjects like Latin, Ancient Greek, music and art, but developed a special interest in astronomy early on.[13] In fact he was something of a child prodigy, having two papers on binary orbits (celestial mechanics) published before the age of sixteen.[14] After graduation in 1890, he attended the University of Strasbourg to study astronomy. After two years he transferred to the Ludwig Maximilian University of Munich where he obtained his doctorate in 1896 for a work on Henri Poincaré's theories. From 1897, he worked as assistant at the Kuffner Observatory in Vienna. His work here concentrated on the photometry of star clusters and laid the foundations for a formula linking the intensity of the starlight, exposure time, and the resulting contrast on a photographic plate. An integral part of that theory is the Schwarzschild exponent (astrophotography). In 1899, he returned to Munich to complete his Habilitation. From 1901 until 1909, he was a professor at the prestigious Göttingen Observatory within the University of Göttingen,[15] where he had the opportunity to work with some significant figures, including David Hilbert and Hermann Minkowski. Schwarzschild became the director of the observatory. He married Else Rosenbach, a great-granddaughter of Friedrich Wöhler and daughter of a professor of surgery at Göttingen, in 1909. Later that year they moved to Potsdam, where he took up the post of director of the Astrophysical Observatory. This was then the most prestigious post available for an astronomer in Germany.[citation needed] At the outbreak of World War I in 1914, Schwarzschild volunteered for service in the German army despite being over 40 years old. He served on both the western and eastern fronts, specifically helping with ballistic calculations and rising to the rank of second lieutenant in the artillery.[8] While serving on the front in Russia in 1915, he began to suffer from pemphigus, a rare and painful autoimmune skin-disease.[16] Nevertheless, he managed to write three outstanding papers, two on the theory of relativity and one on quantum theory. His papers on relativity produced the first exact solutions to the Einstein field equations, and a minor modification of these results gives the well-known solution that now bears his name - the Schwarzschild metric.[17] In March 1916, Schwarzschild left military service because of his illness and returned to Göttingen. Two months later, on May 11, 1916, his struggle with pemphigus may have led to his death at the age of 42.[16] He rests in his family grave at the Stadtfriedhof Göttingen.'

[@pllevi]

I think that authors spend all the energy on content research (which is great) to the point where none is left for pronunciation 🙂 For the future - use wikipedia - they often use phonetic transcription for names for example: /ˈʃvaʁtsʃɪlt/ or (from one of the previous episodes) /ˈjæn tʃɒxˈrɑːlski/

[@AndyRRR0791]

Oh yeah! I'm going to be cheering on Professor Shintake and I have nothing to do with the semiconductor industry beyond using its products. This is just awesome stuff. Godspeed!

[@steveunderwood3683]

Why is that lens configuration called Schwarzschild? A Schwarzschild scheme doesn't have a hole in the middle of the main mirror. The final image is constructed inside the structure, so its used for things like photographic telescopes. What is described in the video is a Cassegrain reflector. Those have a hole in the middle of the main mirror, and a secondary reflector, just like the diagram in the video. The term Cassegrain is used for everything from radio frequencies to light which use a similar hole in the middle of the main reflector and secondary reflector arrangement. At radio frequencies you can easily use polarisation to allow the secondary reflector to be almost as big as the main one.

[@varno]

Karl Swartzchild was a physicist and astronomer, he both discovered the Schwartzchild Metric in General Reletivity and the possibility of black holes, and the Schwartzchild telescope the first telescope to use this specific optical system, he was also a pioneer in the physics of photography. So basically, same guy did a bunch of stuff, though this is definitely closer to a cassegrain system, I think people are calling it this to not evoke the Catadioptric Cassegrains which have a field collector plate and also because Schwartzchild was the first to solve a flat field curvature with two mirrors, which wasn't a problem before photography.

[@ifsey]

Yes, the original schwarzchild mirror is two concave, whereas a cassegrain has a convex hyperbola to redirect the M1 focus to a new conjugate focus. I can't tell if ASML had zero optical designers or if the 7+ mirror system was overengineered for easier alignment and manufacture of the UV mirrors.

[@the_hate_inside1085]

@@ifsey Probably more mirrors were used as band aids to compensate different problems they accounted, and to help correct the image to be as flat, and undistorted, as possible.

[@ifsey]

@the_hate_inside1085 it's certainly more degrees of freedom for controlling imaging aberrations, especially avoiding complicated aspheric or conic forms. I read up some more and it also seems that the thermal deformation from the UV beam inefficiencies also cause problems with highly tuned surfaces. Better to have more numerous loosely toleranced surfaces. I.e. the high power UV light heating up your optical elements

[@the_hate_inside1085]

@@ifsey Yeah, creating some types of big/deeply cut, complicated lens shapes is not exactly easy. Just look at what happened with the Hubble telescope..

[@atefrod680]

I met patrick during the SPIE conference in San Jose last year, he was such a friendly and humble person. Very pleasant experiences.

[@grizwoldphantasia5005]

Evolutionary, not revolutionary, but that doesn't sell headlines. You may think you hate journalists, but you don't hate them enough.

[@rarbiart]

I don't read those sensationalizing tech news anymore. Disruptive revolutions are for gullible venture capitalists.

[@apidas]

breakthrough is always welcome, stock should go up

[@ulogy]

IDK, the machine has cycles which it revolves through, so maybe they were just being too literal.

[@fnorgen]

When everything is hyped into the clouds, reality is bound to be disappointing. Though, a lot of these tech journalists are so sloppy and formulaic that you could probably replace them with an AI without anybody noticing.

[@causewaykayak]

​@@fnorgen Perhaps in some outfits, that has already happened 😂

[@catsspat]

"While cleaning his room." Boys, take that to heart.

[@diegoantoniorosariopalomin2206]

My pestering for you to cover this technology seems to have worked. I should drop at least a few bucks for the channel

[@Basil-the-Frog]

Awesome, Jon! This video shows two ways in which real research happens. First, an idea is triggered by reading something which shows a possible related area advance, followed by a year of studying the current state-of-the-art in the research literature. This is then followed by the discovery of a century-old paper which leads to a contemporary advance. Second, a scienist/engineer has an idea and before building an expensive prototype, they purchase software to do simulations in order to verify the idea and tweak the math. This video should be required watching for all engineering PhD research students and some of us past school. Thank you again for the excellent work, sir! Cheers! p.s. Hopefully someone from ASML, SMIC, Samsung, etc., or a pile of CHIPS-Act money watches the video. Something good may happen. If not, I'm sure something will when the paper gets published.

[@MarkoCloud]

Oh boy! An EUV video! 😮 What an end to a perfect Sunday! 😊

[@PKmuffdiver]

Man, one of your best videos to date! Loved it.

[@Science4Real]

The blend of technological refinement and practicality in Shintake’s design is truly outstanding Clearly this is a bold direction that enhances the efficiency of EUV lithography machines without sacrificing industrial production compatibility

[@Tight_Conduct]

This is exciting, hoping to hear more on this soon!

[@brandonb6164]

This is way better than my idea of using relatively small gnomes to catch and re-emit the EUV photons selectively, thus acting as the mask and collimating optics array.

[@toxiccan175]

But doing so would create many jobs, thus bolstering the gnome economy. How will they compensate the gnomes being put out of theoretical jobs with this fewer-mirror EUV design?

[@marcelo55869]

Don't use gnomes, use Maxell's demons...

[@EG90arg]

THANKS ASIANGOD !!!!

[@lexeindhoven]

Great video, with a good explanation with the basics. Greatings from the Netherlands

[@peterparsons7141]

Excellent presentation filled with relevant information. Your videos have been a tremendous help for me to understand state of the art manufacturing, and lots of other interesting things. I really appreciate your work, thank you.

[@coladict]

Yay for @HighYield shout out!

[@alibizzle2010]

I'm starting to think this whole channel is some really clever AI company marketing scheme. I struggle to believe that one person can release so many high quality videos so regularly without the help of some really good drugs lol.

[@JosephCornishV]

Weed dreams, my dude.

[@100c0c]

He has a team

[@neilrichardson7454]

Thankfully, you're the type of random guy on the Internet whose content I enjoy listening to 😊

[@o224hsday]

And makes us all knowledgeable about things like a standard JORDAN unit kzbin.info/www/bejne/nqvWiIJjn9qih7M

[@mattslaboratory5996]

I'm gradually learning about this stuff from this channel, and it strikes me as the most sophisticated technology of our time. As J. often says, "As you can imagine, this is very difficult." I sure hope people like Dr. Shintake work out ways to make it easier, cheaper, more reliable. We all depend on it so much.

[@kenromaine2387]

It's a long way from the 2" FEP at T.I. that I worked in back in 1973. We used a "contact mask" that covered the complete 2" wafer (all the dies at once). A simple single UV lamp with motor driven metal shutter then exposed the photo resist on the wafer under the mask.

[@lllPlatinumlll]

Thanks for the work. I enjoyed learning about Lithography from you.

[@jtland4842]

I am hopeful that this can be licensed off to Canon or Nikon to help them restart their EUV machine development. While ASML is great in progressing the industry, it is far from ideal to be relying on one company to do it alone in their specific field without any direct competition. Of course I would also love if ASML could use this to improve their systems or develop a different kind if EUV that who knows, could potentially replace the rest of the DUV systems in whole or part some day.

[@rydplrs71]

Knowing just a little about how much effort goes into making an ASML EUV machine and their software let alone keeping it running I’m simply overwhelmed by the thought of even turning this concept into a physical machine. There’s so many suppliers with truly advanced knowledge of the tiniest details all putting in their best efforts for years to get it working. Just the mirrors, you can go to Ziess, or you can go to zeiss, or you can spend decades bringing another manufacturer up to speed and work out the incremental improvements in grinding and coating to dream of reflectivity, let alone a usable amount, then thinning it for atleast one. The e-Beam writer would be a reasonable task, probably just a new stage with z adjustments would work, that might require a new chamber but it’s more a list of tasks then rebuilding hundreds or thousands of entire businesses to get parts that will work to prove the concept.

[@jmirodg7094]

Amazing episode, this brings us to the edge of knowledge. thanks

[@friskydingo5370]

The 3 big problems we need to overcome, in my opinion are. Light source, substrate, mirrors. Any improvement on these items will massively increase output and cost. The hiden gem is someware in R&D. 👍

[@goldnutter412]

This seems awesome, cheers for the content. Clever guy indeed

[@benmcreynolds8581]

It's crazy how much power gets lost from the laser after reflecting off of the mirrors..

[@alexlo7708]

Because it is an extreme high frequency laser. A little bit higher frequency, it will go to X-rays and not reflect but goes through mirror.

[@LBjell]

You know it’s going to be a good video when KZbin hits me with an advertisement every couple of minutes

[@ewerybody]

3:57 "Schwurtz-shilled" yields some close approximation from English on Google Translate. 😉

[@lucaveneri313]

You bake the curvature correction directly in the mask pattern

[@UtahBlender]

Right. It’s a supplement to the industry already in existence, not a replacement. Like, frosting on a boring cake.

[@cabanford]

"Got dodgy" - my all-time favorite Engineering term ❤😂

[@meh2572]

Very informative

[@dziban303]

euv content

[@christophergaspar6520]

i agree with you EUV history should be in the school curriculum, would be my fav class

[@causewaykayak]

How about just "HISTORY" given that most folk know little and rely on pulp sources.

[@popquizzz]

Between the Japanese that know and perfect Optics ad light gathering/disbursement (Canon, Nikon, etc) and German engineering and innovation into optics, (Zeiss, Leica, etc.) and ASML right there in the Netherlands, this is going to be a boom to move to higher yields and greater high production bin counts reducing the costs and complexities of EUV wafer fabrication, it is a shame that this may not make it into foundries until 2035.

[@mhx47]

Well explained. Towards the end I (think) realise, how great this could be. Then you drop example of very common sloppy reporting, someone threatening to overthrow ASML :D That is just nother example I do not consume mainstream media. The sloppiness in reporting work is off the charts for years now.

[@pappaflammyboi5799]

Dispense with all the patent BS, let everyone play ball, and I'd personally invest in such a venture. We need more players. IP utility lockout is lame. It creates exclusivity and harms competition and innovation.

[@marcfruchtman9473]

Thank you for making a very interesting video.

[@TheFutureIsHere808]

hi, amazing video. I personally would be very interested in a video on optical computing :)

[@hanselda]

For a quick read of the paper, I think the biggest technical challenge is to control the defocus. In normal stepper we control the 6 DoF of both the mask and wafer at the ROI, that is where the mask and wafer is illuminated. If we have now 2 illuminated path at mask and wafer, it is generally not possible to keep both path in nm defocus, as both mask and especially wafer is not absolutely flat. It is not technical impossible but we will be talking about something like deformable chuck for wafer and mask. Not completely impossible, just say.

[@AG-pm3tc]

Collaboration between imec and OIST, sounds like a good idea.

[@youcantata]

EUV history in high-school curriculum is too early, too recent and on-going event to teach as history. Just teaching history of semiconductor up to DUV Immersion lithography is enough for high-school common core curriculum. EUV is only for advanced placement courses.

[@quademasters249]

It's like teaching every person to repair cars. Most of them don't need to know. I wish US schools focused primarily on long form reading and math. Once you have reading down, you can learn history on your own if need be. Was a story the other day about how many US students these days go to college having never read full books. They were only assigned snippets in school. If a person can't read, it's pointless to study anything else.

[@DemPilafian]

There's no need to convert metric units to imperial units. Every American who watches videos like this is familiar with metric. By the way, MJ units (2 m or 1 GOAT) are ok.

[@randomchannel-px6ho]

Yeah it's is kind of a myth Americans don't know things like that. At worst by highschool all students wil have been introduced to the metric system as the standard for precise highly technical fields like physics chemistry and biology. Granted not everyone satisfactorially passes those classes but...

[@Intelwinsbigly]

I still like conversions

[@quademasters249]

Yeah. I prefer metric.

[@jimurrata6785]

SI for the win.

[@DemPilafian]

@@Intelwinsbigly Unit conversions are distracting, but they're pretty easy to do. Anyone who wants the conversions is free to do them themselves. Having the conversions in the video is like an annoying fly buzzing around your food.

[@mikeselectricstuff]

Instead of a curved mask, maybe they could use motion control to adjust the position of the mask between exposures to approximate the curve - I realise that even x-y positioning at this resolution is a challenge, but so is pretty much everything in this field.

[@AdityaMehendale]

The mask gets accelerated side-to-side at ~500m/s^2 . (Any ASML'ers lurking here, please correct me!) Making fancy paths within this motion will be nightmarish :)

[@mrlithium69]

@@AdityaMehendale turn on Acceleration Control in Cura :)

[@avp6549]

@@AdityaMehendale yes, they do that so that in 0.08sec the mask hits 88miles/h....and that's when the magic happens

[@mechfan01]

@19:28 Hrmmm, the big red box hides the website. I wonder what it could be "About"...

[@bob_mosavo]

Thanks 👍

[@tsclly2377]

40nm is about where SLC chips stopped (or the most robust). I also think that chip design in a more circular system as opposed to square and rectangular may have electrical advantages (with triangular electrical 'buffers). Imagine a octagon chiplet design or triangular chiplet gate arrays [linear cutting into triangles, then octagonal]. The mathematics then become more aligned with two dimensional vector mathematics.. and with build height to a more 3 dimensional base 3 compute system.

[@p-jbroodbakker1303]

There's a limitation on NA baked into this solution, which might make it harder to optimize further. On the other hand, doubling the light output of the laser is only a few steps away: - higher droplet frequency. - increasing light output per droplet by optimizing in-flight droplet shape. I'm also not certain that the larger mirrors are not making this exponentially more expensive.

[@sloanNYC]

It is funny how insanely complex these things are yet so much of it is basic physics and optics principles at work.

[@davidwilkie9551]

Uninformed questions now matched by slightly informed, but very impressed video watchers. Thanks

[@douro20]

NTT actually makes EUV mirrors. Not sure where they're used but I found that a bit interesting.

[@autohmae]

I remember watching Bent It Like Beckham

[@bodaciouschad]

You don't need variable depth mask imprinting- you need to put the blanks on a shim that rotates them to keep their surface at the focal distance.

[@jeffstaples347]

Thanks shintake-sama.

[@MatthewSuffidy]

When you talk about light loss I'd expect answers more like total internal reflection.

[@techexpert-ww6yq]

🎉🎉🎉❤❤❤❤thanks for this video ❤❤

[@juozasg]

these videos are life

[@jannegrey]

For the algorithm!

[@Quickshot0]

I guess this could help with the need for ever brighter light sources that just keeps bedeviling them right now.

[@CliveBagley]

very interesting - thank you.

[@SilverNuclear]

This is some good analysis

[@knkootbaoat6759]

11:05 It didn't :( but that's because I need more time to understand it, and I'm not interested enough to do so at this moment.

[@sloshy1840]

You are God tier, literally!!!

[@mvadu]

21:31 let's do a Kickstarter now!? 😂

[@DirtyHairy1]

It's a good concept and probably it would make production cheaper in the mod-long run. Maybe this can be included in the next gen of EUV "tools". But if a new technology does not bring a significant leap in chip performance, its better to keep the working designs with which you are able to deliver consistent quality and quantity. Saving a bit of production cost is a viable point to invest in when there is no other, more promising, route to innovate left. Lower prices is currently not the crucial factor in the chip industry. It's mass production with an acceptable level of quality. Source: some dude on the internet

[@danpovey]

If the wafer is so close to the M2 mirror, wouldn't substances ablated from the wafer be deposited onto the M2 mirror?

[@unvergebeneid]

People really tend to underestimate the amount of work it takes to get from an idea to a prototype and from a prototype to a product.

[@tintruder224]

How about spherical masks? Very large radius. Since optics are spherical (or aspherical) you could define zones of adequate zonal "flatness" at the wafer by keeping mask features equidistant from mirror surfaces and exposure timing. Think of it like using a spherical platen to roll an image onto a plate. The flat area of the printed image would also be slightly greater than the straight-on 2D cross sectional area of the mask. Something like how a mercator projection of the earth globe onto a flat map works, but preserving feature size and aspect by rolling projection rather than full field ala mercator which stretches what is further from the center.

[@Joe_VanCleave]

So what keeps the EUV light reflecting from mask to the M2 mirror from entering the central aperture in the M2 mirror and “fogging” the photoresist-covered wafer with out-of-focus light?

[@nijram15]

I think thats a very good point. In optical engineering this is called straylight and usually refers to light bounching of unwanted surfaces.

[@hmson4378]

'Bend it like Beckham' HA HA 😅 Love it!!!

[@martylawson1638]

So a potentially useful optical configuration to pattern the bulk of the chip that doesn't need state of the art resolution. Looks like it would offer a simplified process vs older 193nm tools. (single patterning, no immersion) Cannon or ASML picking this up would be best case for this prof.

[@Xiaotian_Guan]

I don't think bending the mask is a problem at all. Astronomers have been doing this with adaptive optics systems for decades. The simplest way is to pull a slight vacuum behind the mask. Though I'm not sure if the inside of an EUV machine is a vacuum. In that case you need a bunch of small actuators. But that's also a solved problem.

[@horizon7011]

I don't think the problem is if it can be done. It's peoples reaction to wanting to bend an almost million dollar mask that is not designed to be bent.

[@michaelbuckers]

@@horizon7011 The mask hardly costs more than the sheet on which it's printed. The digital design work for the mask costs a lot, making the mask is the cheap part. Also even ceramic items can bend, in fact it's the feature which mineral wool relies upon. Not to mention you can just design it to bend.

[@ersetzbar.]

I wonder if the euv comes in paper boxes with ikea like instructions for assembly

[@fikamatyi2]

Joseph Petzval was not Slovak for sure.

[@rogerbartlet5720]

Those mirrors on the current design cost tens of millions of dollars to make, in fact ASML bought Zeiess to have them make mirrors exclusively for the EXE 5000 family.

[@aapje]

ASML bought a quarter of one subsidiary of Zeiss.

[@rogerbartlet5720]

@@aapje They bought enough to set priorities.

[@hansmuller3676]

Nope - to Split the 100s of Million Dollars of Investment. The big chip makers bought many shares of ASML so they made the Investments down the supply chain

[@rogerbartlet5720]

@@hansmuller3676 Better than spending it on stock buy-backs?

[@MrMulleteer]

TBH if this concept only lives in a optical simulation it really does not matter much. Any simulation, let alone optical is notoriously hard to get to reflect real life problems. Especially in such a delicate fine tuned setup. Physical prototype is needed.

[@LITTLEgiiant]

Well I hope either Nikon or ASML takes interest in the paper so a production machine can be built

[@TimoteiUA]

That's probably a very interesting video. Too bad I'm too busy being in constant epileptic shock to enjoy it.

[@beautifulsmall]

bottle lids have suddenly become attached, that must be an almost world wide agreement. the packaging industry ?

[@IanMott]

We should talk!

[@woolfel]

the michael jordan measurement unit, got me laughing

[@Inceptionxg]

We are struggling with choosing the architecture, so could you do some video about that???

[@inyobill]

I'm surprised that the abstract posited only 10% power savings.

[@lorddorker3703]

I would imagine you could fix the mask curvature issue in the software producing the mask. Process the image to account for the effect then "print" that image instead of the normal one

[@erickhyde9562]

Bend the wafer. It is flexible enough for minor bending. And can be achieved by many methods.

[@Lenskha]

Nice video as always