The Extreme Engineering of ASML’s EUV Light Source
ฝัง
- เผยแพร่เมื่อ 1 มิ.ย. 2024
- After 20+ years of development, extreme ultraviolet lithography has become a commercial reality. As I write these words, multi-million dollar machines from ASML use EUV light to create impossibly small patterns in wafers.
This technological magic requires a powerful heart inside of it. And indeed, there is an amazing system driving ASML's $150 million lithography machine: The EUV Light Source.
In this video, we are going to look at the lasers firing pulses at tin droplets to create the powerful, 13.5 nanometer wavelength light for our latest, greatest microprocessors.
Links:
- The Asianometry Newsletter: asianometry.com
- Patreon: / asianometry
- The Podcast: anchor.fm/asianometry
- Twitter: / asianometry
Subscribe! And remember to sign up to the newsletter: asianometry.substack.com/
Great content as always. One comment: Cymer is like symer, not kymer.
Now do Gigaphoton
Oh my god ,such quality niche content! what's your background and what do you work as , if I may ask? :) I might have found my new role model!
SUbscribed. That was a great video expalanation!
The video is absolutely amazing!
17 mins of this video is much more informative of what we learn in the obsolete uni system in years!
Great job and thanks a lot!
As someone working at ASML, I can say this video is surprisingly accurate.
Ok. I got a question. Will future tech still require chips?
@@dannylo5875 : Maybe we'll go bio, or optical.
@@dannylo5875 For the foreseeable future, yes.
@@TheFulcrum2000 no u don't work there
@@brodriguez11000 This is not maybe anymore.
This is truly an astounding machine. Kudos to all the scientists, engineers, mathematicians, chemists, and precision fabricators for pulling it off.
Dont forget quantum physicists
Wow. I'd heard the phrase 'tin droplets" associated with EUV, but I had no idea they were using one laser pulse to create a convex target for the main pulse. That's absolutely ridiculous.
As an ASML employee, I can say that Cymer deserves all the credit for the light source, that's pronounced Sy-Mir. Cymer began development of EUV from a design perspective back in the mid 2000s. The reason ASML bought Cymer is that the two companies required close cooperation to develop an EUV light source/scanner system and this was technically difficult because of responsibilities to keep technology secrets.
The German Trumpf has some serious contributions as well. They are among the leaders in their field.
Dude, why say you work there when nobody believes that about anyone just online? Just sound smart and credit real sources like physical in person training session 3302? Like I know that also sounds silly, but bro, you just sound like a random douche know it all but just saying, I was there, totally accurate and real
When Cymer was in it's infancy I interviewed with them. I'm a medical laser engineer (BSEE).
End story was I decided I wanted to stay on the medical end of this technology and am glad I did, but who knows where I'd be if I took them up on their offer?
I has *NO* idea this is what it took to make EUV light for this manufacturing process. It's hard to believe, that out of all the people working on this, that this was the solution they landed on. Truly spectacular engineering and science. (if not straight up bonkers)
there are Chinese fanbois who believe that china can copy and build a similar machine in 2 to 4 years, this thing took 40 years from concept to prototype to working samples
@@eduwino151 Yeah but that was from scratch. You realize that with the fact that it's already been done and industrial spionage the whole equation changes, right?
@@kristofferjohnsen4002 doesnt work like that for lithography machines , minus the research and tech transfer you cant do it , China is still struggling with reliable jet engines and those are simple tech compared to semiconductors
@@kristofferjohnsen4002 yes because they have the basic concept down it will only take half the time 20 years.
@@eduwino151 just matters if the can steal the actual relevant information required to make the EUV, if they can it will allow them to cut the time down for their R&D…. I agree with you though about china and their problems with something relatively simple like jet engines….
I heard the quote from an involved engineer that this machine is probably the most advanced piece of hardware ever contructed by a long shot.
The science and engineering involved in this is almost magic and the fact that it sattisfies serial productions demands is mindblowing.
ZEISS posted several videos explaining some of the science in detail if you're interested.
i think only some one-off marvels like space telescopes, supercoliders and fusion reactors can be described as more complex... it is indeed a feat of engineering and human ingenuity and definitelly one of - if not the most advanced industrially used machine
This is definitely up there in terms of complications, collaborations, total human-hours, etc. These machines, each one of them, are absolute mind bonkers accomplishments of engineering.
I've stood beside these new EUV tools and observed them work. Knowing what's going on inside of them blows my mind. It's just incredible.
Your channel is one of the very few channels that bridges the gap between a good lay understanding of the semiconductor industry and a professional understanding of the semiconductor industry. Its a very rare thing to find in any field.
I find it interesting that the "secret" part of "trade secrets" is becoming less and less important as a smaller and smaller number of people or corporations or nations are actually capable of making use of those secrets.
I appreciate the explanations that companies like ZEISS give for the construction and function of their optical systems, but on its own it doesn't really increase your understanding of how the whole process works.
I get a lot of people in the maker space asking questions like "why can't we 3d print our own microchips". This is why. There's no half-assing an EUV light source. Similarly, there's no half-assing ultra-pure water or the purity and precision of reagents and gasses used in semiconductor manufacture. One microscopic droplet of process oil in a gas stream is enough to ruin a processor, or, worse, ruin a photo-mask.
If you're lacking for ideas for videos, one of the things I don't really understand well is how wafers are cut and polished. I'm assuming this has had to keep up with the process nodes as well.
I love that as well it's like "Yeah here's our secrets and research, the fuck you gonna do with this info?". Not that there aren't things to keep secure but still it's insane the degree of advancement you need to be to even begin dreaming of technology such as this even with all of the information on how to do it at your feet, what an astounding technological and international effort it was to get here. (Also amazingly explained as always by Asianometry)
The wafer video is done. It’ll be out in a few weeks.
@@Asianometry Cool! The last I remember about how wafers were cut it involved reels of wire with diamond powder bonded to it. I'd guess there were some kind of acid washing and polishing stages after that. I could probably look it up but I won't. You seem to have a good understanding of the supply dynamics and political interactions that influence the semiconductor industry and thats an important part that wikipedia and research papers aren't even going to touch.
makers are just makers, probably half of them do not study deeply about EE, do not hold a degree of anything related to it, or even something related with physics of semiconductor.. no wonder that question comes up 😂
@@eone199 I think people get the impression that Moore's law somehow applies to 3d printing. We've been stuck at (optimistically) 100 microns for 20 years for 3d printing and 10 microns for 2d printing for about the same amount of time.
For me I guess its interesting to think of how you could pull off making functional electronics "at home" but you're talking about combining molecular beam epitaxy and electron beam lithography and nanometer scale additive manufacturing (which doesn't exist yet, we're still at the micron scale even in research) and now you need tubomolecular pumps and far beyond reagent grade materials and high end CNC machine tools.
So I think you could make your own processor as long as you were willing to invest your life savings in a machine that could produce 1 questionable quality processor per week at a running cost of a thousand dollars a piece.
For the record, though, we're sitting at the 5nm node right now (which doesn't mean 5nm at all, but that's another story) and electron beam lithography has been able to produce sub 10nm features in semiconductors since the mid 1990's. A lot of progress has been made in industry and mass production but in terms of the academic research leading it, far less progress has been made than people assume.
This makes landing rockets sound like child's play! Mind blowing science and engineering!
Looking forward to the Komatsu/Gigaphoton approach video.
I think it’s in the same ballpark. I wonder how one would quantify that though. And I wonder what other projects play in these league. CERN probably ?
Totally insane. Everywhere else in physics, synchrotron radiation is appreciated for its narrow spectrum. But for this application, it has too wide of a spectrum. Incredible.
The synchrotron frequency spectrum IS narrow. The video claims that it has too broad of an angular field of view.
Too divergent spatially
It boggles the mind how many billions of dollars the information conatined in this video would have been worth even 10 years ago!
nope the basic info of EUV has been freely available for years how to build the stuff that makes it work that is where the money is and nobody knows how those companies that supply those parts do it exactly
@@eduwino151 exactly, that’s why the specialized physicist get paid the big bucks to do this work, no way around it…
@@eduwino151 Sure, but the point of the video isn't that the important technologies (prepulse tin) are theoretically possible, but that they ended up being the best way to go. If you could have told tsmc 10 years ago that those two technologies were the solution and to just drop everything else there, I'm sure the value could be measured in the billions.
@@itisinfactpaul2868 Insert head Engineer/manager. "And that's why I get paid what I do"
The clip of the CEO in that wall street movie from a few years back.
While browsing Gigaphoton's site, its EUV is definitely different, as it's using magnetic fields to direct the tin debris/ions away from the mirror. But it lists a mirror lifespan of >3 months, how will it compare to ASML 1 year replacement life for its mirrors
At that point in this clip that was my first thought on what the solution might/would be.
I could've understood magnetic mirrors/lenses for debris/ion mitigation better than the hydrogen gas (or how Asianometry did(n't) explain it in this clip).
I work in the world of cinematography lighting, while I do have a knowledge of « light » and a very basic understanding of laser chemistry I am far far from being an optics engineer or physics expert…. YET the way in which you present and explain the industry needs, challenges, mitigations and solutions were EXTREMELY clear and easy to process….and in the end, HIGHLY ENJOYABLE and EDUCATIONAL. Thank you for the video and for teaching a 58year old man about something so intricate and intriguing. MORE PLEASE !
When I was a physics student some decades ago, a wavelength of 13.5nm was considered "soft X-rays." But "EUV laser" is a lot more palatable than "X-ray laser" I guess...
If we had kept the maser/laser naming scheme alive, we could have had xasers 🥺
Edit: but after watching this video, it doesn't seem to be an actual laser? I think they just use a laser to produce the plasma, but the plasma itself doesn't lase(?)
@@MarcusMedomRyding Correct, they laser a droplet of tin or something and that produces the EUV
It is indeed soft x-rays. Extreme UV is a marketing term. So, you are right, it's a more palatable term.
@@inomo Naming on the EM spectrum seems similar to the musical scale. Soft X-ray and EUV are the same, similar to how Eb and D# are the same. 😅 That’s how I look at it.
Yeah it's not so much a laser as pulsed coherent light.
The quality of videos from Asianometry has always been ASTOUNDING! The research thorough. The presentation clean and engaging. THANK YOU!
Hello! What a very well put together video I’m very impressed. I currently work for trumpf on the co2 drive laser. Ive sent this video to the recruitment agency who hired me they always send applicants videos like this to prep before hand
Please, develop more robust RF generators. Reverse power shouldn't kill them that quickly 😆
Thank you for putting this content together! This clearly shows the insane amount of engineering and science laying the behind the piece of electronics on our hands today.
@17:00 "This last paragraph sounded like a dialogue from a Star Trek movie." You got me here. I was depressed over quarrel with a landlord, you saved my day.
At 16:51 you said *"the modern day moon landing."* Just moments before I was thinking *'This makes the moon landing look easy.'*
Thanks for another interesting video.
Well now I don't feel so bad that my DIY EUV machine was a total failure.
Incredible video! I would say it's your best one until now. It's nice how you went super deep in the subject (for a youtube video) keeping everything clear. Props. Keep up the good work.
What an amazing explanation of an incredibly complex process. For all the advances in computational modeling, there are still many applications where empirical experiments are still necessary, and it’s in this intersection of empirical and theoretical research that a lot of amazing solutions come up
Great job of ASML ! 🇳🇱
Whenever I see the innards of an ASML stepper, it makes me think of those hugely complicated machines they used to keep Akira cryogenized.
Is there one here? Time, pls?
The fact this works at all is mind blowing.
Wow! This video sort of flipped my world of understanding regarding chip manufacture upside down. We've come a long way since discovering fire and casting shadows on the cave wall.
The most interesting thing is that light hold the key to the next phase of electronics or photonics more properly, as we now are at the cusp of digital optical transistors and Qbit transistors made from light sensitive protein. This is my companies work and we’ll be coming to Taiwan soon to bring these to production reality. Our work will include super computer for development, laboratory and more.
Please elaborate more. One question I have would be if in your opinion we could see such a product in a niche application on the market that would be better in one or multiple categories than traditional silicon and what those categories would most likely be and how this would be achieved compared to the present.
Or do you have some whitepaper or Research-/Businessplan I could read to catch up on?
Sign me up for an SDK…..
A) Holy Shit
B) Can they run Crisis?
please elaborate. You do have a tech company doing this?
This is great. I don't think I have seen all this information covered in a single source before. Now I really understand why these machines are so expensive and take so long to make.
Mindblowing! Both the engineering behind EUV lithography and the quality of your videos buddy!
Amazing video! Thank you! This was so well explained! The quality of research you put in could not have been easy to parse, let alone explain, especially in such a way a dumb dumb like me can even attempt to grasp!
Bravo, and thank you! I have always been fascinated by the EUV topic, but have
Found it either impenetrable, or incredibly surface level, and lacking the depth and explanation of each step of the process. Then you went even above and beyond by explaining the problems faced at each step in the design history! I don’t even wanna know how many white papers and documents you must have had to go through researching just one of these videos!
P.S. On a side note: What was your research process, and how many papers did you have to read on average to write each of these scripts John?! I have become absolutely fascinated by the work involved in your script and writing&research process, almost more than the topic of the videos at this point!
Care to share a little bit about that? I’m sure many others would love to know as well! Many subscribers must find the process as astonishing as I do!
Thanks again!
Machines like that definitely make fusion feel achievable
After watching almost all your videos I signed up to your newsletter. Your channel is amazing and your humour is so sharp 😂 I wish I could support you more. Keep up the great work. Cheers. Glen from Sydney.
As a programmer, few of the many university lectures on how cpus/silicon circuits work enlightened me about how it all truly works as much as your channel has!
I am always intimidated by real engineers on this level of proficiency. With software development, you always have comparatively more margin for errors when you design a system. Engineering like this that requires you to get things right on the first try with so much investment involved is crazy.
Awesome work, again! I was employed in some sections of semiconductor industry and I'm very glad to find such very understandable videos about highly complex themes without the common inaccuracy of modern days medias. Thanks for your work!
What an amazing system and great job at explaining it so simply.
Bro I’ve been following you since you had no subs. You turned 100 to 149k subs real quick! Let’s see that writing style, in-depth analysis and entertaining commentary turned up! Let’s go 200k!
Thanks for watching!
Another home run --- every single one of your videos are gold.
Big fan of your channel after checking out a good portion of it. I have to say thank you for enlightening me to some truly tectonic movements in science and industry both historically and currently.
you make the best content on TH-cam. Absolutely unrivaled in production quality and the topics
Fantastic, another example of why I am a happy Patreon supporter. Great place to learn!
Hitting liquid tin droplets 💧 with two separate lasers simultaneously at perfectly calculated angles, 80 times a second, to create a light conversion efficiency six times greater than the YAG laser… WOW
50000 times a second, not 80!
Brilliant! I absolutely love high tech like this and you've been able to put it all in simple words for a lay man like me. I knew that ASML machines are notoriously difficult to manufacture, but didn't know the details behind it. Thank you for sharing
That’s crazy. Shoot a pre-pulse then smack it with a big pulse. That’s straight laser DP right there no cap 🤣
this is the most detailed how ASML EUV works. finally i found it, thanks.
Fantastic explanation once again, thanks John 🙏
This reminds me of a scene in the Expanse season 4 where the Rocinante's reactor stops producing plasma from fuel pellets. It literally shows a pellet falling and lasers hitting it but failing to create plasma.
Inertial confinement fusion!!! I was also just reminded of that from this video
There are prototype machines that does this. (not successful enough though)
Thank you so much for the educational videos. I don’t begin to understand a lot of what you present, but I always feel that I have gained something valuable. A few weeks ago I decided to cancel my cable service and put @ least that much money into Patreon, to support the creators of the TH-cam channels I enjoy. Asianometry is one of the first channels I felt I needed to support. Thanks again.
Extreme is an understatement. All this to only create a specific light is INSANE!
The video is absolutely amazing!
17 mins of this video is much more informative of what we learn in the obsolete uni system in years!
Great job and thanks a lot!
Thank you for the deep coverage of the technology. I enjoyed it very much. Great work. You’ve earned a new subscriber.
This Video Explained it completely !! Thanks for making it !!!
Love your videos! You always seem to find the right topics
Very awesome technology analysis. Great video. Love ❤️
Your best and most insightful post so far. I particularly liked the conclusion, because it is something I have spent a career arguing in a quite unrelated field. The guys who control the money in this world (including government money) underrate the importance of getting the theoretical basic science right for industrial R&D. Engineers who say "best to just do a seat of the pants trial and error approach" really do not understand how technology actually works.
That is bonkers. Utterly mind-blowing machine. I love this channel.
Thank you for posting this. I'm looking forward to see your video on Gigaphoton.
As a lens technician I found this video very interesting. Great work.
That was an awesome overview of the ASML system, thank you. On to 200k.
I got the opportunity to hack apart one of the tin covered mirrors in my undergrad. Unintentional tin coating is still very much an issue today
Outstanding presentation! Thanks!
Great conclusion paragraph- very stylish
Brilliant episode, thanks!
Loved ur way of explaining things ❤️.
Great video. Thank you for the knowledge. Will watch your other videos and I will continue to learn
The quality of videos from Asianometry has always been ASTOUNDING! The research thorough. The presentation clean and engaging
Amazing content, wish you talked more about Cymer as a company.
You do such a good job always a pleasure watching
What a great video, thank you. I made the first sets of air bearings that the laser assembly rides in so it can be steered at the tin. I honestly had no idea the scope of this project was so massive. Cymer found a little company in NH called Nelson Air. It was a big project for our 4 man shop. One of the more challenging things I’ve machined.
Air bearings… Didn’t even know such a thing exists
If you're at liberty to, please tell us more it sounds fascinating
Very awesome and engaging documentary! 👌🏼 Channel subscribed!
Phenomenal to me that they're doing lithography
Now just imagine how they were doing things like 7nm with 193nm light on DUV. Double patterning and weird interference patterns on the mask, Intel tried quad patterning on their original 10nm design but it introduces too many opportunities for error. Although they had many other issues/techniques that ended up abandoned like contact over gate (a very dense transistor design that was too sensitive to defects and was switched out for gate all around)
@@Jaker788 it's like trying to do brain surgery with a battleaxe.
Superb presentation!
very interesting stuff , love the vids , please do the video on GIGAPHOTON approach as well
Thanks. Awesome learning about what seems to be one of the most complicated engineering by humans.
My buddy works for ASML and loves his job. Top notch company and they're expanding into the US.
The LHC is arguably just as insane piece of engineering. The difference is that ASML are knocking out several of these machines every year. Just taking once of these machines from concept to manufacture is insane enough. To make them day after day on a production line is mind-blowing. Kudos to ASML, Zeiss and Cymer for the epitome of extreme engineering, and to myriad of physicists and engineers who brought this to fruition.
What the hell are you folks smoking?
Very interesting. I wonder if an approach used by Becton Dickinson around 1980 to steer droplets either is used by ASML, or could be. After forming droplets with a Piezo, the droplets were hit by a laser and optically scanned (diffraction&absorption). The spectral results were instantly analyzed and used to apply a positive, neutral or negative charge to a pair of plates.
As the droplet passed between the plates, the charge would then SORT each droplet into one of three vials. This served as an alternate means for concentrating a desired component, when other methods (like variations in density) would not work.
The analogy for ASML would be to used the charged plats to steer or align the tin droplets, and perhaps also serve to minimize the travel of tin debris.
AFAIK there are machines that allow for the same thing as described but with photons/lasers.
but since this is done 10000 times per second and these plused lasers are on the femtosecond scale floatation or gravity cancellation is not needed. they are in fact so fast and precise that it has been used to produce slow motion videos of single photons traveling, thats billions to trillions of fps, the slomo guys have a very good video on that topic.
@@G4m3G3ni3 About a decade ago, I visited the femtosecond photography labs at MIT.
th-cam.com/video/Y_9vd4HWlVA/w-d-xo.html
@@gregparrott Wow thats super cool! :)
What do you think of my theory? Since all parameters in the EUV machine are known or measurable and the random factor of tin contamination has been engineered down to a predictable degree any additional method of levitation would add another complication and thus reduce efficancy of the machine.
Now another question that comes to mind now is if the horizontal spacing of the tin droplets should be detected from distance or artificially manipulated by the method of dispension.
I think droplet size is controlled by tin dispensing nozzle type+pressure+viscosity+temparature and cavity athmosphere. Since it is a two dimensional line with 5x the droplets of the actual pulse frequency they only expect 20%of droplets to be useful. I guess it its almost trivial to use highspeed cameras because models with multi million fps and very accurate interface that provide a very high line resolution - atleast high enough to determine good droplets, time the first pulse,determine correct tin formation and time the second pulse.
I guess the performance increase mentioned in the video will come from expirience "in the field" and subsequent data collection and process optimization.
@@G4m3G3ni3 What you're saying is interesting, but I'm missing some of the details and have to conjure up.
We're both assuming that the tin 'droplets' are formed in situ, which implies it is molten when it enters the nozzle. While it likely is molten, it is possible that they're actually small pellets.
Assuming it's molten, the Piezo actuators used in fuel injection could be (highly) redesigned for metals. For starters, tin's density is 7.3X that of water, ~10X that of gas and ~9X that of Diesel. As you say, the piezo's characteristics, the size/shape of the nozzle's cavity, and the electrical pulse's shape can influence the resulting droplet.
Saying it is a 'two dimensional line' is confusing. By definition, a line is one dimensional, and a plane is two dimensional.
Also, I found nothing in the video which suggests "5x the droplets of the actual pulse frequency". At 14:14, he says they're hitting "50,000 bullets a second with a laser, twice (twice because two lasers hit each droplet). He mentioned the droplets are traveling at 80 m/s and are separated by 1.6 mm. If you crunch those numbers, this affirms there are 50,000 droplets per second.
Thank you for these amazing topics you cover.
Cyclotrons and synchrotrons are not the same type of accelerator, despite both being "doughnut" shaped -- their principle is not the same: cyclotrons use a time varying magnetic to speed up charged particles by magnetic induction, whereas a synchrotron uses electric fields to speed up particles in linear stretches, and magnetic fields at the "joints" to bend the trajectories -- synchrotrons are not actually circular, but rather they are many-sided polygons. In synchrotrons, the X-rays come out at the vertices that join two linear stretches.
In fact, the LHC is a synchrotron, and it also produces X-rays the same as those synchrotrons which are used explicitly for that purpose: the X-rays are a byproduct of how synchrotrons work. The LHC is not simply related to the ones that are used for their X-rays, it is one of them, albeit used for a completely different purpose.
And great video, by the way. Despite my minor "complaint", you explain things in a very straightforward fashion. Keep up the good work.
Cyclotrons use a static magnetic fields and a RF source for accelerating charged beams. Betatrons use time varying magnetic fields to accelerate particles.
"Cyclotrons and synchrotrons are not the same type of accelerator, d"
He didn't say they were, though. He said the synchrotron developed from the cyclotron, which is correct.
@@Eletronicafgyou're correct. I mixed them up -- been a few decades since I studied them. I should do a bit of refreshing of some accelerator physics.
@@isodoublet yes, you are right, I got the betatron and the synchrotron confused. As for the rest, I was trying to add more information -- the LHC is a synchrotron, but as far as I know, it's not used as an x-ray source.
And of course, Asianometry's videos are excellent regardless.
@@gildardorivasvalles6368 Oh, don't get me wrong, I think your post was very informative.
Fantastic video, thank you.
Good job on the video - keep-em coming!
Cool best information on this tech to date, thanks for your attention to details 😀
That's an insanely complex light source. Imagine they figure out a way simpler method after all that? That would be funny and a good thing.
I mean.. basically its a particle source.. tin just has the right frequency
that thing is literally a mini particle accelerator it is the simpler method
Almost unbelievable that these kinds of machines can be dreamed up, fabricated & assembled, fine tuned & then work.
Very good video, many thanks indeed.
This was very interesting. Thankyou.
Kinda want to find one of these ultra-clean mirrors and lay a big greasy fingerprint on it
Thank you for sharing this
2 years on you almost got the 20k likes :) Love your content brother!
And this is just one of many reason why I have been into physics and all the sciences since I was a kid!
OMG. This is far beyond what I had imagined.
Even if you understand every bit of this, it still feels incomprehensibly magical.
Just subscribed to your channel based on this amazing video. This was brilliant. Thank you 💡💡🙏🙏
You have to hit 50k tin droplets a second, shot at 80m/s with a laser, twice. ASML: "hold my beer!"
Well done video! Really good
great video, especially at the end
Definitely the most complex machine ever built that wasn't a one-off (like the Large Hadron Collider or ISS for example).
I see we're well short of 20k likes but I'd still love to see what Japan has built to rival this-and, hey, Asianometry is the name right...
yes, pulsed lasers has ""more"" energy output than continous
the popular laser engravers using pulsed lasers
from 20khz up to 4000-8000 khz
at 20 khz a 20 watt input can result several kw output in the momentum
in a very tiny spot like 0,01 mm
20+ kW in, 250W out. Oh well, at least the factory has no problem heating the place.
Thank you for this great video!
Wow, I didn't know this had been achieved. I remember reading a bit about the promise of Xenon years ago. I was unaware what was going on with tin. I was always interested in mirror making for telescopes. To think that the first pulse creates a curved surface is amazing. Then the second pulse strikes this curved surface squarely to create EUV....WOW! What an incredible accomplishment....