I work for on of the support manufacturers you listed, Tokyo Electron. ASML's scanners really are a sight to behold as they're being installed, truly marvels of engineering. Our tools are extremely well built as well, and are fascinating to watch when they're working. The speed with which they transfer wafers to the many steps of the photolithography process, including the scanner, is something out of a sci-fi film.
I work for another support manufacturer, also for the wafer stage, seeing it in the testing tools is insane. Mindblowing how fast yet precise the movements are
I've been on the prototyping/building/design side of of Intels EUV Vaccum Control System for almost 8 years now, multiple revisions, tweaking, process updates, it never ends. But I'm the only one in my group that knows what the ASML Lithography machine actually is, and that I tie into that machine to make it function. All because 8 years ago I was asked if I want to take on the challenge of prototyping the first version based on terrible drawings and documentation, and I said yes, and somehow the whole project dropped into my lap instead of getting laid off when they moved production out to the new Oregon Facility. I went out there for 3 months to train my replacements, and came home to a work from home job I wasn't expecting, it all worked out in the end 🙂
Yup that they create the tiny bits that make silicon into a powerful tool in and of itself is a little left of outright magic. I can only imagine when they start making monolithic nitride devices that use photons instead of electrons. 😮❤
Your content never fails to impress me. You have brought immense amount of information to us and amazing accuracy. I wonder how you managed to do such comprehensive research unless you have very reliable industry connection. Just want to say thank you and amazing work.
@@gilligan87 Oh really? That'd explain a lot :-) But even allowing for that, I'm just blown away by the depth of research and information he brings to each new video. I can't imagine how he can possibly manage to do this level of research and production and (apparently) at the same time hold down a full-time job doing something else! :-0
Any time I hear about the EUV manufacturing process I can't help but feel that every tool at every step of the process is way too complicated to scale well. The amount (and quality) of engineers necessary to maintain even one of them would be a nightmare--not to mention the potential supply chain issues with these ultra specialized parts.
@@ClockworksOfGL I'm not convinced they will be economically viable. Have you ever seen one of these tools in person? They're absolutely ridiculous monsters.
I’m always amazed when somebody thinks the chip process is too complicated and expensive. The number of transistors in a small chip is crazy impressive. Billions. It is complicated and expensive for a reason.
Scalability by multiple stand alone processes maybe? Much like Apple did back in the magnesium case era, they stood up a few thousand CNC millls to make the cases individually. You get some economy of scale advantages by using a single brand and model of CNC machine.
@@akatsuki6371 They are economically viable. They are using this monster machines for several years now. The foundries raise their wafer prices but the chipmakers have enough margin to absorb the increase in price. It is very economically viable in fact that intel already ordered the version 2 an even larger monster machine.
Please make videos about why those EUV alternatives didn't materialize, that would be awesome. Some of them, like x-ray and electron beam, would allow to print feature sizes much smaller, right? Can they still happen?
Electron beams can already make features much smaller than EUV. Its how a lot of the masks are made. Its problem as a technology currently is that it is way to slow to be used for wafer production.
EUV had already surpassed anything they hoped to achieve with proximity X-Ray and e-beam is direct write and was never going to meet throughput requirements.
Back in the early 90's, EUV was called Soft Xray. I worked on component optics fab and test for an experimental small field exposure tool Bell Labs was developing.
@@kellensanna Indeed this has been true since long time ago - for example, I remember that already about y. 2000 ~10nm features were (slowly) printed using a modified electron microscope under university lab conditions. This was obviously way before EUV was available, but just for comparison contemporary (DUV) AT1100 machines could do 'only' CD of 100nm but with decent throughput of about 90 of 300mm wph.
I'd say even more so because there are typically often far more failures than there are successes. They offer far larger sample sizes to learn from. And the lessons you can learn from them are likely to be more universal than the successes.
Not sure this was a true failure for Nikon, it is just ASML/Intel are so in-bed with each other, penetrating the chip manufacturing market for companies that can afford $200M machines is quite limited. Although Nikon may have lagged technologically, just a bit (not much), it is the networking between the two dominate players in this Tech Space that became Nikons Achilles heel.
@@harthenry I'm prerry sure the lesson is, "get your shit together and gather a sufficient war chest to fund your R&D if you want to be cutting edge and an industry leader." The war chest could be any mix of cash, miscellaneous capitol assets, facilities, tallent, and customer consultants, but you have to have a sufficient war chest. Nikon/the Japanese reasearch associations just had insufficient reasources applied into beating ASML.😅
Holy crap, I never imagined the level of complexity that goes into making chips. I knew it was difficult, but never to that extent. That was super interesting video, thanks!
At the 12:53 mark. That's the MET 5 tool. I worked on both MET 3 and MET 5. Also, the exposure shots on the MET 3 were 200 x 600 microns. We exposed test patterns on experimental photoresists and provided metrology via SEM for dose and focus as well as process windows. I was there between 2013 and 2021.
I am super interested why the other alternatives to EUV failed (like proximity X-ray). There is great video covering one of the EUV alternatives: "Why 157nm Lithography failed"
@@NoNameAtAll2 Yes, why 157nm failed, however that wasn't an EUV alternative, it was supposed to be an easy stop gap between DUV and the next gen. But he wants info on the other alternatives that were higher on the list, like proximity x-ray being considered the #1 choice over EUV at #3.
You have explained most key functions of a litho machine, but one thing I cannot get my head around is the precision motion systems inside them that have to reliably and repeatably reposition the wafer to the nm, that kind of precision is just generally unheard of.
There are 2 components: a measurement system like an interferometer and motor systems that position the stage. The biggest secret is to do this all at very high speeds...
@@tommy2cents492 The resolution of those motors and gears must be insane! with non-existent backlash to reliably hit those exact spots again and again on the wafer, especially with multi-patterning, it wouldn't surprise me if they have two modes, a low resolution mode for large and fast axial movements and an ultra high-resolution, high-precision mode for honing on the exact spot it needs to be at the end of a large low-resolution travel move I can imagine it's also a very dynamic closed-loop system that requires constant real time micro-adjustments that might be guided by some kind of etched landmark on the silicon?
@@ChimpyChamp if someone asked me to name a technology they thought came from aliens, it would be scanner stage positioning. You look at the requirements and can only conclude it shouldn’t work but it does. At a very simplified level, they use bands of closely spaced electromagnets to move the stage (the entire stage is levitating over a near perfectly smooth block of granite). The closed loop control uses an interferometer and powerful signal processing to feedback an electrical signal to those magnets that results in oscillation damping and acceleration control. The part of every stepper/scanner you don’t see is the rows of 2m tall control racks supporting it. They’re usually tucked away in the sub-fab, quietly and frantically sending out a flood of information to the control systems. It is truly mind boggling.
Fantastic videos. Well done! Can you do one on the failed 5th generation computer program. I grew up in that era and it seemed that Japan would produce what we are now seeing with chatGPT and DeepMind. That never happened. It was yet another massive consortia effort
@@BR-hi6yt No AI model is "woke". Only the dataset is they are trained on can be "woke". If anything actually diversity in datasets is a positive starting point, because otherwise the trained model ends up doing exactly the same things that humans do when overexposed to the same things ad infinitum. It forms biases. Which isn't good for a model that is going to be used by a multinational business. The whole reason that white people from rural towns with mostly white people have trouble differentiating different black peoples faces is because our brains suffer this problem - it's intrinsic to neural networks to become weighted on their training data and form bias on those weights.
Googles Ceo Sandar explained on 60 Minutes, how after awhile the computer starts spitting out false info and garbage not fit for a black comedy. Truly sad what perfection means, it’s all a lie!
I love your videos. I have not watched them all, but have you covered the success of ASML and Cymers' DUV contributions up to the point ASML bought Cymer in 2013?
Xray proximity's primary downfall was the mask. It was a thin membrane with Au or W absorber which was not rigid/ stable enough to meet pattern distortion specifications.
The serious problem of Japan is that bureaucrats are totally useless. No company can get along with costly R&D itself alone... Please refrain from the 5th generation as requested below. It was an another nightmare,,.
Aye. But it seems that Intel move really sealed its fate from the economics perspective. In hindsight, it might have been a poor long term decision for Intel and other chip manufacturer to be left with a single high end equipment manufacturer monopolizing the market. If Intel signaled they would also be interested in Japanese EUV machines, we might have two competitors for chip manufacturers to select and drive competition.
@@syjiang They probably trasure more control over competition. Japan is an ally, sure, but the ties with europe are stronger and the US can control EU better, unfortunately....
Another fascinating piece, great work as always! One note: You said that the discs that picked up the molten tin were 20 meters in diameter, is that correct? If so, that’d require an absolutely enormous high-vacuum system to house them! I’m curious where the carbon contamination on the mirrors came from: Maybe minuscule amounts of diffusion-pump oil? I’m intrigued as to where the hydrocarbons might have entered the system. I was in the camera biz at the time, and the global crash of 2008 hit Nikon pretty hard too, so that might have been a factor in limiting their appetite for lithography R&D as well. Nikon’s an interesting and sad story that would make an interesting video of its own (if you haven’t already done one). A lot of things combined to create an almost perfect storm of problems for them in the 2016-2017 time frame. In 2017 they took an extraordinary accounting charge for their semiconductor lithography division, essentially saying in financial terms that they’d never recoup the capital cost of their manufacturing facilities even in terms of gross revenue, let alone profits :-0 Leading up to that, their camera division had two disastrous technology failures in 2016-2017, both of which could be laid at the feet of their development partners, but that Nikon ultimately paid the price for. The first was their ill-fated “Key Mission” action camera line. This consisted of three models, the 80, 170 and 360, the numbers referring to the angular field of view. (The 360 captured 360-degree stills and videos.) The hardware and optics of these were excellent, well ahead of anything GoPro or others had at the time, with the optics in the 360 being little short of amazing. Unfortunately, the firmware and software were another story entirely, to the point that they were almost unusable. (The WiFi connectivity was particularly bad.) I don’t know but suspect that some chunk of the software development was contracted out, but however it was done, it killed the products. This was especially unfortunate because Nikon had invested a huge amount in development, having decided that action cams were a critically important market for them to enter. They also spent some enormous amount in a global marketing campaign to launch them. It’s possible that they might have eventually recovered and gotten things working, but as we now know, this was just as the action camera market hit saturation and sales were falling off a cliff. The second critical blow was to their planned DL camera line. These were very high-spec compact cameras (meaning with non-interchangeable lenses) that were intended to compete with Sony’s very successful RX-series models. The specs looked great, and the optics would undoubtedly have been first-rate. They went through the entire development process, right to the point of being ready to go into mass production, only to discover that the (unknown) third party responsible for the image processors couldn’t deliver on them (!). They were so far along they’d even placed ad contracts for the product launch, only to have to cancel them a month or two later, before any ads aired. I never found out what the specific problem was with the processor chips, but I’m guessing that the supplier designed them around a process node that wasn’t stable yet so they couldn’t get the yields they needed to manufacture them. (This is entirely speculation on my part, all I know for sure is it was the processors that killed the line. They were originally slated to launch in June 2016, but it took until February 2017 for Nikon to officially announce their demise. Both sets of products held a lot of promise and I’m sure would have been quite profitable if they’d been able to successfully produce them. I wonder where Nikon would be today if they hadn’t lost so much investment and opportunity cost on these two products. As a longtime Nikon fan, it saddened me to see them have such struggles :-/ (The whole evolution of the camera business is a pretty interesting subject; there have been a lot of twists and turns along the way as the tech moved from analog to digital.)
I was surprised to learn ASML has really only been in their forefront position since the latter years of the 2010s i.e. about half a decade now. The way they're talked about nowadays you'd think they were this hitherto little known (cos semiconductor supply chains only became a hot topic recently due to geopolitical tensions) but key player in the global tech economy, but no they only really rose to prominence recently (rose, not founded, that happened all the way back in 1984). EUV is literally the one big thing that makes them important now, which they got into thanks to acquisitions, they weren't anything too important before it. Which also means if the technology in semiconductors changes again, ASML will be left in the dust (unless they can use their newfound position to keep up with the changing trends in the industry).
ASML was and is huge in all previous UV step and repeat systems. They just have competition in that arena. There are ASML I-line steppers where I work that were built in the early 1990s with Zeiss optics stamped “made in West Germany”. ASML is famous now because they’re the only game in town for EUV and that makes them a strategic resource at a national security level. But they’ve been doing business in this space for well over 30 years.
@@Grak70 Yeah they've been doing business since 1984. But they weren't so talked about before cos, as you say, there were plenty of alternatives. They weren't anywhere near as big either. EUV practically made ASML what it is now.
@@ArawnOfAnnwn correct, but they also ship dozens of twinscan platforms a year at all wavelengths. Most of them are I-line or KrF. Characterizing ASML as a one trick pony with EUV is just not true.
so what is the boundary between DUV and EUV? I am seeming occasional references that products on the 7nm and earlier nodes are DUV while 4nm and later are EUV
it seems that the gaps in having four entities integrating all the cutting-edge components of the EUV system defeated the Japanese before the break with Intel. One would think if they were more mature in their grasp of the technology earlier on they could have succeeded in overcoming the many milestones involved.
Hopefully US govt will fund its science & technology departments a lot more. These govt ran institutions are a lot less risk averse than corporations, so they're willing to fund bleeding edge tech that may or may not be fruitful.
@@alexmcmahon2810 then take this to your senators and lobby them to increase funding to the tech and research fields have them get spending close to that of the military budget and youd literally have warp drives and flying cars in 2 years
DOE is also responsible for technology that made the shale boom possible in the US. It's reasonable to be against fossil fuels but it did provide the US oil independence.
I keep coming to the idea that these technologies are extremely fragile and easy to lose. A little less cooperation, a solution not shared, political dysfunction, any number of phase changes that may occur, that would break our ability to create such tiny wonders. A friend stated X needed to invade Y because Freedom. When asked where cellphones would come from after the invasion. The answer given was, "Why would that change?" So much is taken for granted.
"that would break our ability to create such tiny wonders" It would still happen eventually unless we suffer some worldwide cataclysm. The nature of an international system of academic research publication is what has enabled us to gain a stratospheric increase in technological prowess since the industrial revolution. Every new innovation that makes communication between those academic institutions easier from international post to telegraphs to telephones to the internet has served to accelerate that cycle of innovation (of course the occasional war too).
Here's a reccomendation: I get topics like this in my recommendations list. I have no idea what EUV is. It would be good to describe what the video is about in the beginning.
Can i inquire one thing? Why did the EUV consortium pass their R&D to ASML? Wiki stated that ASML joined the consortium in 1999 so I am curious why the Japanese did not join as well and why the foreign participation restriction wasn't applied to ASML? Was it how the Japanese firm and trade practices was structured that limited their participation?
The more I watch these kinds of videos the more I realise just how a clusterfuck of a waste of money it was to spend 44 billion on twitter. That amount of money could have single handedly created a slew of thriving cutting tech industries practically from nothing.
It’s moments like this that make me laugh when people claim laissez faire capitalism will always properly allocate resources. It may make progress in fits and starts, but it can also be obscenely wasteful.
@@TheDuck1234 So find the technological idea and researchers wich make it happen. Chances are there that it didn't work out even several times. Some people are very good in convincing. Than this can go for a long time. I think the more wealthy you are the more contact you make with these people.
I disagree with your assessment. It has already moved the masses and broken the iron grip of globalist hegemony on public opinion that cannot be undone. In the long term, this is worth infinitely more that any losses in said purchase. It will take a couple years to bear fruit, but the change in mental direction was/is critical for the species and the planet as a whole. There have already been white papers by companies like blackrock basically forecasting the end of globalism.
That's amazing content. Wow, so Japan was close. They were the leaders in the previous technology generation of lithography. They started behind the race for EUV, had some help from the Americans, had overly complex management, and basically lost out because of market conditions and the economy. The 2011 Tohoku earthquake might have also affected this timeline as well, because of political attention to the earthquake response, nuclear disaster, cleanup and rebuilding, plus the power outages. Now it's China's turn. Can they spend this kind of time and money to catch up? Especially with no official assistance from other nations.
Yes. For China it's existential. Plus, the road to achieve EUV is already known. What China needs to do is develop the process to make all the components. It's an engineering problem instead of a research/development problem. Since US bans China from purchasing EUV equipment, there is no need to be concerned with IP issues either.
@@xuansu9036 Sadly, China never worries about IP issues anyway. Infiltrate and copy has always been the way it makes advances from military equipment to space tech and chips.
China has no choice but to make highly advanced lithography machines and other IC chip manufacturing machines by itself as it aims to be able to produce the most sophisticated logic and memory chips possible. The United States is determined to prevent other countries from supplying China with such equipment...
@@xuansu9036 Not really, The main focus Chinese EUV light source was from 2016 with DPP method this was only changed the double shot method in 2021. So my guess is the benefit of DPP is just too alluring even though existing experience has shown it to be a dead end path. Perhaps the original strategy is to bang on it while importing existing ASML machines for downstream customers, and this only got serious became more of a industrial rather than science project when China can't import any for it's chipmakers.
@@mattropolis99 I think IP is a western phenomenon. If not, does the West pay China monthly licenses for using technologies like paper, porcelain, maths, gunpowder, movable type, etc. ?
So the end of the video is all but saying that a Chinese effort to obtain EUV lithography will take 5-10 billion dollars and 10-15 years. I'd personally be surprised if it took the same amount of time. Even assuming that specific details aren't obtained through whatever methods there's at least one branch of the potential technology paths that they know they don't have to go down based on what EUV LLC succeeded with. That alone is going to shave off a decent amount of cost and time.
No, he did not say that. add the time and dollar amount to what he said next, that all the info related to said project is offered up to China by the americans and dutch. Not happening.
@@Grak70 China's Changchun Institute of optics and fine mechanics already built a prototype EUV back in 2017 similar to ASML's alpha demo tool in 2006. ASML's machine didn't reach mass production before 2019. So its pretty safe to assume that a Chinese EUV will come out by this decade.
Mice, how did pixart take over Avago and get all the mice sensors, Agilent and HP. looks like a few patents were involved. would love to know more. mouse sensor with mythic.
Very likely there will be no post EUV lithography, cause smaller you get, bigger problems with electrons migration become, making chips less durable, but with higher price. Would you like to have 1nm CPU with lifetime of 2-3 years and 1500-2000$ price or 20nm CPU with 10 years lifetime and 500-1000$ ?
@@konstantingrudnev8374 I probably didn’t elaborate that well but I meant what the future of computation will look like in the future. Silicon based processors have brought us so far, what comes next?
@@ominollo I don't know what the next breakthrough tech gonna be, but meanwhile it's gonna be hybrid approach of existing technologies. For example CPU with RAM on the same die, or CPU + AI module + RAM on the same die, this will give some boost while they seek breakthrough tech
Side note: IBM (back when they were a major fab power), bet big on X-Ray machines. The epilog to that chapter appeared in Spectrum magazine, who detailed that a visit to the IBM fab saw those machines sitting out back under tarps. End of dream.
I am very curious about the so called NIL technique that was announced by Canon. Will Canon win back the leading position by realizing and commercialising the technique? I just wonder if there's any chance that you would share on this topic
@@Grak70 I dunno..... Canon seems pretty confident and are putting a lot of faith and resources into solving it's problems. Maybe it won't work out. Maybe it will.
The monopoly in the industry could be detrimental to the resiliency of it. It's clear that these things have become primordial assets with enormous geopolitical value and there is a clear effort to prevent the advancement of alternative technologies for strategic reasons.
If we look to Shuji Nakamura who won the Physics Nobel prize for his work inventing the blue LED, it was a small team of inventors who did not have significant support who invented the tech. Large consortiums are slow and I think rarely succeed because of the internal inertia and politics. MITI should have tasked multiple teams to work on the various sub systems. Consortiums don't allow for fast decisions, etc.
These are not technologies that can be made by just one research team. It is an extremely complex set of technologies, all niche, which have to work together perfectly for even a prototype to work. Consortium is the only way. Plus developing even a single technology component is extremely expensive, and so requires governmental help.
@@slash2bot Yeah, but they still could not produce gallium nitride LEDs which are now powering lighting systems globally. 3 watt lights that are brighter than old-style >10 watt strip lights.
I'm wondering weather it was an actual colapse or a choreographed theft of an idea. Like the method for making silicon monocristals was developed in Poland. (Czochralski process)
The closing comment certainly emphasises the difficult position in which China finds itself, and the challenges it faces to develop home-grown equipment to counter international equipment sales bans
It is interesting how this critical process technology is only available to the cutting-edge fabs from a single vendor today, ASML. I am also curious about the less sexy, but also limited sources of semiconductor lead frames. Are there more than two producers left of this fundamental product to the process? Last time I knew anything about it, Intel and AMD had to go to Mitsui or Toppan, and no one else, for this component.
Zeiss technology in ASML is main component, Japanese smartphones manufacturers uses Zeiss lenses. No technological component was present in Japanese's EUV failure just short deadline, but maybe some isolation from technology was the cause of failure.
Japanese manufacturers tend to vertically integrate all processes within their own companies. I've heard that unlike Japanese photolithography equipment manufacturers, ASML is Philips-led and has excellent software collaboration with internal component manufacturers.
I'm not familiar with a .01 of this. It took me half the video to figure out that this was about semiconductor production. I have some terms to look up.
ahhh, im just a lowly mfg/design engineer for consumer light fixtures, but i do appreciate the beautiful complexity of producing things like this! maybe in a different life, i try harder... but damn, it's easy designing luxury consumer products.
Why did proximity x-ray lose out to EUV? Especially since it had a majority backing. I wonder if Japan would've had more success with PXL rather than racing against ASML.
X-Ray mask technology didn’t ever really work and the resolution of x-ray was quickly overrun by traditional UV scanners in the early 2000s so it was always going to be behind the curve.
At minute 2:18, that letter showing where it all started is a marker in history. Intel, AMD, Motorola, joining national labs, (VNL), Lawrence Livermore, Lawrence Berkeley and Sandia National Laboratories to develop EUV is compelling. ASML was brought in later. Then at timestamp 11:18 a letter granting ASML into the fold to mass produce the device with a factory also in the USA, I guess there was a reason ASML did not build that factory.
I really like your videos, my father worked for several Japanese companies, NEC, Nikon, I know he also works with several universities, etc... My father was involved in some of the processes of the machinery you mention, especially in defects ... I really liked it because some of the solutions you mentioned came from my father's group. My father always said that the Japanese were hard workers, but sometimes they weren't very smart. They had a hard time being imaginative in finding solutions to problems. My father then went to a European company, although I think he was working for IBM or Intel... and then he retired. But this video has made me very funny, I still remember my father's discussions about some of those issues and how hard it was for them to solve the problems.
Seems like most assessments concluded that proximity x-ray would have been the best, and most scalable technology. But having to build synchrotrons scared people away. I think in the end the joke may be on them. EUV is expensive, inherently not very scalable, and ultimately took decades to develop. Hard to see how the same investment in x-ray based lithography wouldn't have everyone better positioned by now...
China is developing a synchrotron radiation based steady state electron microbunching EUV, different from the Free electron laser based Laser pulsed plasma EUV used by ASML. The SSMB EUV can give a much high power radiation along with a much more stable and efficient performance at a reasonable cost. Tsinghua University along with Helmholtz-Zentrum Berlin has already verified it's proof of principle. Further research is ongoing at Tsinghua University, a prototype is supposed to come out by 2026 and commercialization within this decade.
First rule of Use Of Acronyms: at first use, expand and explain definition. I thought we're talking electric utility vehicles (which is the first result in google btw), not the lithography method for semiconductors! 🙄
But what happens if another government happened to get it's hands on Japan's EUV research? It would be a huge leg up for them & cut the time estimate significantly, wouldn't it?
Ending the Nikon program around 2011/2012 might also be linked to the enormous costs of the march 11th 2011 Fukushima nuclear disaster, effectively killing all Japanese high expense R&D programs.
@@vincentlin9350 Are you confusing the EUV machine with the Nanoimprint machine? I know that Canon is building Nanoimprint machine which can be a competitor with EUV and its is in trial run state
Do you have any suspicions regarding major "disruptive" processes or "architectures" for replacements for integrated circuits as they seemingly are approaching physical miniaturization limits?
Silicon photonics looks likely to take at least part of the market, they're already appearing in optical switches for fibre optic networking gear- czcams.com/video/29aTqLvRia8/video.html & czcams.com/video/t0yj4hBDUsc/video.html
I work for on of the support manufacturers you listed, Tokyo Electron. ASML's scanners really are a sight to behold as they're being installed, truly marvels of engineering. Our tools are extremely well built as well, and are fascinating to watch when they're working. The speed with which they transfer wafers to the many steps of the photolithography process, including the scanner, is something out of a sci-fi film.
I work for another support manufacturer, also for the wafer stage, seeing it in the testing tools is insane. Mindblowing how fast yet precise the movements are
I've been on the prototyping/building/design side of of Intels EUV Vaccum Control System for almost 8 years now, multiple revisions, tweaking, process updates, it never ends. But I'm the only one in my group that knows what the ASML Lithography machine actually is, and that I tie into that machine to make it function. All because 8 years ago I was asked if I want to take on the challenge of prototyping the first version based on terrible drawings and documentation, and I said yes, and somehow the whole project dropped into my lap instead of getting laid off when they moved production out to the new Oregon Facility. I went out there for 3 months to train my replacements, and came home to a work from home job I wasn't expecting, it all worked out in the end 🙂
Yup that they create the tiny bits that make silicon into a powerful tool in and of itself is a little left of outright magic. I can only imagine when they start making monolithic nitride devices that use photons instead of electrons. 😮❤
@@michaelkeudel8770 damn dude your life is so much more interesting than mine
Your content never fails to impress me. You have brought immense amount of information to us and amazing accuracy. I wonder how you managed to do such comprehensive research unless you have very reliable industry connection. Just want to say thank you and amazing work.
He has the connection indeed ;)
@@totenkopfgrgdfhb1336 you’d be surprised how much info you can get by just asking.
He has stated in other videos that he has a family member who works for TSMC
@@gilligan87 Oh really? That'd explain a lot :-) But even allowing for that, I'm just blown away by the depth of research and information he brings to each new video. I can't imagine how he can possibly manage to do this level of research and production and (apparently) at the same time hold down a full-time job doing something else! :-0
Depth of research is amazing 😍😍😍
Any time I hear about the EUV manufacturing process I can't help but feel that every tool at every step of the process is way too complicated to scale well. The amount (and quality) of engineers necessary to maintain even one of them would be a nightmare--not to mention the potential supply chain issues with these ultra specialized parts.
As long as it makes economic sense, they’ll find a way.
@@ClockworksOfGL I'm not convinced they will be economically viable. Have you ever seen one of these tools in person? They're absolutely ridiculous monsters.
I’m always amazed when somebody thinks the chip process is too complicated and expensive. The number of transistors in a small chip is crazy impressive. Billions. It is complicated and expensive for a reason.
Scalability by multiple stand alone processes maybe? Much like Apple did back in the magnesium case era, they stood up a few thousand CNC millls to make the cases individually. You get some economy of scale advantages by using a single brand and model of CNC machine.
@@akatsuki6371 They are economically viable. They are using this monster machines for several years now. The foundries raise their wafer prices but the chipmakers have enough margin to absorb the increase in price. It is very economically viable in fact that intel already ordered the version 2 an even larger monster machine.
Please make videos about why those EUV alternatives didn't materialize, that would be awesome. Some of them, like x-ray and electron beam, would allow to print feature sizes much smaller, right? Can they still happen?
Electron beams can already make features much smaller than EUV. Its how a lot of the masks are made. Its problem as a technology currently is that it is way to slow to be used for wafer production.
EUV had already surpassed anything they hoped to achieve with proximity X-Ray and e-beam is direct write and was never going to meet throughput requirements.
Back in the early 90's, EUV was called Soft Xray. I worked on component optics fab and test for an experimental small field exposure tool Bell Labs was developing.
@@kellensanna Indeed this has been true since long time ago - for example, I remember that already about y. 2000 ~10nm features were (slowly) printed using a modified electron microscope under university lab conditions. This was obviously way before EUV was available, but just for comparison contemporary (DUV) AT1100 machines could do 'only' CD of 100nm but with decent throughput of about 90 of 300mm wph.
Learning failures is as important as learning success.
I'd say even more so because there are typically often far more failures than there are successes. They offer far larger sample sizes to learn from. And the lessons you can learn from them are likely to be more universal than the successes.
@@Girder3 that, and with hindsight you can see why you failed but you don't necessarily see why you succeed.
Not sure this was a true failure for Nikon, it is just ASML/Intel are so in-bed with each other, penetrating the chip manufacturing market for companies that can afford $200M machines is quite limited. Although Nikon may have lagged technologically, just a bit (not much), it is the networking between the two dominate players in this Tech Space that became Nikons Achilles heel.
@@harthenry OMG it was a huge failure for Nikon. ASML/Intel need so many lessons until they succeed in penetrating
@@harthenry I'm prerry sure the lesson is, "get your shit together and gather a sufficient war chest to fund your R&D if you want to be cutting edge and an industry leader." The war chest could be any mix of cash, miscellaneous capitol assets, facilities, tallent, and customer consultants, but you have to have a sufficient war chest. Nikon/the Japanese reasearch associations just had insufficient reasources applied into beating ASML.😅
Holy crap, I never imagined the level of complexity that goes into making chips. I knew it was difficult, but never to that extent. That was super interesting video, thanks!
Should watch his other video, euv is just crazy
Yes, it's not magic however. Enough time and funding will get the technology rolling.
Mee too ,crazy chip tech whirlwind
And no mention of the extraterrestrial intelligent life really behind the designs.
いつもながら、豊富な情報量ですね。勉強になります。ところでAsianometryさんは、日本語の資料を直接お読みになるのでしょうか？
日本の半導体露光装置の研究史のような、高度に専門的な内容を、英訳された資料だけで、ここまで詳細に調査するのは難しいのではないかと推測します。
アイコンの奈良の鹿ちゃんが読んでくれているのかな？🤔笑
At the 12:53 mark. That's the MET 5 tool. I worked on both MET 3 and MET 5. Also, the exposure shots on the MET 3 were 200 x 600 microns. We exposed test patterns on experimental photoresists and provided metrology via SEM for dose and focus as well as process windows. I was there between 2013 and 2021.
I am super interested why the other alternatives to EUV failed (like proximity X-ray). There is great video covering one of the EUV alternatives: "Why 157nm Lithography failed"
I actually work directly for one of the former Intel guys who killed 157. 😆 it’s a fascinating story.
Oh I want your hair and EUV ohhhh please I need to eat hair
"there is great video"
made by this very channel?
@@NoNameAtAll2 Yes, why 157nm failed, however that wasn't an EUV alternative, it was supposed to be an easy stop gap between DUV and the next gen. But he wants info on the other alternatives that were higher on the list, like proximity x-ray being considered the #1 choice over EUV at #3.
@@Grak70 Please illuminate us a bit.
You have explained most key functions of a litho machine, but one thing I cannot get my head around is the precision motion systems inside them that have to reliably and repeatably reposition the wafer to the nm, that kind of precision is just generally unheard of.
There are 2 components: a measurement system like an interferometer and motor systems that position the stage. The biggest secret is to do this all at very high speeds...
@@tommy2cents492 The resolution of those motors and gears must be insane! with non-existent backlash to reliably hit those exact spots again and again on the wafer, especially with multi-patterning, it wouldn't surprise me if they have two modes, a low resolution mode for large and fast axial movements and an ultra high-resolution, high-precision mode for honing on the exact spot it needs to be at the end of a large low-resolution travel move
I can imagine it's also a very dynamic closed-loop system that requires constant real time micro-adjustments that might be guided by some kind of etched landmark on the silicon?
@@ChimpyChamp if someone asked me to name a technology they thought came from aliens, it would be scanner stage positioning. You look at the requirements and can only conclude it shouldn’t work but it does.
At a very simplified level, they use bands of closely spaced electromagnets to move the stage (the entire stage is levitating over a near perfectly smooth block of granite). The closed loop control uses an interferometer and powerful signal processing to feedback an electrical signal to those magnets that results in oscillation damping and acceleration control. The part of every stepper/scanner you don’t see is the rows of 2m tall control racks supporting it. They’re usually tucked away in the sub-fab, quietly and frantically sending out a flood of information to the control systems. It is truly mind boggling.
@@ChimpyChampThere are no gears.
Fantastic videos. Well done! Can you do one on the failed 5th generation computer program. I grew up in that era and it seemed that Japan would produce what we are now seeing with chatGPT and DeepMind. That never happened. It was yet another massive consortia effort
GPT is overblown in my opinion. It can't tell when it is producing nonsense, and it does that quite often.
ChatGPT is woke too.
@@BR-hi6yt No AI model is "woke".
Only the dataset is they are trained on can be "woke".
If anything actually diversity in datasets is a positive starting point, because otherwise the trained model ends up doing exactly the same things that humans do when overexposed to the same things ad infinitum.
It forms biases.
Which isn't good for a model that is going to be used by a multinational business.
The whole reason that white people from rural towns with mostly white people have trouble differentiating different black peoples faces is because our brains suffer this problem - it's intrinsic to neural networks to become weighted on their training data and form bias on those weights.
Googles Ceo Sandar explained on 60 Minutes, how after awhile the computer starts spitting out false info and garbage not fit for a black comedy.
Truly sad what perfection means, it’s all a lie!
I love your videos. I have not watched them all, but have you covered the success of ASML and Cymers' DUV contributions up to the point ASML bought Cymer in 2013?
Xray proximity's primary downfall was the mask. It was a thin membrane with Au or W absorber which was not rigid/ stable enough to meet pattern distortion specifications.
Gigaphoton. The most badass company name I've ever heard. Ever.
Right outta a Godzilla movie right?
The serious problem of Japan is that bureaucrats are totally useless. No company can get along with costly R&D itself alone... Please refrain from the 5th generation as requested below. It was an another nightmare,,.
Aye. But it seems that Intel move really sealed its fate from the economics perspective. In hindsight, it might have been a poor long term decision for Intel and other chip manufacturer to be left with a single high end equipment manufacturer monopolizing the market. If Intel signaled they would also be interested in Japanese EUV machines, we might have two competitors for chip manufacturers to select and drive competition.
@@syjiang They probably trasure more control over competition. Japan is an ally, sure, but the ties with europe are stronger and the US can control EU better, unfortunately....
Another fascinating piece, great work as always!
One note: You said that the discs that picked up the molten tin were 20 meters in diameter, is that correct? If so, that’d require an absolutely enormous high-vacuum system to house them!
I’m curious where the carbon contamination on the mirrors came from: Maybe minuscule amounts of diffusion-pump oil? I’m intrigued as to where the hydrocarbons might have entered the system.
I was in the camera biz at the time, and the global crash of 2008 hit Nikon pretty hard too, so that might have been a factor in limiting their appetite for lithography R&D as well.
Nikon’s an interesting and sad story that would make an interesting video of its own (if you haven’t already done one). A lot of things combined to create an almost perfect storm of problems for them in the 2016-2017 time frame. In 2017 they took an extraordinary accounting charge for their semiconductor lithography division, essentially saying in financial terms that they’d never recoup the capital cost of their manufacturing facilities even in terms of gross revenue, let alone profits :-0
Leading up to that, their camera division had two disastrous technology failures in 2016-2017, both of which could be laid at the feet of their development partners, but that Nikon ultimately paid the price for.
The first was their ill-fated “Key Mission” action camera line. This consisted of three models, the 80, 170 and 360, the numbers referring to the angular field of view. (The 360 captured 360-degree stills and videos.) The hardware and optics of these were excellent, well ahead of anything GoPro or others had at the time, with the optics in the 360 being little short of amazing. Unfortunately, the firmware and software were another story entirely, to the point that they were almost unusable. (The WiFi connectivity was particularly bad.) I don’t know but suspect that some chunk of the software development was contracted out, but however it was done, it killed the products. This was especially unfortunate because Nikon had invested a huge amount in development, having decided that action cams were a critically important market for them to enter. They also spent some enormous amount in a global marketing campaign to launch them. It’s possible that they might have eventually recovered and gotten things working, but as we now know, this was just as the action camera market hit saturation and sales were falling off a cliff.
The second critical blow was to their planned DL camera line. These were very high-spec compact cameras (meaning with non-interchangeable lenses) that were intended to compete with Sony’s very successful RX-series models. The specs looked great, and the optics would undoubtedly have been first-rate. They went through the entire development process, right to the point of being ready to go into mass production, only to discover that the (unknown) third party responsible for the image processors couldn’t deliver on them (!). They were so far along they’d even placed ad contracts for the product launch, only to have to cancel them a month or two later, before any ads aired. I never found out what the specific problem was with the processor chips, but I’m guessing that the supplier designed them around a process node that wasn’t stable yet so they couldn’t get the yields they needed to manufacture them. (This is entirely speculation on my part, all I know for sure is it was the processors that killed the line. They were originally slated to launch in June 2016, but it took until February 2017 for Nikon to officially announce their demise.
Both sets of products held a lot of promise and I’m sure would have been quite profitable if they’d been able to successfully produce them. I wonder where Nikon would be today if they hadn’t lost so much investment and opportunity cost on these two products. As a longtime Nikon fan, it saddened me to see them have such struggles :-/
(The whole evolution of the camera business is a pretty interesting subject; there have been a lot of twists and turns along the way as the tech moved from analog to digital.)
Nikon was a world-class business. Sad to hear about those failures and thank you for highlighting them.
Keep up the good work
Hey, where did you get the defect width / scattered intensity graph at 9:29? Would be very useful for me...
I was surprised to learn ASML has really only been in their forefront position since the latter years of the 2010s i.e. about half a decade now. The way they're talked about nowadays you'd think they were this hitherto little known (cos semiconductor supply chains only became a hot topic recently due to geopolitical tensions) but key player in the global tech economy, but no they only really rose to prominence recently (rose, not founded, that happened all the way back in 1984). EUV is literally the one big thing that makes them important now, which they got into thanks to acquisitions, they weren't anything too important before it. Which also means if the technology in semiconductors changes again, ASML will be left in the dust (unless they can use their newfound position to keep up with the changing trends in the industry).
ASML was and is huge in all previous UV step and repeat systems. They just have competition in that arena. There are ASML I-line steppers where I work that were built in the early 1990s with Zeiss optics stamped “made in West Germany”.
ASML is famous now because they’re the only game in town for EUV and that makes them a strategic resource at a national security level. But they’ve been doing business in this space for well over 30 years.
@@Grak70 Yeah they've been doing business since 1984. But they weren't so talked about before cos, as you say, there were plenty of alternatives. They weren't anywhere near as big either. EUV practically made ASML what it is now.
@@ArawnOfAnnwn correct, but they also ship dozens of twinscan platforms a year at all wavelengths. Most of them are I-line or KrF. Characterizing ASML as a one trick pony with EUV is just not true.
@@Grak70 Oh he cos you me correct ship I-line trick pony Germany cos he supply hot topic cos
Indian copes and seethes about europeans out doing asians.
At first I thought this a video about Japan making an “EUV” like the “Electric Utility Video” a la the Chevy Bolt EUV.
~ 2:00 - Yeah, i remember when proximity X-ray lithography using illumination from synchrotron was all the rage.
so what is the boundary between DUV and EUV? I am seeming occasional references that products on the 7nm and earlier nodes are DUV while 4nm and later are EUV
Damn you are CRANKING the videos out. This is one I’ve been looking forward to.
As always, impressive videos with detailed information. ¡Gracias por compartir!
it seems that the gaps in having four entities integrating all the cutting-edge components of the EUV system defeated the Japanese before the break with Intel. One would think if they were more mature in their grasp of the technology earlier on they could have succeeded in overcoming the many milestones involved.
It's always interesting to see where the US DOE pops up. Which is basically everywhere cutting-edge innovation occurs.
Hopefully US govt will fund its science & technology departments a lot more. These govt ran institutions are a lot less risk averse than corporations, so they're willing to fund bleeding edge tech that may or may not be fruitful.
@@sevrent2811 Indeed! They're amazing institutions. We should be doubling their budgets if you ask me.
@@alexmcmahon2810 then take this to your senators and lobby them to increase funding to the tech and research fields have them get spending close to that of the military budget and youd literally have warp drives and flying cars in 2 years
DOE is also responsible for technology that made the shale boom possible in the US. It's reasonable to be against fossil fuels but it did provide the US oil independence.
and yet a dutch company successfully make a EUV, not a US company, why is that?
I keep coming to the idea that these technologies are extremely fragile and easy to lose. A little less cooperation, a solution not shared, political dysfunction, any number of phase changes that may occur, that would break our ability to create such tiny wonders.
A friend stated X needed to invade Y because Freedom.
When asked where cellphones would come from after the invasion.
The answer given was, "Why would that change?"
So much is taken for granted.
"that would break our ability to create such tiny wonders"
It would still happen eventually unless we suffer some worldwide cataclysm.
The nature of an international system of academic research publication is what has enabled us to gain a stratospheric increase in technological prowess since the industrial revolution.
Every new innovation that makes communication between those academic institutions easier from international post to telegraphs to telephones to the internet has served to accelerate that cycle of innovation (of course the occasional war too).
Good to see the reference to Biolante! Not only a gifted scholar but a man of sophisticated tastes as well.
Here's a reccomendation: I get topics like this in my recommendations list. I have no idea what EUV is. It would be good to describe what the video is about in the beginning.
If you don't already know what EUV is, then your views are obviously not wanted.
Japanese have always been best at incremental improvements. That's why I love their cars.
Downside is the possibly to adopt to paradigm shifts.
😂lmao asian nzis
Can i inquire one thing? Why did the EUV consortium pass their R&D to ASML? Wiki stated that ASML joined the consortium in 1999 so I am curious why the Japanese did not join as well and why the foreign participation restriction wasn't applied to ASML? Was it how the Japanese firm and trade practices was structured that limited their participation?
In light of recent developments, do you plan to make a video on Nano imprint lithography?
"The contraption was indeed quite nifty" ... that understatement of the year was indeed quite nifty. Love these videos :)
The more I watch these kinds of videos the more I realise just how a clusterfuck of a waste of money it was to spend 44 billion on twitter. That amount of money could have single handedly created a slew of thriving cutting tech industries practically from nothing.
It’s moments like this that make me laugh when people claim laissez faire capitalism will always properly allocate resources. It may make progress in fits and starts, but it can also be obscenely wasteful.
The 44 billion didn’t disappear, a lot of other people have it now. Just try to convince them to pay for your “oh so important “ project.
@@TheDuck1234 So find the technological idea and researchers wich make it happen. Chances are there that it didn't work out even several times. Some people are very good in convincing. Than this can go for a long time. I think the more wealthy you are the more contact you make with these people.
Same
I disagree with your assessment. It has already moved the masses and broken the iron grip of globalist hegemony on public opinion that cannot be undone. In the long term, this is worth infinitely more that any losses in said purchase. It will take a couple years to bear fruit, but the change in mental direction was/is critical for the species and the planet as a whole. There have already been white papers by companies like blackrock basically forecasting the end of globalism.
That's amazing content. Wow, so Japan was close. They were the leaders in the previous technology generation of lithography. They started behind the race for EUV, had some help from the Americans, had overly complex management, and basically lost out because of market conditions and the economy. The 2011 Tohoku earthquake might have also affected this timeline as well, because of political attention to the earthquake response, nuclear disaster, cleanup and rebuilding, plus the power outages.
Now it's China's turn. Can they spend this kind of time and money to catch up? Especially with no official assistance from other nations.
Yes. For China it's existential. Plus, the road to achieve EUV is already known. What China needs to do is develop the process to make all the components. It's an engineering problem instead of a research/development problem. Since US bans China from purchasing EUV equipment, there is no need to be concerned with IP issues either.
@@xuansu9036 Sadly, China never worries about IP issues anyway. Infiltrate and copy has always been the way it makes advances from military equipment to space tech and chips.
China has no choice but to make highly advanced lithography machines and other IC chip manufacturing machines by itself as it aims to be able to produce the most sophisticated logic and memory chips possible. The United States is determined to prevent other countries from supplying China with such equipment...
@@xuansu9036 Not really, The main focus Chinese EUV light source was from 2016 with DPP method this was only changed the double shot method in 2021. So my guess is the benefit of DPP is just too alluring even though existing experience has shown it to be a dead end path. Perhaps the original strategy is to bang on it while importing existing ASML machines for downstream customers, and this only got serious became more of a industrial rather than science project when China can't import any for it's chipmakers.
@@mattropolis99 I think IP is a western phenomenon. If not, does the West pay China monthly licenses for using technologies like paper, porcelain, maths, gunpowder, movable type, etc. ?
Please slow down! You're pumping out so much good content it's unbelievable. My brain is expanding too quickly lol
very interesting (as usual) and well presented (as usual). thank you for all the work and then sharing. I learned a lot (as usual).
So the end of the video is all but saying that a Chinese effort to obtain EUV lithography will take 5-10 billion dollars and 10-15 years. I'd personally be surprised if it took the same amount of time. Even assuming that specific details aren't obtained through whatever methods there's at least one branch of the potential technology paths that they know they don't have to go down based on what EUV LLC succeeded with. That alone is going to shave off a decent amount of cost and time.
Let’s see if they build a KrF or ArF scanner worth two farts first…
China needs to build the industry to build those machines first... they can't have that as long as Communism is ruling the country
@@elchippe ok, it’s 6 years later. Where’s the production tool?
No, he did not say that. add the time and dollar amount to what he said next, that all the info related to said project is offered up to China by the americans and dutch. Not happening.
@@Grak70 China's Changchun Institute of optics and fine mechanics already built a prototype EUV back in 2017 similar to ASML's alpha demo tool in 2006. ASML's machine didn't reach mass production before 2019. So its pretty safe to assume that a Chinese EUV will come out by this decade.
TechAltar channel upload your exact video on the same day, did he have your permission? Can you report for copyright?
Your content ist awesome
You bring me great joy when I feel very down
Thank you
Be well
At 18:30 "two twenty meter diameter wheels". Is that a misspeak?
Mice, how did pixart take over Avago and get all the mice sensors, Agilent and HP. looks like a few patents were involved. would love to know more. mouse sensor with mythic.
Do a video on what happened to x-ray lithography efforts .
Fascinating story!
It would be nice to put together a video to discuss the future (post EUV) lithography 😉
Very likely there will be no post EUV lithography, cause smaller you get, bigger problems with electrons migration become, making chips less durable, but with higher price. Would you like to have 1nm CPU with lifetime of 2-3 years and 1500-2000$ price or 20nm CPU with 10 years lifetime and 500-1000$ ?
@@konstantingrudnev8374 I probably didn’t elaborate that well but I meant what the future of computation will look like in the future. Silicon based processors have brought us so far, what comes next?
@@ominollo I don't know what the next breakthrough tech gonna be, but meanwhile it's gonna be hybrid approach of existing technologies. For example CPU with RAM on the same die, or CPU + AI module + RAM on the same die, this will give some boost while they seek breakthrough tech
@@konstantingrudnev8374 yeah, this seems a reasonable scenario. We’ll see 🙂
For a split second I thought it would be a Europa Universalis video...
Side note: IBM (back when they were a major fab power), bet big on X-Ray machines. The epilog to that chapter appeared in Spectrum magazine, who detailed that a visit to the IBM fab saw those machines sitting out back under tarps. End of dream.
18:29- 20 metres in diameter? Perhaps 20cm or mm?
Why is there no explanation what EUV stands for, anywhere in the video?
I am very curious about the so called NIL technique that was announced by Canon. Will Canon win back the leading position by realizing and commercialising the technique? I just wonder if there's any chance that you would share on this topic
Imprint is cute but it can’t really overcome its throughput problem. Maybe a niche market but never gonna make consumer processors or dram with it.
@@Grak70 I dunno..... Canon seems pretty confident and are putting a lot of faith and resources into solving it's problems. Maybe it won't work out. Maybe it will.
@@Grak70it is possible to make CPU with it just not ideal
@@_Chad_ThunderCock “possible” is not necessarily “manufacturable”. If you work in a fab, you know this. I do.
Sick delivery, bro
ASMLがCymerを買収した結果から考えても、日本メーカーの光源開発が苦しくなるのは分かっていたと思う。
最先端半導体は開発費が民間企業では手に負えない額になっている。
Can you di a video on asian CNC machining industries?
The monopoly in the industry could be detrimental to the resiliency of it.
It's clear that these things have become primordial assets with enormous geopolitical value
and there is a clear effort to prevent the advancement of alternative technologies for strategic reasons.
Is there a path to X-ray lithography? Or is it just not needed because the molecules required to make a transistor already at it's limit?
If we look to Shuji Nakamura who won the Physics Nobel prize for his work inventing the blue LED, it was a small team of inventors who did not have significant support who invented the tech. Large consortiums are slow and I think rarely succeed because of the internal inertia and politics. MITI should have tasked multiple teams to work on the various sub systems. Consortiums don't allow for fast decisions, etc.
These are not technologies that can be made by just one research team. It is an extremely complex set of technologies, all niche, which have to work together perfectly for even a prototype to work. Consortium is the only way. Plus developing even a single technology component is extremely expensive, and so requires governmental help.
@@slash2bot Yeah, but they still could not produce gallium nitride LEDs which are now powering lighting systems globally. 3 watt lights that are brighter than old-style >10 watt strip lights.
10:30 I'm waiting for "NERV" to pop up...
Wow - gripping account. Thx so much.
I hope you will cover KLA Corp (REBL program) and Mapper LLC ebeam direct write lithography.
Second this in general. KLA’s story just as a company is long and interesting.
Good job with the research! I can imagine it was a lot of work.
I'm wondering weather it was an actual colapse or a choreographed theft of an idea. Like the method for making silicon monocristals was developed in
Poland. (Czochralski process)
that Godzilla is comedy gold 20:15
2:08 "DOI EUV RND project"
ha
The closing comment certainly emphasises the difficult position in which China finds itself, and the challenges it faces to develop home-grown equipment to counter international equipment sales bans
How does he get all this in depth information
awesome channel. thank you.
Could you do a video about the rise and fall of MITI?
20:19 that poster tho
It is interesting how this critical process technology is only available to the cutting-edge fabs from a single vendor today, ASML. I am also curious about the less sexy, but also limited sources of semiconductor lead frames. Are there more than two producers left of this fundamental product to the process? Last time I knew anything about it, Intel and AMD had to go to Mitsui or Toppan, and no one else, for this component.
Zeiss technology in ASML is main component, Japanese smartphones manufacturers uses Zeiss lenses. No technological component was present in Japanese's EUV failure just short deadline, but maybe some isolation from technology was the cause of failure.
Japanese manufacturers tend to vertically integrate all processes within their own companies. I've heard that unlike Japanese photolithography equipment manufacturers, ASML is Philips-led and has excellent software collaboration with internal component manufacturers.
That first image is the engine of my Toyota Tundra.
I'm not familiar with a .01 of this. It took me half the video to figure out that this was about semiconductor production. I have some terms to look up.
ahhh, im just a lowly mfg/design engineer for consumer light fixtures, but i do appreciate the beautiful complexity of producing things like this! maybe in a different life, i try harder... but damn, it's easy designing luxury consumer products.
I'm waiting for the DIY version to appear - it can't be THAT hard - lol.
Why did proximity x-ray lose out to EUV? Especially since it had a majority backing. I wonder if Japan would've had more success with PXL rather than racing against ASML.
X-Ray mask technology didn’t ever really work and the resolution of x-ray was quickly overrun by traditional UV scanners in the early 2000s so it was always going to be behind the curve.
At minute 2:18, that letter showing where it all started is a marker in history. Intel, AMD, Motorola, joining national labs, (VNL), Lawrence Livermore, Lawrence Berkeley and Sandia National Laboratories to develop EUV is compelling. ASML was brought in later.
Then at timestamp 11:18 a letter granting ASML into the fold to mass produce the device with a factory also in the USA, I guess there was a reason ASML did not build that factory.
At minute day second 2:20 timestamp a letter granting it to all start Livermore Livermore AMD EUV so IDK WTF YOU ARE TALKING ABOUT
@@TimPerfetto I'm saying it was an American invention and ASML was brought in later.
@@Erik-gg2vb OMG so it was you?
_"I guess there was a reason ASML did not build that factory."_
Probably the cost of the workers is even greater than those of Netherlands...
@@ThePowerLover ASML’s factories are in Netherlands, not Belgium. IMEC is in Belgium.
Thanks!
I really like your videos, my father worked for several Japanese companies, NEC, Nikon, I know he also works with several universities, etc... My father was involved in some of the processes of the machinery you mention, especially in defects ... I really liked it because some of the solutions you mentioned came from my father's group.
My father always said that the Japanese were hard workers, but sometimes they weren't very smart. They had a hard time being imaginative in finding solutions to problems.
My father then went to a European company, although I think he was working for IBM or Intel... and then he retired.
But this video has made me very funny, I still remember my father's discussions about some of those issues and how hard it was for them to solve the problems.
[country][tech][status] You sure do make these video titles easy to put into an excel sheet.
I love that Japan took alternative approach.
Thank you for this video
Why is there no news on e-beam lithography?
"The contraption was indeed quite nifty" :)
Very impressive video!
Thank you!
Seems like most assessments concluded that proximity x-ray would have been the best, and most scalable technology. But having to build synchrotrons scared people away. I think in the end the joke may be on them. EUV is expensive, inherently not very scalable, and ultimately took decades to develop. Hard to see how the same investment in x-ray based lithography wouldn't have everyone better positioned by now...
China is developing a synchrotron radiation based steady state electron microbunching EUV, different from the Free electron laser based Laser pulsed plasma EUV used by ASML. The SSMB EUV can give a much high power radiation along with a much more stable and efficient performance at a reasonable cost. Tsinghua University along with Helmholtz-Zentrum Berlin has already verified it's proof of principle. Further research is ongoing at Tsinghua University, a prototype is supposed to come out by 2026 and commercialization within this decade.
First rule of Use Of Acronyms: at first use, expand and explain definition. I thought we're talking electric utility vehicles (which is the first result in google btw), not the lithography method for semiconductors! 🙄
lol that biolante poster is my phone wallpaper
But what happens if another government happened to get it's hands on Japan's EUV research? It would be a huge leg up for them & cut the time estimate significantly, wouldn't it?
Next video idea :Current day leaders in xray lithography
Ending the Nikon program around 2011/2012 might also be linked to the enormous costs of the march 11th 2011 Fukushima nuclear disaster, effectively killing all Japanese high expense R&D programs.
glad they failed, the EU needs to have some champions in tech and ASML is doing that for us.
On a different note, what is your thought on a new series on *Telecom equipment giants engaged in 5G* 📱🛰️ like Ericsson, Nokia, Qualcomm, Cisco?
A source said the Nikon recently installed one EUV litho system in TSMC's factory(Hsin-chu?) for trial run.
Do you have a link ?
@@simanhtranao7859 correct the message, it's from Canon, not Nikon.....which I was just told few days ago. sorry no link.
@@vincentlin9350 Are you confusing the EUV machine with the Nanoimprint machine? I know that Canon is building Nanoimprint machine which can be a competitor with EUV and its is in trial run state
Do you have any suspicions regarding major "disruptive" processes or "architectures" for replacements for integrated circuits as they seemingly are approaching physical miniaturization limits?
Silicon photonics looks likely to take at least part of the market, they're already appearing in optical switches for fibre optic networking gear-
czcams.com/video/29aTqLvRia8/video.html
&
czcams.com/video/t0yj4hBDUsc/video.html
WOW - thanks for sharing. Don't ever count out the Japanese.
EVU = Extreme ultraviolet lithography.
This technique is used for production of high density chips, like processors.
20:50 Fun fact: siloxane is the fancy name for silicone, the stuff your spatula is made of.
No it's not🤣
@@titanicisshit1647 silicones are polymerized siloxanes. Close enough...
@@RonJohn63 so for you water is the same thing as CO2?
@@RonJohn63 it's the opposite🤣
@@titanicisshit1647whatever, Poe-boy.
Learned something new today. 😎😎😎🔥🔥🔥
I like the last statement
OMG I LOVE ALL UR VIDEOS
Chock full of valuable information. Wish you had spent 3 seconds to say Extreme UltraViolet Lithography. And maybe talk about what EUV is used for.