The Unreasonable Effectiveness of Atomic Layer Deposition

[@andersjjensen]

As a programmer your videos about the never ending toils and tribunals of the lithography world reminds me not to wasteful with system resources. It's clear that at some point we're going to hit a wall called "physics don't care about your computing needs", so while I've seen stupendous performance gains from my Commodore 64 to my Ryzen 7950X3D that party is tapering off.

[@clintcowan9424]

Moore's law prob has a limit. Quantum physics might break that physics, who knows

[@SammyGDude]

@@clintcowan9424 I think that people forget how small chips are currently. Even when we hit that limit, a double in size every every 2 years for the first 10 years will result in chips the size of an orange; I doubt that for our lifetime computing power to size to power requirements will pose an issue.

[@andersjjensen]

@@clintcowan9424 Moore's law is an observation that turned into a cadence target. Moore actually revised it twice in the early years. And, depending on how exactly you interpret it, it tapered off a decade ago. But it is precisely quantum phenomena that starts screwing things up. Since electrons have this uncanny habit of "teleporting" themselves through barriers classical physics say they can't penetrate, at a certain point the thickness of the oxide insulator, needed for ever smaller gates, becomes a problem. We're already having massive leakage problems. Once we start reaching those thicknesses that is going to get exponentially worse.

[@andersjjensen]

@@10001000101 Room temperature superconductors won't help us unless we also find lower leakage gates and a much faster switching semiconductor. And even then we eventually run into the problem that signals propagate at the speed of light (in that medium), so if we want faster... there is only smaller distances to help us, as we cant speed up light.

[@autohmae]

Depending on the interpretation of Moore's Law, in machine learning, Moore's Law is alive and well, actually, things are going faster than Moore's Law: "Subsequently, the overall trend speeds up and doubles every 4 to 9 months,” they say. That significantly outperforms Moore's Law". This is because of the huge amount of money pouring into machine learning in the past years... Those in the comments saying: quantum computers, ...: remember quantum computers are not general computers, they are highly specialized. Probably even more specialized than the hardware used for machine learning.

[@andrewpinto50]

Always nice to see the industry I’m in being represented so well

[@Facetime_Curvature]

Ditto

[@CraigChrist8239]

Were either of you the guy at 3:05 who's eyes glazed over and approved a project anyway?

[@curiodyssey3867]

You wish

[@tidel32]

Yes Yes, it has to do with pizzas and cake, don't let the narrative distract you from the images! :P :)

[@ntabile]

Instrumentation Oy! What a name!

[@KomradZX1989]

Next video should be called: “The reasonable effectiveness of Asianometry’s videos to hold my interest without fail” 😂❤ Great video man, your tickle my nerdy interests EVERY SINGLE VIDEO. 11/10

[@justinklenk]

I really love your channel. It's just a spectacularly direct, no-bullshit, yet polished conveyance of what we really yearn and love to learn. Thank you. 👍

[@robert-wr9xt]

Pure red meat or if you prefer pure pea protein. ps burn that steak good

[@FloridaMeng]

Here here!

[@AllocatorsAsia]

Praise you Jon for breaking down complex topics for simpletons like me to understand. You the real MVP

[@Nick-gj6je]

Joking aside, regardless of background, anyone with the genuine curiosity required to watch these videos is not a simpleton.

[@omniyambot9876]

This is not complex but interesting and sometimes hard to find/combine history..

[@Nick-gj6je]

@@omniyambot9876 while I agree with you, try telling my girlfriend that

[@dieSpinnt]

@@Nick-gj6je She obviously can't talk for herself. But why is she good as an example? Oh, maybe we can detect "simpletons" by the fact that they can't understand that people have different interests and EXPERTISE (PERIOD!!!!)? Btw, can you please explain to me why you feel so "special"? Combining your stereotypes with your misogynistic views and bragging about that online isn't really the pinnacle of human culture. Wait! No! You're disgusting! By the degenerated eff ... how can be the properties (if meant derogatory or not) of your partner be an answer to a question that @omniyambot9876 never asked. Sorry, I may be totally wrong. Possibly you are just joking on the cost of your girlfriend. In front of total strangers ... As......!!! ( It is unbelievable, what socially underdeveloped disgusting animals wander the Interwebs nowadays )

[@samfedorka5629]

I was waiting for the ALD episode after I listened to the Ian Cutress interview. Always good to hear more about semiconductor processing techniques that I work with.

[@dosgos]

Ian wrote some great articles at AnandTech. I just found his KZbin channel. Thanks!

[@arandomperson8336]

I love your videos but as a professional chemist I feel compelled to point out that ALD is an adsorptive process. When something is absorbed is goes into the bulk material, like dipping a sponge in water. When something is adsorbed it adheres to the surface but doesn't go into the bulk material.

[@robertb6889]

As someone who has worked directly with ALD for memory chips, you did a great job discussing the topic. And there are actually quite a few more applications and materials out there (that I probably can’t detail on an open web page due to IP concerns.)

[@scottieb6442]

ALD for thermoelectric devices and as a potential boon for advancing the refractory industry seem interesting.

[@joseestrada2542]

Lol cant believe asianometry knows about olympia tool set. Made me feel proud of the group im working with hehe. Hopefully the predecessor to olympia takes off. Wish us luck! :)

[@craigslist6988]

curious, how wasteful of the reactants are purge cycles normally and is the drum method more wasteful of the reactants and it just isn't very expensive for them?

[@alvinnorin8820]

I was JUST going to search if there were any videos on this topic, and then just before I search, I see this recommended to me; uploaded 20 minutes ago.

[@stevenjackson8226]

Super cool. I add another thanks for doing this. I saw epitaxial deposition layering way back in the early 1980s at Bell Labs, Murray Hill, NJ. I was taken through the process by the engineers producing GaAs substrate FETs for use in Hughes Aircraft satellites. It's amazing how much further this has improved and what can be done today.

[@stevesloan6775]

One minute in and I’m already refreshed how you paid respect to the people that first had the concept. So many videos on KZbin miss out the most critical information. It’s seems so disrespectful to me. Thank you. 🇦🇺🤜🏼🤛🏼🍀😎

[@christopherleubner6633]

Man a lot has changed since running CVD machines in the nineties and early 2000s. Used to make arsenide laser chip wafers. Once the trimethyl aluminum injector misfired adding two extra shots to a batch of 635nm laser wafers and since it was for the active layer we went with it to see what would happen. At first we thought they were bunk but when we lowered the temperature of the test die to about 15 degC we got pure orange laser light at 612nm. If you got it extremely cold like -77 deg C the wavelength dropped to 602nm which was bright orange.The place was Polaroid lasers then Boston lasers if you were wondering ❤

[@tylermyers3106]

Excellent video as always! I love that I work in ALD. It's an elegant technique with seemingly endless applications, and the community is one of the best scientific communities I've had the pleasure to interact with. I run a podcast interviewing ALD folks, called ALD Stories. Some excellent people to listen to!

[@jmlinden7]

9:18 source-to-drain leakage is a different phenomenon than gate oxide leakage. If you use the light switch analogy, source-to-drain leakage would be the amount of electricity flowing into the lightbulb even when the switch is off, while gate oxide leakage would be if your finger could zap electricity into the lightbulb when you touch the switch. A better designed gate oxide helps reduce both of these leakages though

[@adambarker3130]

Excellent, as ever. I was expecting a mention of a much older technique for molecular deposition : Langmuir-Blodgett layers. It was started off by a German teenager (Agnes Pockels) doing beautiful experiments in her kitchen and finally sending her results to Lord Rayleigh, who got them published in Nature. It's a great story.

[@lindsay.newman]

en.wikipedia.org/wiki/Langmuir-Blodgett_film

[@e1n17g13l1i14sh]

Cool story, thanks for sharing!

[@JudgeJor]

Good stuff! It could have been emphasised that ALD specifically uses self-limiting surface reactions. This means that each cycle creates precisely a certain thickness of film and that the deposition is extremely conformal. These characteristics are especially useful for high aspect ratio structures or cavities (which were discussed).

[@richardwhitehouse8762]

I just wanted to say, as someone who really didn't pay enough attention to chemistry and physics at school but who is interested now in how things like semi conductors work, thank you for producing this amazing content. It gives a complete non-techhie like me a chance of grasping something about it. So, many, many thanks. PD I loved the bit about the Finnish Dr looking at the periodic table for inspiration. Its a bit the same as when as a musician you look at a score and find the underlyning patterns.

[@fiasco2003]

I really enjoyed "The Unbearable Lightness of Being". The sequel has been a long time coming. So I hope that my expectations for "The Unreasonable Effectiveness of Atomic Layer Deposition" are not unwarranted.

[@yyyyyeeeee4060]

It's a reference to a famous 1960 article titled The Unreasonable Effectiveness of Mathematics in the Natural Science - highly recommend a read.

[@stevengill1736]

OK, cool, but is there really a sequel? ;^[}

[@fiasco2003]

Aha. That oblique reference went straight over my head. Thanks!!@@yyyyyeeeee4060

[@lawrencefrost9063]

Ey it was invented in Finland? In my location? This makes me happy. Great video.

[@Inkkari9]

Same - sama :D

[@mdharrisuiuc]

I used to manage an ALD about a decade ago. They're really simple to run and maintain, the only real issue being sustained chamber temperatures causing the chamber o-ring seal to degrade over time. Periodically testing the films (we were primarily doing Al2O3 with trimethyl-aluminum and water) for carbon contamination in a ICP-MS is a good check to replace the seal, although it is easier to just replace the seals somewhat frequently.

[@EkiToji]

Well pump maintenance can be an issue because even with pretty good filters or trying to thermally decompose the precursors you can still end up with deposition on the pump. Granted they're not even particularly expensive pumps since you don't need or even want high vacuum.

[@chillphil967]

love your channel. just added the “bell” to my notification settings. thx from detroit 🤙

[@dadawoodslife]

I remember working at Nottm University, Physics Dept in the 1980s when they were doing MBE. I never know how groundbreaking it was, this just made me realise. The same department also invented NMR imaging, what we call MRI today.

[@yuema2078]

Good video as usual. A couple of suggestions: add some equipment suppliers information will be nice (ASM, Beneq, Picosun, which was bought by Applied Materials). Also ALD is a self limiting process therefore it does not have deposition uniformity concerns as in PVD or CVD. This unique feature could be mentioned as it gives ALD a huge advantage over CVD. Lastly, on spatial alternation method, it does not have to be rotation drum + multiple chambers - there have been single chamber linear systems used in the industry for at least 7, 8 years. An example is PERC process in thin film solar panel manufacturing equipment, where substrates move back and forth in a liner chamber with zones separated by purge gas curtains.

[@alfredcpr837]

Hello Yuema, we see many different ALD players in the industry (Also LAM, Eugene...) but can we assume that ASM and Oxford Instruments own the best solution (Aspect Ratio / Deposition Rate / Step Coverage) ? Because every time you listen AMAT, they are pretty confident to their ability to gain market share, that's why they bought Picosun right ?

[@mikebarushok5361]

I remember when the higher quality and less expensive electroluminescent panels replaced incandescent illuminated switch and circuit breaker lighting in general aviation aircraft. I never heard anything about the technological break through required to make them practical.

[@hansbleuer3346]

Hut ab. Didaktisch hervorragend gestaltetes Video.

[@tykjpelk]

Oh my god I love ALD almost as much as I love atomic layer etch. There are things I want to do with this technology that were discarded by better nanotechnologists just because it's too damn slow for the scales needed by photonics. Still hopeful.

[@MarcCastellsBallesta]

Your video would have been so helpful when I was studying nanomaterials back at university... Really well done!

[@teddy0819tw]

I love all these series. Will there be another nice introduction on atomic layer etching?

[@sprk373]

He "kind of" looks like Colonel Sanders? Thats a 100% match in my book. Great job dude as always

[@woolfel]

another good video explaining some complex stuff.

[@da1otta]

It never ceases to fascinate me what humanity is capable of achieving.

[@brodriguez11000]

World peace.

[@artbanks27]

I was wondering if this was even possible I'm glad it is. Thanks for your explanation mate.

[@hai-duynguyen8429]

This speaks to me on a number of levels. I don’t work on ALD but epitaxy. I can appreciate the difficulty of nano scale growth

[@AaronSchwarz42]

Most excellent description of ALD with graphics professor John :)

[@mururoa7024]

Thanks for breaking it down for us in such simple and understandable terms. 👍👍

[@redlogicsquare]

Valentin Borisovich Aleskovsky is also winner of most Russian name award.

[@craigslist6988]

just FYI there is a difference between absorption and adsorption, and this is one of the times where it's useful to know the difference. The molecules in ALD are adsorbed.

[@zallen05]

I would love to see you cover PVD as well.

[@drmosfet]

When seeing your thumbnail photo I thought this might be about taking that electron tunnelling microscope technique of making atoms sized greeting cards to the next level in manufacturing. But still very interesting.

[@GodmanchesterGoblin]

Candidate for KZbin "Video Title of the Year"! And a very informative video too.

[@KevinLyda]

Wild, I didn't know chip makers and lawyers both depended on depositions...

[@SUNRISE-ADVENTURES]

TOP NOTCH WORK!!! This stuff makes my brain hurt...LOL

[@largezo7567]

I'm a Finn, but I had never heard of this remarkable invention before.

[@tylermyers3106]

Tuoma Suntola earned the Millennium Prize for the ALD invention!

[@brodriguez11000]

Interesting how the inability to patent and hence facilitate a transition from academia to commercial made all the difference. Something some in the states complains about because all academic work should belong to the public.

[@MarcCastellsBallesta]

​@@brodriguez11000 I had a boss years ago who told me that the technology transfer from academia to industry is called is called "The valley of death". Many industries have needs academia has already solved, but either they don't communicate enough or there's a scaling problem.

[@xanovaria]

Finn Fett, Boba Fett’s silicon brother.

[@aloksharma4611]

Thanks!

[@usercid640]

Man, you're learning more than me about deposition

[@nayhem]

Cake in an ALD demonstration, but no mention of flouring the pan. Those first layers of grease and flour can literally make or break the cake.

[@cpt_bill366]

They missed the opportunity to name a company ALD Incorporated, or ALDI

[@miinyoo]

I'm a little surprised no one has yet figured out a reliable way to "air brush" atoms onto a substrate. Of course I don't mean using actual air as the motive medium since it has to be done in as near perfect a vacuum as possible. Argon might be too big of an atom to use so Helium seems a superior option since it can escape minute barriers. Pure H2 could be cheaper if the substances aren't proton reactive at the temperatures of adequate deposition. Either way, one would need to ensure the "bubbles" are so small that they don't affect the final product. Sans that, annealing can help remove the nano-cavities created by the distribution "air". Or, as is done with particular plasmas, an electrostatic "air brush" working with ions. Tune the power down so instead of crashing into a substrate displacing other atoms, they land softly accompanying a neutralizing spray of electrons to lock it all in place. Purge, rinse, repeat in very much the same idea as creating seasoning on a cast iron pan. Of course I may be way off base since we're dealing with so many variables on an atomic scale most of which aren't relevant once you get to the micron scale.

[@mduckernz]

You could consider the plasma technique mentioned to be roughly equivalent

[@al4x5]

I really like your videos. I think they are always very information rich and easy to understand. Could you make some videos about different wafer bonding techniques one day perhaps?

[@LuciFeric137]

Love these historical looks

[@TomLynchIsAwsome]

BANGER VIDEO

[@benjaminlynch9958]

Dude, 2 months ago?!?!

[@pamir8337]

Very informative video, thank you!

[@KepZubel]

I am assuming you will release a video on Huawei’s 7nm chips, right?

[@Jaxon220]

I work for a company that makes deposition equipment and I’m only slightly disappointed I didn’t see one of our systems in this video 😢

[@EkiToji]

At around 5:44 you mean to say adsorption and not absorption. Absorption has a fluid permeate into a substance but adsorption is surface adhesion.

[@FeintMotion]

she layer on my deposits until I atom

[@BlahBlahBleh317]

The only KZbin channel currently worth watching!!

[@robert-wr9xt]

Mass content week. Awesome.

[@MrSuperBrite]

Sun in Suntola is pronounced the same way Sun in Sun Yat-Sen.

[@ag3127]

Can you talk about the chip used by Huawei 60 Pro - who made them?

[@davieb8216]

The stock footage you used at 6:20 didn't demonstrate your point very well. The first pour looked smoother relative to its thickness.... Anyway thanks for making this stuff interesting.

[@jaakkopontinen]

How does the application of a layer reagent end up as 1 atom thick on the substrate, or on top of previously laid layers, in the first place? What's stopping n^x atoms from piling up before the purge? What did I miss? Thanks!

[@brandonpalafox8932]

The chemical commonly used is HMDS. Which has a byproduct of ammonia so we gotta test for amines after application.

[@budlaumer]

Thanks for making me think.

[@KomradZX1989]

Wait so they were working on flat panel displays way back in the 1970s?!?!?! MIND BLOWN 🤯

[@nexusyang4832]

ngl, i think i may have to listen to this lecture a few times to sink in.....

[@temperultor7602]

Will you be doing atomic layer etching next?

[@emeraldwind]

Are you going to release a video about Huawei's new 7nm phone?

[@frittsdonoghue478]

Easily your best titled video

[@cosmicmuffet1053]

We have reached the edge of comprhensable limits.

[@KokkiePiet]

Could you do a video on optical coatings, like on lenses or heat reflective windows?

[@tomholroyd7519]

as they shrank, they shrunk

[@quark31]

What is the role of ASM Intl here in ALD? How do you view it?

[@fundingsecured7072]

Whats your take on Huawei's new CPU? Can you make a video about it?

[@SixTough]

So the layers just adsorb on the substrate? Holding on with hopes and dreams?

[@alexandruraresdatcu]

I would be a bit skeptical about russians claiming to have invented something a couple of years before someone actually getting a patent for that process....

[@awah4676]

I really have no use for this information... But I'm gonna watch it anyway

[@Renormal]

I operated an MBE machine during my internship.

[@ibnu7942]

i don't understand a thing but i still watch this till the end

[@brodriguez11000]

Like creating a fancy cake.

[@badrinair]

Thank you

[@AaronSchwarz42]

Molecular precision manipulation of atoms to make very high precision products like chips or chipsets or IC or integrated circuits (eg ECU, GPU, SOC, Sensors, ETC) // the most complicated & complex products made!

[@werre2]

The company is Väisälä

[@mefobills279]

The substrate has to be hydrogen terminated before dep. TSMC especially has tried to poach or work around my patent.

[@craigslist6988]

what patent is that?

[@mefobills279]

@craigslist6988 low temp oxidation of silicon using chemical oxidation and etch, then terminate with hydrogen. Basically, if the surface isn't perfect, then the dep doesn't take. The oems for ald especially want it. Doesn't matter. When you are an employee, you sign your inventor rights away. I did have to defend the invention, even to the point of having my logbooks examined.

[@craigslist6988]

@@mefobills279 that sounds very basic though.. guessing it's the exact conditions and chemical. Still, wouldn't this patent be out of protection by now?

[@mefobills279]

@craigslist6988 I was vague. Yes, it gets more complex, using ozone, etc. All invention can be reduced to some simple concept. Still active patent. Generally, big companies cudgel each other because they mutually poach. This is why Google bought Motorola to then move in on Apples I phone turf. The Chinese, especially poach. You have to sue, and that takes big bucks. Like Elon says, patents are a cook book, yet Tesla still gets them.

[@craigslist6988]

@@mefobills279 all good. And I don't know of anyone who would file for a patent to protect IP if they thought they could keep it as trade secret. Even patents are not always a cookbook. I have mostly patents that are useless for actually making the things, because key aspects of the method to apply the ideas are left out as trade secret. I don't care to look into Tesla's patents but design, especially mechanical, is impossible to hide so I could see patenting novel inventions there. I wish the patent system was somehow more effective because I know a lot of the details get reinvented as trade secrets in many companies and all it serves to do is let large companies monopolize it. Or in some cases they just invent it and then sit on it until it gets forgotten, even the inventor cannot tell anyone or use it because as you said all companies require full sign over.

[@dndboy13]

look the topic is interesting and all but there's a part of me deep down that's thrilled about a Soviet Commissar Sanders lookalike.

[@JamesOliverLindsey]

why do the atoms want to assemble DNA strands etc? does everything just boil down to some 1s and 0s math equation that defines all interactions?

[@LorenStClair]

THX great topic

[@mennowitteveen3313]

I am testing timestamp comments. 6:31

[@williamwoosleyiv6150]

modern lithography probably has some odd connection with the inception of lasagna. or some other more ancient layered recipe, with an understanding of science/engineering LOL.

[@AlexGheorghe]

waiting to hear the credits just to learn about the fate of hair and makeup!

[@skaltura]

LOHJA: Cool how close you came, but the J is hard, as is the H. loHJa is closer on the pronounciation. Still, good effort! :)

[@zailoleite6770]

I want the recipe for that starfish pizza and I want it now

[@BobBob-nr1zt]

When paired with "have", "shrank" becomes "shrunk". :-)

[@kevin_6217]

Thicknesses is thickness, absolute. Thinness is RELATIVE, it's a comparison.

[@str0fix]

It’s effectiveness is not unreasonable. The only thing behind this is a reason.

[@Asura_vithu]

Seems like you're philosophy buff as indicated by your title

[@sayrerowan734]

It's shrunk, not shrank, in that instance. But a very informative video, as usual.

[@d.jensen5153]

Past participle vs past tense.

[@werre2]

TORILLE!

[@mansurtxafapapaias3517]

The planes fly due to a inestable system. The chips works due a inestable & doped systems.

[@hurrikkkanes2533]

hail the soviet scientists! the japanese also “discovered” euv from soviet scientific papers!