@@Asianometry I happen to have a crucible with molten eutectic tin+lead to help when I must "tin" a lot of wire ends. I'd have to be very very tired to mistake it for an ink reservoir.
Its hard to think that they could have kept going after RAM and Electroma come out. That really seemed to be it for them. No more nostalgia for the duo as just that.
my father designed machining for intel for board assembly and placing silicon wafer. said at the time it was almost impossible to do at the scale and had no clue how it pulled it off. He didn't even understand the engineering he did, but it worked. This was in the 90s. He still keep parts of the prototype around. Kinda cool.
@Dr. Irina Luminesk "without knowing how they track us." pretty much every web developer, of whom there are literal millions, know how facebook tracks users, it's nothing marvelous. Bosses of millions of different websites that help facebook track your presence literally ask their hired web developers to plug in tracking scripts from facebook. It all kinda started from the Facebook like button that at some point had spread the internet. Even though this button disappeared the scripts are still used only that more often now they are hidden. Website owners plug these trackers because it helps them make few extra bucks from advertising partnerships.
I used to work for a semiconductor company. And was once given the opportunity to visit Shin-Etsu in Japan and see their production line for SOI wafers. That was a great opportunity as I have now seen the entire manufacturing process: seeing the production of wafers, visiting a wafer fab, and have visited assembly and test sites.
"Paper is like an iPad that doesn't run out of battery." Wow, that's amazing, I wonder why Apple makes iPads instead of paper when it doesn't run out of battery.
The problem with paper is that it has a high per-frame cost as it is single use only and that it has a high display latency when a printer is used to update the screen or when a the display is manually scribed with a piece of graphite.
Thank you for this. I finally got an answer for why wafers were circular if it was obviously more efficient to print square chips on a square substrate
It would be less efficient actually. The optics is round, so you would just lose part of its field, not get. Square substrate would be not only more difficult to get and process, it would be impractical for cleaning, where you need to rotate them. The corners would be more susceptible for damage and cracks. Square substrate is bad idea on so many levels.
You did a great job describing the Cz single-crystal-silicon ingot pull ;) I worked at ARCO Solar a few years, then Siemens purchased ARCO. People don't realize or appreciate the amount of energy and effort it takes to grow an ingot. The melt of silicon on the crucible is rotated say, CCW while the seed is rotated CW. The seed is oriented a specific way ( the term 1-1-0 rings a bell ), and as I recall 4mm x 4mm x 100mm. I was there when we changed from ID saws (more kerf loss) to wire saws (the carbon slurry was with Mobil One) We could cut 4 ingots on one wire saw, yielding about 800-1200 wafers per 36 hours where the ID saw took about 48 hours for one ingot yielding about 200-250 wafers. We made 105mm wafers (as I recall), then 125mm after I left.
Gallium Arsenide (GaAs) were considered to be the future in the 1990s. They are still around as a niche product. Another niche is Silicon Carbide (SiC) for power electronics.
And now molybdenum disulfide is showing up for power electronics, though currently on smaller stuff like 45w chargers. I don't think any cars or chargers for cars or forklifts have used it yet. There's potential to make it useful for logic, but there's a lot of work to be done in order to make it so.
@@Jaker788 are you sure about that? Power semiconductor technology is my field of expertise and I have not heard about any research into MbSO2 I would be interested if you can point me towards any research you are aware of
I worked in the 90's with both GaAs and Si. GaAs logic devices were much faster but the yield rate was diabolical. If I recall it was as low as 50%. That and material cost meant it was just far too expensive to be used for commercial grade chips. The metallisation layer we used was gold, so no doubt that made GaAs more expensive. I remember doing one design and the engineering samples (a small pre-production quantity) cost £1000 each, this is 1990's prices.
You might not believe this but I worked at a company in the 90's that got a job from Sanyo Solar in Carson Ca. (LA county) to make ingot transfer carts for ingot right out of the growing machines. They were lined with 1/4" (6.25 mm) kelvar blanket material. They made Solar panels here in the US! Or at least the silicon wafers. That was a different era. Paying people a bowl of rice for a days work became the trend.
Let's say someone starts a silicon ingot production company and pays their employees rediculously generous salaries They pay employees $1 billion in total per ingot produced. Of course you can't sell that ingot for less than a billion if you plan to recoup the costs of the labor that went into making it...or you'll go bankrupt, and everyone loses their jobs. Let's say a separate company buys those ingots and makes wafers from that ingot, and also pays their employees a billion in total per batch of wafers produced (from 1 ingot). So now that batch cost 2 billion (in just labor). A fab buys those wafers, and makes micro chips from the wafers...adding another billion in labor costs to that batch. Then you see the CPU in a store and complain that it's too expensive priced at $3 million each. How do you suppose we make that CPU cheaper for you?
He did poczochrał Czochralski name and made Czouchralski out of him. :P Gonna bet that he used google translate to read it because google translate is reading it as Czo-chra-lski, voice sounds like a native Polish speaker but in 3 separate files. btw. "czochrać"= rub, tatter, tousle, scratch, ripple. But most of the time you can hear it in Poland via word "poczochrany"(for man) or "poczochrana"(for woman) when someone is trying to describe a mess of hair, so i am not surprised that Czochralski invented his metod by messing of the pen...
I actually found a YT video of Polish guy to figure out how to pronounce his name correctly: ruclips.net/video/Qg48Xe6U5qs/видео.htmlsi=FUDlfES9ilYAgFEh&t=85
Jan Czochralski is currently the most cited Polish scientist all because of an accidental revelation. The method is probably semiconductor's equivalent of penicillin and just as much crucial.
Hi, enjoyed the Wafer video. I worked at a (at that time) Monsanto Silicon Wafer plant in Spartanburg SC. we built a new plant there and made 4" and 5" wafers. At that time (early 80s) that was about the largest being produced commercially. I actually have quite a few samples, lapped, polished as well as a short piece of rod and some chunks of polysilicon from a quartz crucible, actually have a couple of those also. Was a great business, but moved on in my career to other businesses.
I'm an old American koot of 62, I was working in the burgeoning digital and mixed signal industries in California in the late 70s though the 90s I enjoy this channel greatly. It saddens me to see where we have fallen to in America even with the luxuries we now enjoy from the brow sweat of our patriarchs
Growing up my mom used to work in a fab in SoCal and she would sometimes bring wafers home and show me what it is she did for work (I was a kid, had no clue wtf she was saying it was way too complicated). I'm guessing the wafers she brought me were faulty orwere thrown out for QC issues since they seem kinda expensive. Her company (Microsemi) was snapped up by an Israeli firm and she switched industries shortly after. I have so much respect for her, can't believe humans went from cavemen to turning sand into these hypercomplicated semiconducters. If you tried explaining this process to someone not too long ago they would just chalk it up as being magic.
It turns out we are so good at fabricating silicon, and have such a close understanding of the properties and isotopic distribution of it, that we now use it as the physical translation of the new SI kilogram.
Silicon is awesome, and there was a project to make an exceptionally round 1 kg sphere of isotopically pure silicon. But that is not how the kilogram is defined. The modern definition of kilogram (in force since 2019) is based on "Kibble Balance" -- a special instrument which reproduces kilogram based on the value of Planck's constant and other fundamental physical units defined through their respective standards.
@@cogoid sorry, I was less than clear. I am aware that the kilogram is defined by Planck's constant, but i was referring to how we have used our exceptional knowledge of silicon as a part of checking that process.
This probably won't be your most popular video but its incredibly important. I'll have to do some reading on US domestic silicon boule production. It seems odd that we would be financing a $12bn TSMC plant in AZ while also having the source silicon produced in the US without producing 300mm wafers here as well (we very well may, I just don't know). If you're running low on ideas, electron beam lithography has been able to produce nanometer(ish) feature sizes for almost 30 years. I know the DoD has been funding it for at least that long but I don't know what secret squirrel types of products have been being produced by EBL. -edit Oh, also, yeah... Graphene. A lot of people (not you, other people) make the mistake of thinking a good semiconductor is automatically equally good at things like forming insulating layers or accepting metalization layers. Decades have gone into developing dielectrics and processes explicitly for silicon that are completely incompatible with graphene. Lets go full TL;DR. Graphene might be another good topic. I'm assuming that semiconductor grade graphene has very little to do with the idealized graphene sheets people think of. I would think it would have to be at least cross linked graphene. I don't think you can just stack a bunch of graphene layers on top of each other and expect them to stay stuck together through even just operating temperature differentials. Also: high-k dielectrics. That was on the radar like a decade ago at the 45nm node, but everything seems to have gone quiet. I don't think its because things are actually quiet though. I think high-k dielectrics are now very highly guarded industrial secrets.
The sheer complexity of all this going into the first IC’s (nevermind subsequent advances) makes me terrified of how much knowledge would be lost if there was ever a significant disruption to civilization. Back to boolean gates and relays, assuming you even have enough population density and law and order to justify the effort.
Relax -- look for "I, Pencil" by Leonard Reed (I think). A page or so, and there is at least one video of it. It's all about how no single person knows how to make something as simple as a pencil, once you get into the chain saws to cut down the trees (who makes them?), the mines, the ships to transport everything ... in effect, if civilization collapses enough to lose the knowledge of how to make integrated circuits, it will probably also lose the ability to make even pencils. No halfway collapse, it's all or nothing. There was one person who actually did try to make a pencil from scratch. I think it took him about 8 months, and that was with a lot of short cuts. It's a fascinating study in what we know.
Great subject! As a fresh grad, I worked in diffusion process before moving on to device fab, this video is a great primer the covers the fundamentals.
In the category of alternative materials to silicon, unmentioned are diamond (carbon) and gallium arsenide. There was also a vogue for sapphire as a base material (coating of the wafer prior to patterning), which if this old man recalls was SOS or Silicon on Sapphire.
Awesome technical video! You explain the subject very well. You are never boring, and your pace and tone are incredible. I am surprised Bloomberg hasn't snapped you up yet. Thanks for all your hard work. The research and time to make and edit the video should be applauded. 👏
Im following you for quite a while now. Your way of presenting is very unique. You aren't overly enthusiastic (or actually almost not at all) but especially combined with your writing, it just works. The editing, while pretty much basic, also fits your style very well (pls keep including those bad jokes, in writing and in picture x) )
A nice review of the type of place I worked at as a Electrical Maint. Tech. in the late 90's. All of our stuff was 200mm but there were a couple of experimental 300mm pullers in our building being run on and off at the time.
The silicon demand problem isn't solved by graphene either. Graphene nanoelectronic devices still require a silicon substrate. Graphene is useful when it's as flat as possible, silicon, more specifically silicon dioxide, is not flat, but provides a good starting point. Atomically flat flakes of other crystals, Hexagonal Boron Nitride (hBN) being one, are placed atop the silicon substrate. This is achieved by peeling off thinner and thinner layers of hBN crystals from a larger parent crystal. Graphene is then transferred onto the hBN using a similar method. On the hBN, graphene can lay falt and it's tremendous properties can be utilised. So although graphene is an interesting and exciting advancement for nanoelectric and spintronic devices, we're a long way from a silicon free world. Silicon itself is too versatile a substrate and the manufacturing processes are too well developed to transition to other materials just yet. If you're interested in more, check out the Graphene research group at the University of Manchester. I used to work in the condensed matter research team there under Dr. Ivan Vera Marun. My research involved investigating magnetotransport and thermoelectric effects in ferromagnetic channels on hBN substrates. It's cool stuff!
Given the amazing quality of this channel, I am utterly mystified as to why it isn't vastly more popular than it already is. This is prima facie evidence that most homo sapiens are basically stupid.
I wasn't paying attention for a short time until I heard you say, "paper is like an iPad that doesn't run out of battery!" You get a video like from me for this.
You left out Gas Plasma Vapor etching. It was big back in the late 60's and 70's. It cleaned, polished and scribed lines that allowed the wafers to be snapped without any substantial damage. I would say continuous improvements in the etching process is what has driven the industry forward.
I disagree. Etching is just one small part of the complete manufacturing process of integrated circuits. The thing that has really made the difference and enabled more transistors to be integrated in to a "chip" is the photolithography. The etching is the next stage of processing after photolithography. The achievements have been in going to smaller and smaller wavelengths of light in the photolithography and still assuring the image definition when the mask pattern is projected onto the etch resist coating covering the wafer.
I read the title as Hubble Silicon Wafer and waited all the way through your interesting presentation and when you said that was it, I had to check the title again and realize my mistake. Thank you
ABSOLUTELY MINDBLOWING! What's even more amazing, considering the modern miracles of science which goes into creating these microchips, is that one microchip doesn't cost as much as a house. If you were to travel back 100 years, you wouldn't even be able to explain the concept of a microchip to another person. It would have been far beyond what even science fiction was capable of imagining. To the untrained brain, mine for example, just one microchip is as mind-boggling as the depths of the universe.
I think it would be interesting, in future, to have a similar video, but on GaAs or other III-V wafers production. P.S. I enjoyed the primer's humor for gen Z of what a pen is.
I kinda wanted to know pricing on like a wafer, or one of the silicon crystals. I was looking to redo the bannisters on my staircase and silicon crystals would be a perfect fit.
How your channel don't have more subs, is beyond me! You do amazing work over and over! i feel like i even wrote you this before! But was true then as well!
I'm curious: is there a reason why the wafers aren't cut at an angle instead of perpendicularly to the growth axis? You'd get fewer but larger wafers. i donno if the tradeoff makes sense but considering most of the the losses happen at the edges it might make more sense to have a larger area on fewer slices. maybe the crystal orientation limits how you can cut it? you could also cut it lengthwise to get more rectangular wafers which would be a lot more suited for rectangular chips but each slice would get smaller in width from the largest in the center. again maybe more efficient use of the wafer surface would offset the annoyance with having to handle larger wafers of varying sizes? maybe they could all be cut into strips of more uniform width to make further processing more consistent?
Surely they have done the numbers and practical questions and answers already. Every step in the process is optimized. And one of the ways to even start getting anything useful out is to create one process that exactly can be replicated for multiple wafers. Starting to cut diagonally makes the shape not round. Edges and machine to cut it has to line up and be tailored etc etc. Ones getting a cylinder shape they just cut the thing up into as close to perfect equal wafers. With as little waste or defects as possible. Really 350mm should be more then enough. There is so much software optimization and changes needed to our lives that really it might be a good thing to stop complexity from growing and instead make sure stuff works well and efficient. The only people in need for faster chips really are the ones running A.I and in general automation to track and make money out of peoples private life. And instead of accepting that maybe just maybe focusing on software and sustainability is more important then making wafers even harder and more complex to make. Like get rid off the people and machines (and the knowledge) in the few places making this wafers? Ooo bad times. Might it be better to make sure we are not trusting 2 places in the world to sustain basically all our wafer supply??? Instead of wanting it to be even more complicated to produce?
I would like a video going deeper into the silicon used for solar panels, developments in that field and the differences with electronic grade silicon. Kinda like a side video to this one.
Hope you enjoyed the video. Check out other deep dives here in the playlist: ruclips.net/p/PLKtxx9TnH76RiptUQ22iDGxNewdxjI6Xh
I just want to thank you for yet another very interesting video. 🙏
@Asianometry you seem to have found the right idea for your april 1st video :-)
man this is epic now I know how cpu is so important
@@masternobody1896 and you haven't seen much yet... It's barely blowing on the surface.
@@leyasep5919 lmao then tell me
I hate it when I accidentally dip my pen in to a crucible of molten tin
Happens all the time
@@Asianometry I happen to have a crucible with molten eutectic tin+lead to help when I must "tin" a lot of wire ends.
I'd have to be very very tired to mistake it for an ink reservoir.
Czochralski czochrał his pen in the wrong crucible... nothing out of the ordinary for him... 😆
8'957'656th invention due to a mistake.
'make it look like an accident' :)
0:06-0:09 are the best 3 seconds of a technical video ever
I want to see him rap in his video.
0/10 video needs more demonic mechanical babies.
Harder, better, faster, stronger. Can't believe they split.
:(
Damn dude why would you remind me out the blue like this.
One more time, a celebration
You know we're gonna do it right
DID THEY??
Its hard to think that they could have kept going after RAM and Electroma come out. That really seemed to be it for them. No more nostalgia for the duo as just that.
"Like all minerales and people, quartz is influenced by the enviroment in which it's grown". So double true. 👌
Deep
@Dr. Irina Luminesk I freaking died. Absolutely lapped up whatever the hell that was.
@Dr.Kay_R Ikr, paper runs on green energy.
Hank Schrader approves.
as someone working at ASML I scream internally when I see those wafers out of a cleanroom and someone touching them barehanded.
but they look SOOO GOOOD !
anyway I feel for you.
Should I invest in ASML?
@@AndrewRosales-y1hRespect the hustle
my father designed machining for intel for board assembly and placing silicon wafer. said at the time it was almost impossible to do at the scale and had no clue how it pulled it off. He didn't even understand the engineering he did, but it worked. This was in the 90s. He still keep parts of the prototype around. Kinda cool.
@Dr. Irina Luminesk "without knowing how they track us." pretty much every web developer, of whom there are literal millions, know how facebook tracks users, it's nothing marvelous. Bosses of millions of different websites that help facebook track your presence literally ask their hired web developers to plug in tracking scripts from facebook. It all kinda started from the Facebook like button that at some point had spread the internet. Even though this button disappeared the scripts are still used only that more often now they are hidden. Website owners plug these trackers because it helps them make few extra bucks from advertising partnerships.
Always blown away by the research and numbers here. Would love to get some insight on the research process behind these videos
His research process was covered on episode 24 of the Compounding Curiosity podcast, by the way
@@shazmosushi thanks for a great recommendation
@@shazmosushi thanks for the info.. Just downloaded it
Loved the dry humor of that pen joke.😙👌
I feel bad for kids growing up who don't understand what a pen or paper is.
I used to work for a semiconductor company. And was once given the opportunity to visit Shin-Etsu in Japan and see their production line for SOI wafers. That was a great opportunity as I have now seen the entire manufacturing process: seeing the production of wafers, visiting a wafer fab, and have visited assembly and test sites.
0kll
I definitely want to visit Soitec one day, their SOI manufacturing process seems really cool.
"Paper is like an iPad that doesn't run out of battery." Wow, that's amazing, I wonder why Apple makes iPads instead of paper when it doesn't run out of battery.
well, Apple can control what you install, do, watch, pay... Paper is too liberal :-D
Ibm made a paper think pad. Definitely superior to the ipad, and much cheaper to produce. It's a shame technology is getting worse.
@@gunner75171 oh, was that an April fool's joke ?
Then they cannot charge you for battery replacement.
The problem with paper is that it has a high per-frame cost as it is single use only and that it has a high display latency when a printer is used to update the screen or when a the display is manually scribed with a piece of graphite.
It would be nice to have a video on the industry of solar panels
I did a video on the Chinese solar industry a while back. Worth checking out
yeah let's talk solar valley rather than silicon valley!!
Thank you for this. I finally got an answer for why wafers were circular if it was obviously more efficient to print square chips on a square substrate
It would be less efficient actually. The optics is round, so you would just lose part of its field, not get. Square substrate would be not only more difficult to get and process, it would be impractical for cleaning, where you need to rotate them. The corners would be more susceptible for damage and cracks. Square substrate is bad idea on so many levels.
Problems would be created for the spin coaters which spin on etch resist on the wafer with a square wafer.
Originally, wafers were square.
...by trimming a square out of a round disc.
You did a great job describing the Cz single-crystal-silicon ingot pull ;) I worked at ARCO Solar a few years, then Siemens purchased ARCO. People don't realize or appreciate the amount of energy and effort it takes to grow an ingot. The melt of silicon on the crucible is rotated say, CCW while the seed is rotated CW. The seed is oriented a specific way ( the term 1-1-0 rings a bell ), and as I recall 4mm x 4mm x 100mm. I was there when we changed from ID saws (more kerf loss) to wire saws (the carbon slurry was with Mobil One) We could cut 4 ingots on one wire saw, yielding about 800-1200 wafers per 36 hours where the ID saw took about 48 hours for one ingot yielding about 200-250 wafers. We made 105mm wafers (as I recall), then 125mm after I left.
Gallium Arsenide (GaAs) were considered to be the future in the 1990s. They are still around as a niche product. Another niche is Silicon Carbide (SiC) for power electronics.
And now molybdenum disulfide is showing up for power electronics, though currently on smaller stuff like 45w chargers. I don't think any cars or chargers for cars or forklifts have used it yet. There's potential to make it useful for logic, but there's a lot of work to be done in order to make it so.
@@Jaker788 are you sure about that? Power semiconductor technology is my field of expertise and I have not heard about any research into MbSO2 I would be interested if you can point me towards any research you are aware of
I worked in the 90's with both GaAs and Si.
GaAs logic devices were much faster but the yield rate was diabolical. If I recall it was as low as 50%. That and material cost meant it was just far too expensive to be used for commercial grade chips.
The metallisation layer we used was gold, so no doubt that made GaAs more expensive.
I remember doing one design and the engineering samples (a small pre-production quantity) cost £1000 each, this is 1990's prices.
The common joke back then was that GaAs was the semiconductor of the future and always would be.
@@deang5622 So basically silicon is a poor man's material? :|
You might not believe this but I worked at a company in the 90's that got a job from Sanyo Solar in Carson Ca. (LA county) to make ingot transfer carts for ingot right out of the growing machines. They were lined with 1/4" (6.25 mm) kelvar blanket material. They made Solar panels here in the US! Or at least the silicon wafers. That was a different era. Paying people a bowl of rice for a days work became the trend.
Let's say someone starts a silicon ingot production company and pays their employees rediculously generous salaries
They pay employees $1 billion in total per ingot produced.
Of course you can't sell that ingot for less than a billion if you plan to recoup the costs of the labor that went into making it...or you'll go bankrupt, and everyone loses their jobs.
Let's say a separate company buys those ingots and makes wafers from that ingot, and also pays their employees a billion in total per batch of wafers produced (from 1 ingot).
So now that batch cost 2 billion (in just labor).
A fab buys those wafers, and makes micro chips from the wafers...adding another billion in labor costs to that batch.
Then you see the CPU in a store and complain that it's too expensive priced at $3 million each.
How do you suppose we make that CPU cheaper for you?
-at 2:43, it is pronounced "peg-muh-tight" that they mine the quartz from. Pegmatites are any rock with very large crystal sizes. -Geologist
It isn't an Asianometry video without a mispronunciation.
Just wait until you get a Polish person watching this.
Standing ovation Sir! And it's a miracle that you didn't break your tongue at the Polish guys name! :P
He did poczochrał Czochralski name and made Czouchralski out of him. :P
Gonna bet that he used google translate to read it because google translate is reading it as Czo-chra-lski,
voice sounds like a native Polish speaker but in 3 separate files.
btw.
"czochrać"= rub, tatter, tousle, scratch, ripple. But most of the time you can hear it in Poland via word "poczochrany"(for man) or "poczochrana"(for woman) when someone is trying to describe a mess of hair, so i am not surprised that Czochralski invented his metod by messing of the pen...
I actually found a YT video of Polish guy to figure out how to pronounce his name correctly: ruclips.net/video/Qg48Xe6U5qs/видео.htmlsi=FUDlfES9ilYAgFEh&t=85
Best video ever. Extremely informative, and the purity of the memes is off the charts!
Jan Czochralski is currently the most cited Polish scientist all because of an accidental revelation. The method is probably semiconductor's equivalent of penicillin and just as much crucial.
Gen-Z-splaining pen and paper was hilarious
Hi, enjoyed the Wafer video. I worked at a (at that time) Monsanto Silicon Wafer plant in Spartanburg SC. we built a new plant there and made 4" and 5" wafers. At that time (early 80s) that was about the largest being produced commercially. I actually have quite a few samples, lapped, polished as well as a short piece of rod and some chunks of polysilicon from a quartz crucible, actually have a couple of those also. Was a great business, but moved on in my career to other businesses.
the subtle references at the start of your videos are amazing
the freedom units at 9:30 hurt my soul
🇺🇸🇺🇸
I'm an old American koot of 62, I was working in the burgeoning digital and mixed signal industries in California in the late 70s though the 90s I enjoy this channel greatly. It saddens me to see where we have fallen to in America even with the luxuries we now enjoy from the brow sweat of our patriarchs
I find it amazing that an entire supply chain + production process that is less than a lifetime old has changed the way every human lives.
Growing up my mom used to work in a fab in SoCal and she would sometimes bring wafers home and show me what it is she did for work (I was a kid, had no clue wtf she was saying it was way too complicated). I'm guessing the wafers she brought me were faulty orwere thrown out for QC issues since they seem kinda expensive.
Her company (Microsemi) was snapped up by an Israeli firm and she switched industries shortly after. I have so much respect for her, can't believe humans went from cavemen to turning sand into these hypercomplicated semiconducters. If you tried explaining this process to someone not too long ago they would just chalk it up as being magic.
It turns out we are so good at fabricating silicon, and have such a close understanding of the properties and isotopic distribution of it, that we now use it as the physical translation of the new SI kilogram.
Silicon is awesome, and there was a project to make an exceptionally round 1 kg sphere of isotopically pure silicon. But that is not how the kilogram is defined. The modern definition of kilogram (in force since 2019) is based on "Kibble Balance" -- a special instrument which reproduces kilogram based on the value of Planck's constant and other fundamental physical units defined through their respective standards.
It's not the basis of kilogram.
Kilogram is now defined using universal constants
@@cogoid sorry, I was less than clear. I am aware that the kilogram is defined by Planck's constant, but i was referring to how we have used our exceptional knowledge of silicon as a part of checking that process.
Thanks!
“Paper is like an iPad that doesn’t run out of battery” -Asianomotry, 2022
As someone just beginning to fall in love with the world of engineering and electronics this video was fucking phenomenal thank you
then watch the whole archive of this channel.
Then find other channels, RUclips has many electronics design specialists !
Thanks, there is a specialty producer in Southern Oregon as well producing silicon wafers for specialized applications.
Cut it, etch it, grind it, clean it, crystalize it, polish it.
Technologic.
Technologic.
omg, i LOVE LOVE LOVE these silicon wafer vids!!!! keep the comming!!!
This probably won't be your most popular video but its incredibly important. I'll have to do some reading on US domestic silicon boule production. It seems odd that we would be financing a $12bn TSMC plant in AZ while also having the source silicon produced in the US without producing 300mm wafers here as well (we very well may, I just don't know).
If you're running low on ideas, electron beam lithography has been able to produce nanometer(ish) feature sizes for almost 30 years. I know the DoD has been funding it for at least that long but I don't know what secret squirrel types of products have been being produced by EBL.
-edit
Oh, also, yeah... Graphene. A lot of people (not you, other people) make the mistake of thinking a good semiconductor is automatically equally good at things like forming insulating layers or accepting metalization layers. Decades have gone into developing dielectrics and processes explicitly for silicon that are completely incompatible with graphene.
Lets go full TL;DR. Graphene might be another good topic. I'm assuming that semiconductor grade graphene has very little to do with the idealized graphene sheets people think of. I would think it would have to be at least cross linked graphene. I don't think you can just stack a bunch of graphene layers on top of each other and expect them to stay stuck together through even just operating temperature differentials.
Also: high-k dielectrics. That was on the radar like a decade ago at the 45nm node, but everything seems to have gone quiet. I don't think its because things are actually quiet though. I think high-k dielectrics are now very highly guarded industrial secrets.
EBL is used to make masks for photolithography.
This is what is great about the internet, thanks Jon. Knowledge of how our world works, well presented, easily referenced.
The sheer complexity of all this going into the first IC’s (nevermind subsequent advances) makes me terrified of how much knowledge would be lost if there was ever a significant disruption to civilization. Back to boolean gates and relays, assuming you even have enough population density and law and order to justify the effort.
Relax -- look for "I, Pencil" by Leonard Reed (I think). A page or so, and there is at least one video of it. It's all about how no single person knows how to make something as simple as a pencil, once you get into the chain saws to cut down the trees (who makes them?), the mines, the ships to transport everything ... in effect, if civilization collapses enough to lose the knowledge of how to make integrated circuits, it will probably also lose the ability to make even pencils. No halfway collapse, it's all or nothing.
There was one person who actually did try to make a pencil from scratch. I think it took him about 8 months, and that was with a lot of short cuts. It's a fascinating study in what we know.
I love the primer on pens.
Solid gags snuck into this one, the saw reference had me in stitches
That pop music intro though...
Appealing to the youths
Great subject! As a fresh grad, I worked in diffusion process before moving on to device fab, this video is a great primer the covers the fundamentals.
Very well-curated video. Me being an avid semiconductor industry enthusiast found this video to be pretty impressive.
In the category of alternative materials to silicon, unmentioned are diamond (carbon) and gallium arsenide. There was also a vogue for sapphire as a base material (coating of the wafer prior to patterning), which if this old man recalls was SOS or Silicon on Sapphire.
GenZ approved ✅
Awesome technical video! You explain the subject very well. You are never boring, and your pace and tone are incredible. I am surprised Bloomberg hasn't snapped you up yet. Thanks for all your hard work. The research and time to make and edit the video should be applauded. 👏
Im following you for quite a while now. Your way of presenting is very unique. You aren't overly enthusiastic (or actually almost not at all) but especially combined with your writing, it just works. The editing, while pretty much basic, also fits your style very well (pls keep including those bad jokes, in writing and in picture x) )
Daft Punk sends their love.
YEAH! I caught that too. I think I only caught that because I just listen the 2007 Alive album.
Tons and tons of information that I could not find in any other video tutorial. GREAT JOB! Thanks!
Silicon wafers are my most favorite snack.
“Paper is like an iPad that doesn’t run out of battery” 😂
Great video 👍🏾
Had me dying!!!🤣🤣🤣🤣🤣🤣🤣
He roasted gen Z to the Bone... Lmao
The funny thing is he's probably right. Imagine gen z looking at vacuum tubes.
@@KarrasBastomi ok boomer
@@hotosabjatahaiprabhoo..kar8081 Easy there, we are just as much ok boomer as you are. It was one hell of a joke.
the gen-z shade, lmao
I currently work at a fab and am using this video as part of a "this is what I work with day to day" explainer for my husband lol. A+ dude.
NICE #DAFTPUNK REFERENCE
A nice review of the type of place I worked at as a Electrical Maint. Tech. in the late 90's. All of our stuff was 200mm but there were a couple of experimental 300mm pullers in our building being run on and off at the time.
I have not been in the industry for decades, but I think Synthetic Diamonds have potential.
The silicon demand problem isn't solved by graphene either. Graphene nanoelectronic devices still require a silicon substrate. Graphene is useful when it's as flat as possible, silicon, more specifically silicon dioxide, is not flat, but provides a good starting point. Atomically flat flakes of other crystals, Hexagonal Boron Nitride (hBN) being one, are placed atop the silicon substrate. This is achieved by peeling off thinner and thinner layers of hBN crystals from a larger parent crystal. Graphene is then transferred onto the hBN using a similar method. On the hBN, graphene can lay falt and it's tremendous properties can be utilised. So although graphene is an interesting and exciting advancement for nanoelectric and spintronic devices, we're a long way from a silicon free world. Silicon itself is too versatile a substrate and the manufacturing processes are too well developed to transition to other materials just yet.
If you're interested in more, check out the Graphene research group at the University of Manchester. I used to work in the condensed matter research team there under Dr. Ivan Vera Marun. My research involved investigating magnetotransport and thermoelectric effects in ferromagnetic channels on hBN substrates. It's cool stuff!
what "silicon demand problem"?
Thermoelectric effect - cool stuff. Loving the pun Mr. Owens 🤪
Struggling to remember now, there used to be a materials science building, was it a joint venture between Owens and UMIST I think?
0:05 T E C H N O L O G I C
Oh nice, by the way good job fixing the audio level issue, I can finally enjoy your videos without needing to smash my volume to 100% 😍
Excellent fountain pen explanation! : ))
I have liked all the videos I've watched from your channel, but this one I think was the most interesting!
Great video as usual John
Informative and insightful, as always. Hats off!!!!
This is incredible information sir!
Thank you for the video, very informative and interesting!
was bored with older videos, i am a gen z electronics and computer engineer. this one was very entertaining and informative at the same time
Excellent explanation, thank you very much.
Paper tutorial nice, now I know
Given the amazing quality of this channel, I am utterly mystified as to why it isn't vastly more popular than it already is. This is prima facie evidence that most homo sapiens are basically stupid.
That was proper good mate, nice one! Hi from London! Subbed.
Very informative. Thank you
Thanks!! Been wanting this one for a while
I have never seen a guy on youtube so much into semiconductors and silicon wafers :D
Didn't think it would be so funny and entertaining thanks.
I wasn't paying attention for a short time until I heard you say, "paper is like an iPad that doesn't run out of battery!" You get a video like from me for this.
love the low key humor
I think the thing I learned the most from this video is what is pen and paper, none of my classmates know
"Paper is like an iPad that does not run out of battery"!
Silicon: “Crush it, grind it, legalize it.”
You left out Gas Plasma Vapor etching. It was big back in the late 60's and 70's. It cleaned, polished and scribed lines that allowed the wafers to be snapped without any substantial damage. I would say continuous improvements in the etching process is what has driven the industry forward.
I disagree. Etching is just one small part of the complete manufacturing process of integrated circuits.
The thing that has really made the difference and enabled more transistors to be integrated in to a "chip" is the photolithography. The etching is the next stage of processing after photolithography.
The achievements have been in going to smaller and smaller wavelengths of light in the photolithography and still assuring the image definition when the mask pattern is projected onto the etch resist coating covering the wafer.
Thank you for sharing this content
I read the title as Hubble Silicon Wafer and waited all the way through your interesting presentation and when you said that was it, I had to check the title again and realize my mistake. Thank you
That sarcasm on fountain pens 😂😂😂
Very Interesting. Thank you
ABSOLUTELY MINDBLOWING! What's even more amazing, considering the modern miracles of science which goes into creating these microchips, is that one microchip doesn't cost as much as a house. If you were to travel back 100 years, you wouldn't even be able to explain the concept of a microchip to another person. It would have been far beyond what even science fiction was capable of imagining. To the untrained brain, mine for example, just one microchip is as mind-boggling as the depths of the universe.
The intro cracked me up, great job as always :D
7:02 Best part of the entire video.
I think it would be interesting, in future, to have a similar video, but on GaAs or other III-V wafers production.
P.S. I enjoyed the primer's humor for gen Z of what a pen is.
I kinda wanted to know pricing on like a wafer, or one of the silicon crystals. I was looking to redo the bannisters on my staircase and silicon crystals would be a perfect fit.
A wafer costs like 400 dollars
Reject silicon boules are like $10 a pound.
Once again great video! Thanks for your effort.
How your channel don't have more subs, is beyond me! You do amazing work over and over! i feel like i even wrote you this before!
But was true then as well!
ShinEtsu has also a very large site in the Rotterdam harbour.
April 1st : Asianometry publishes a video that compares a revolutionary technology that rivals the Ipad. Totally biodegradable, made of wood fibers...
Wow ... masterful.
Thanks for the excellent video.
Try Graphene for 450mm more economical and greater surface area for other applications...
Would love to hear your take on minimal fab
I'm curious: is there a reason why the wafers aren't cut at an angle instead of perpendicularly to the growth axis? You'd get fewer but larger wafers. i donno if the tradeoff makes sense but considering most of the the losses happen at the edges it might make more sense to have a larger area on fewer slices. maybe the crystal orientation limits how you can cut it? you could also cut it lengthwise to get more rectangular wafers which would be a lot more suited for rectangular chips but each slice would get smaller in width from the largest in the center. again maybe more efficient use of the wafer surface would offset the annoyance with having to handle larger wafers of varying sizes? maybe they could all be cut into strips of more uniform width to make further processing more consistent?
Surely they have done the numbers and practical questions and answers already. Every step in the process is optimized. And one of the ways to even start getting anything useful out is to create one process that exactly can be replicated for multiple wafers. Starting to cut diagonally makes the shape not round. Edges and machine to cut it has to line up and be tailored etc etc. Ones getting a cylinder shape they just cut the thing up into as close to perfect equal wafers. With as little waste or defects as possible. Really 350mm should be more then enough. There is so much software optimization and changes needed to our lives that really it might be a good thing to stop complexity from growing and instead make sure stuff works well and efficient.
The only people in need for faster chips really are the ones running A.I and in general automation to track and make money out of peoples private life. And instead of accepting that maybe just maybe focusing on software and sustainability is more important then making wafers even harder and more complex to make. Like get rid off the people and machines (and the knowledge) in the few places making this wafers? Ooo bad times. Might it be better to make sure we are not trusting 2 places in the world to sustain basically all our wafer supply??? Instead of wanting it to be even more complicated to produce?
I would like a video going deeper into the silicon used for solar panels, developments in that field and the differences with electronic grade silicon.
Kinda like a side video to this one.
I wish these videos were shown at schools
00:06 my man went full technologic
7:01 --- HAHAHAHAHAHAHAH AHMAZING