Jay Taylor's notes
back to listing indexUltrapure Water for Semiconductor Manufacturing - YouTube
[web search]
Original source (www.youtube.com)
Clipped on: 2024-09-26
Ultrapure Water for Semiconductor Manufacturing
109,672 views views
Description
109K views 2 years ago
It is the purest water you will ever know. And every day, chip factories are sloshing their wafers with it.
Ultrapure water or UPW is an industry term. A term that describes its product quite well. Water with purity requirements so strict, you're more likely to win the national lottery than to find a non-wat …
...more
Top Comments
@notanymore9471 | I work at a fab in facilities and just started taking care of the water systems and this video really helped understand wtf is going on. Thank you for the great research!
2 years ago
@Kevin-jb2pv | My dad spend about 10 years designing these ultrapure water purification systems for Evoqua/ Siemens water technologies/ a half dozen other names while he was there, IIRC. So when you say that the "killer particle size limit" got small enough that lasers could no longer detect them, I actually remember him talking about what a headache this was. I've actually seen some of these systems you showed pictures of, or newer models. They used to make water purification systems for both semiconductors and pharmaceuticals, and oh boy, pharmaceuticals sounded like a lot less of a bitch to deal with than semiconductors. Pharmaceutical water just needed to be sterile and not have any contaminates that could chemically react with anything and cause unwanted (possibly dangerous) byproducts during production, or otherwise cause yields or quality to be inconsistent (you really, really don't want your medicine to be even a fraction of a percent more or less potent than you're expecting when you're talking about making bulk chemicals that then have to be divided into millions or billions of exact milligram or microgram doses). When it came to semiconductor water, he used to say that the water has to be so pure, that if you were to blow up a single die on a wafer to the size of a football field, then it would only take an impurity the size of a single marble to scrap that unit, and he had to make sure that it wasn't his water that was responsible for letting any of these "marbles" get through.
At my job I measure sodium impurities in my process in PPT, parts per thousand. He also used to measure in PPT, but his was parts per trillion, and he was dealing with a lot more difficult impurities than just simple table salt.
To get an idea of how pure this water is, he used to say that if you filled a kiddie pool with this water and stood in it, you could shove a high-voltage transmission line inches away from your ankles and you wouldn't even feel a tingle, because there was simple nothing dissolved in the water that could conduct any current.
I am intimately familiar with the Intel 14nm problems they were having back then, too. One time around 15 years ago, now, in the summer my dad needed to drive down to Intel's Phoenix fab (from Colorado Springs) to deal with a problem they were having with his water system they had just installed. He asked me to go with him for the company and to split the driving with him (it's about a 12 hour drive each way). He thought it was going to be a "drive there, look at it for about 15 minutes, press a couple buttons, drive back" thing, but... Nah. I wound up stuck sitting in the parking lot in his Volvo for 4 or 5 hours in the Phoenix heat with nothing at all to do. I couldn't go into the building because I didn't have security authorization, and I couldn't really go in and page him because the campus is huge and there's clean rooms and stuff, so me paging him to the front could easily waste an hour or everyone's time after he had to unsuit for the clean room, walk way the hell across the huge campus to the front, probably just tell me that I couldn't go in, anyways, and then walk all the way back, suit back up, and get back to what he was doing. So I was just stuck with a non-smart phone, a Nintendo DS that was starting to give me a headache from how long I had been staring at it, the AM radio. I had the keys, and I turned on the AC, but lemme tell you, an AC system built in the 90's to deal with the weather conditions IN SWEDEN is absolutely not up to the task of fighting against the midday heat of the Arizona Desert in the middle of summer while idling in the middle of a hot parking lot. So I wound up basically pacing around a lot, turning the car off and on to keep the AC from overheating the idling engine, and eventually detailing the interior of his car with a paperclip and some napkins out of sheer boredom. That was a pretty miserable day. Worst part was that Whataburger was closed by the time we left so I still haven't gotten to try that.
He asked me to go a couple more times and I found other shit to do. But I certainly heard all about the problems that 14nm was causing.
And, yeah. He has (or _had, maybe, he left that job about 3 years ago) that book you showed at the end. I looked at it for about 15 seconds and decided that it was all just bullshit cryptic jargon written that way to conceal the fact that they were actually performing an old school purification ritual by sacrificing goats and virgins under a blood moon to make this shit work.
2 years ago
@csours | sometimes I think about a portion of society bootstrapping itself back to our level of technology. Getting to early MOSFET ICs could probably be done with the economic power of a population of one hundred housand people, but it takes an advanced world economic system to get to the nanometer level we're at now, even if you had all the reference manuals in the world
2 years ago
@innovatingtechnologist5385 | I work as a wet process engineer in development for EUV. Water purity is critical and something I work on regularly. This summary is brilliant and very accurate. One thing you missed is filter analysis. There are nm scale filters on tools in the fab that will catch most final particles coming from facilities. After running for sometime you can cut them open and analyze them, which solves the dilution problem of particle analysis. There are also filters that are tuned to chemically absorbed specific contaminants, based on your local water quality issues (metals, silica, etc). Great video!
2 years ago
@michaelvanderzee269 | I'm a Canadian engineer who did work in Kaohsiung, Taiwan many years ago installing a large networked SCADA system at several treatment plants for the Taiwan Water Company. Those facilities I worked on are probably are the precursors to supply the raw water used by TSMC for their fabs on the island. Cool.
2 years ago
@microdesigns2000 | Everybody in our engineering department, some in the maintenance, and a few managers are spending most of our time trying to improve yield in our product line. We sputter large 4x3 meter “objects”. A single defect can scrap a unit costing us several thousand dollars and all the wasted time. Since the line isn’t exactly screaming fast, this costs us millions. I am thankful that we have made great effort on this problem from many angles, including ones that I wouldn’t have thought of because we have a really talented team. I most recently finished a procedure for tracking down the cause of a single little scratch. Since the production line is over half a mile long, it was no small task. It took months to make a decent system so a future engineer will be able to make quick work of it. Tomorrow I will introduce new PLC software that will reduce processing times at one our bottleneck areas. I’m excited about that because my single software edit will save about $300-500k per year.
We utilize chilled DI water for many processes, pure but not like the UPW in this video. Many places we use DIW for cooling high voltage electrical circuitry because pure water doesn’t conduct electricity, imagine that.
In this video there is discussion about water purity to avoid sparking junctions on the wafer. I question how they prevent static charge build-up in the water itself.
2 years ago
@Muonium1 | Where I work we use 18.7MΩ UPW for ultrasonically washing massive coffee table size glass substrates before they're loaded into metal oxide electron beam vapor deposition chambers to be multi-layer dielectric coated with Bragg diffraction interference polarizers, anti-reflection coatings, and mirrors used in the construction of ultra-high power neodymium glass lasers for inertial confinement fusion research. The particle counter on the final deionized water loop from Particle Measuring Systems that detects down to 30nm diameter particles is the size of a desktop PC and costs about as much as my house. The software that runs the counter from the same company is actually amazingly crap and looks like something from the windows 3.1 era.
Also PS, I really wish the ridiculous and very old myth of ultrapure water being "so toxic and corrosive because it sucks minerals out of everything it touches" would finally die. It's not. I've drank it. It's just water, it tastes great, I didn't die. The instant it touches your tongue, or you, or ANYthing except the hyperpure fluorinated polymer pipes that it runs through, it's not ultra-pure anything anymore, it's just normal water again with normal water's normal properties. One of those properties is actually really interesting - people always say water is clear and it just looks blue because it's reflecting the sky or has junk dissolved into it or whatever, but that's wrong, it's ACTUALLY blue. When we fill a super shiny stainless steel ultrasonic washer with it the color gradually gets bluer and bluer as it gets deeper and there's obviously no particle impurities in there doing it. All the color is coming from the third overtone of the stretching mode of the hydroxy group part of the water molecule absorbing a tiny amount of far red light in the tail of the absorption peak.
2 years ago (edited)
@SGresponse | Before you explained the source of Urea in Intel's water my thoughts went to "a disgruntled janitor pissed into the ultra-pure water tank." That would have been A STORY.
2 years ago
@akashtripathi5347 | I work in facilities in a semiconductor fab. This video clearly explains the everyday process and efforts we make to achieve strict UPW specifications and supply continuous supply on a 24*7 basis. As shown in the video, the real problem we face is in the analysis of the purity of water due to very presence of impurities in ppb or ppt levels. The analyzers or instruments use high-tech metrology techniques to detect impurities; sometimes, they show errors and need maintenance. Repair and maintenance of these hi-tech equipments is a big task in itself. Due to disruption in the global supply chain in the semiconductor sector, long lead time in procurement, and minimal industry players, we face problems in sending analyzers for calibration or maintenance or even procurement of spares is getting complicated. At last, I would say, producing a UPW water isn't easy and need a lot of effort.
2 years ago (edited)
@paintballthieupwns | Your dry humor is more deadly than a killer particle :) Love your delivery style. I never understood why so much water was needed - thanks for the understanding
2 years ago