12

This is a question about material use specifically in semiconductor manufacturing.

CNN's Ukraine halts half of world's neon output for chips, clouding outlook begins:

Ukraine's two leading suppliers of neon, which produce about half the world's supply of the key ingredient for making chips, have halted their operations as Moscow has sharpened its attack on the country, threatening to raise prices and aggravate the semiconductor shortage.

Some 45%-54% of the world's semiconductor grade neon, critical for the lasers used to make chips, comes from two Ukrainian companies, Ingas and Cryoin, according to Reuters calculations based on figures from the companies and market research firm Techcet. Global neon consumption for chip production reached about 540 metric tons last year, Techcet estimates.

and later says:

Before the invasion, Ingas produced 15,000 to 20,000 cubic meters of neon per month for customers in Taiwan, Korea, China, the United States and Germany, with about 75% going to the chip industry, Nikolay Avdzhy, the company's chief commercial officer, said in an email to Reuters.

The CNN article mentions lasers, and it certainly can't be HeNe lasers (e.g. as used in interferometric nano-positioning stages) because they are sealed and long-lived and no where near consumable on this scale.

Is neon perhaps used in large volumes in KrF or ArF excimer lasers for DUV lithography? Or perhaps is it really used in reactive ion etching instead and they've got it wrong? Helium is often used in etch gas chemistries as both a buffer and a way to limit and control the plasma temperature, but I'd never heard of neon being used.

Question: What are the ways in which semiconductor-grade neon is critical for manufacturing?

uhoh
  • 3,399
  • 2
  • 23
  • 63
  • 2
    @DKNguyen Excimer lasers for lithography are nasty, ugly, monstrous beasts and need their gas loads cleaned and purged regularly. They are not sealed glass tube lasers like HeNe or Ar+ lasers. – uhoh Mar 12 '22 at 05:13
  • 2
    @DKNguyen The Asianometry guy did a video on this recently, according to him (his info is okay much of the time) apparently it (with current practice) is consumed (contaminated and replaced) regularly. It's possible to recover and re-purify it, at least in theory, but not economical because the fab has to install expensive equipment under the floor etc. – Spehro Pefhany Mar 12 '22 at 05:13
  • @SpehroPefhany parenthetical comment: I see references to Asianometry videos in several different SE sites now; it's quite an interesting channel. and I agree that whenever something is mentioned that I know about it generally checks out. – uhoh Mar 12 '22 at 05:15
  • 2
    Interesting. I didn't realize they needed replenishing. – DKNguyen Mar 12 '22 at 05:45
  • could be HeNe - retooling to the latest most efficient/in demand chips is important too. whatever the case, we have a global economy so disturbances in one part cause disturbances in others. – Abel Apr 17 '22 at 23:23
  • @Abel I don't think so. HeNe lasers are sealed glass tubes with quite a tiny amount of neon inside (e.g. micrograms), and you buy them and they work for decades. It has to be some application where neon gas is a consumable and constantly used. "Semiconductor grade" suggests to me it's some kind of process gas. – uhoh Apr 17 '22 at 23:28
  • 1
    (+1) Good question. I used to run XeCl in my excimer laser. These lasers are finicky and need considerable upkeep and maintenance: the windows and electrodes get dirty rather quickly. I had to use 99.999% pure Xe: the cheaper 99.995% purity was too dirty to work *at all*. I have not kept up with the latest fab processes, but I suspect the neon is not being used to make the exciplex itself, e.g., NeF or whatever. If I find anything, I will post it here. I would be astonished if a HeNe laser could be used: they are intrinsically limited in power by deactivation at the walls of the laser tube. – Ed V Apr 18 '22 at 00:51
  • 1
    @EdV ya KrF* and ArF* lasers for DUV lithography abound, and He is a common additive to etch gas mixtures, not as a reactant but to redistribute energy somehow. But I've never seen Ne being used as a consumable in a fab. – uhoh Apr 18 '22 at 00:57
  • 1
    Yeah, maybe the Ne is just an essential buffer gas. Given that the 3 nm process chips are just starting to be made, I am guessing the neon is being used in the 5 nm fab. I will dig around a bit and update if I find anything. Too bad you cannot just stop by TSMC and get the info from the experts! – Ed V Apr 18 '22 at 01:09
  • I thought maybe the molten tin droplet EUV 13.5 nm process, https://www.protocol.com/enterprise/intel-euv-fab-chips, might use neon as a buffer gas, but apparently the tin droplets are in vacuo. Nice little mystery. – Ed V Apr 18 '22 at 01:25
  • 1
    @EdV See [Neon crisis worsens, Cryoin and Ingas stop production](https://semiwiki.com/forum/index.php?threads/neon-crisis-worsens-cryoin-and-ingas-stop-production.15651/) and Asiometry's [The Semiconductor Neon Shortage](https://youtu.be/SwcCC3tKZ3E?t=175) – uhoh Apr 18 '22 at 01:32
  • Let us [continue this discussion in chat](https://chat.stackexchange.com/rooms/135624/discussion-between-ed-v-and-uhoh). – Ed V Apr 18 '22 at 01:34
  • somewhat related: [What are the primary industrial uses of gallium and germanium in electronics?](https://electronics.stackexchange.com/q/672932/102305) – uhoh Jul 05 '23 at 23:53

1 Answers1

10

Neon is used the buffer gas in most excimer lasers. Both KrF and ArF excimer lasers contain roughly 98-99% neon, with the balance being the actual krypton or argon (and of course fluorine).

Due to its nature as a carrier gas, it doesn't just need to be any old neon gas, it must be extremely pure 'semiconductor grade' neon, which very few companies are able to produce. The purification process is long and complicated and there is not a lot of demand from other industries. There is a small demand from the medical industry, but it is minor compared to the semiconductor industry.

It is worth emphasizing that Ukraine didn't produce 50% of the global neon supply, but rather 50% of the global semiconductor grade neon. As stated earlier, purifying neon to this grade is not easy and and the few companies that can do it will require the better part of a year to increase their production capacity to pick up the slack.

It is also worth noting that the gas in excimer lasers is a consumable. In fact, something called 'gas lifetime' is one of the primary performance metrics for an excimer laser. Commercial ArF excimer lasers typically achieve only 1,000,000 shots before requiring that the gas be purged completely and refilled with a fresh mix of mostly neon with a little bit of argon and fluorine of course.

The gas is not actually consumed, but each shot dislodges a small amount of impurities from various sources like the electrodes, and these impurities are quite opaque to 193nm light. Eventually, the concentration of impurities grows too high and the laser will require too high of a voltage to achieve the needed dose to the wafer, and at that point, the gas is simply purged and replaced. Purifying the used gas would be the same challenge and difficulty as purifying the raw neon in the first place, so little is solved by reusing the gas. In fact, it might even be recycled, but the purifying process would be made no easier or cheaper for it.

metacollin
  • 27,884
  • 4
  • 64
  • 119
  • Informative and thorough answer, thank you! There's some additional supporting sources cited in [this comment](https://electronics.stackexchange.com/questions/611700/what-are-the-ways-in-which-semiconductor-grade-neon-is-critical-for-manufacturin/620549#comment1625771_611700). – uhoh May 21 '22 at 21:41