Neon, Ukraine, And The Global Semiconductor Industry

On our news feeds and TV channels at the moment are many stories concerning the war in Ukraine, and among them is one which may have an effect on the high-tech industries. It seems that a significant percentage of the world’s neon gas is produced in Ukrainian factories, and there is concern among pundits and electronics manufacturers that a disruption of this supply could be a further problem for an industry already reeling from the COVID-related chip shortage. It’s thus worth taking a quick look at the neon business from an engineering perspective to perhaps make sense of some of those concerns.

As most readers will know from their high school chemistry lessons, neon is one of the so-called inert gasses, sitting in the column at the extreme right of the Periodic table. It occurs in nature as a small percentage of the air we breathe and is extracted from the air by fractional distillation of the liquid phase. The important point from the above sentences is that the same neon is all around us in the air as there is in Ukraine, in other words, there is no strategic neon mine in the Ukrainian countryside about to be overrun by the Russian invaders.

So why do we source so much neon from Ukraine, if we’re constantly breathing the stuff in and out everywhere else in the world? Since the air separation industry is alive and well worldwide for the production of liquid nitrogen and oxygen as well as the slightly more numerous inert gasses, we’re guessing that the answer lies in economics. It’s a bit harder to extract neon from air than it is argon because there is less of it in the air. Since it can be brought for a reasonable cost from the Ukrainians who have made it their business to extract it, there is little benefit in American or Western European companies trying to compete. Our take is that if the supply of Ukrainian neon is interrupted there may be a short period of neon scarcity. After that, air extraction companies will quite speedily install whatever extra plant they need in order to service the demand. If that’s your area of expertise, we’d love to hear from you in the comments.

Here at Hackaday we are saddened beyond words at what has happened in Ukraine, and we hope our Ukrainian readers and those Ukrainian hackers whose work we’ve featured make it through safely. We sincerely hope that this madness can be ended and that we can mention the country in the context of cool hacks again rather than war.

If you are interested in the strategic value of inert gasses, have a read about the global helium supply.

Header image: Lestat (Jan Mehlich), CC BY-SA 3.0.

34 thoughts on “Neon, Ukraine, And The Global Semiconductor Industry

  1. Not for nothin- I hope the readers, hackers, and everyone else including the folks who I dont agree with on anything make it through this. No one deserves this. Slava Ukraini!

    1. There are air separation units across the planet as noted. But as also stated there are not many set up for the collection of neon. It takes the shut down of a plant and about 25 days of work to modify the cold box section where the distillation columns are located. Then the infrastructure outside of the cold box needs to be built to refine the neon to microprocessor grade and ultimately to storage tanks ready for shipment. It’s a significant outlay of capital and would only be pursued at large facilities located in a geographical location near the demand.

      Given typical business processes, material lead times, engineering, and construction it would be at least an 8 month process from concept to start up – with a focused team. With constraints on getting anything industrial fabricated these days being a major constraint that would be hard to overcome.

      The article did not mention the mass of neon that needs to be replaced but it’s unlikely that the US plants can replace it until say 2024.

  2. Not Neon specifically, but I’m right now learning the technology of this company:

    They’ve got a process that extracts CO2 from the air using an electrolytic cell, in a much cheaper fashion than distillation. Their original paper is here:

    The process is simple and apparently works with CO2 at any concentration, meaning that they can extract it from (industrial) combustion sources or take it directly from the air.

    There are a lot of stranded energy sources such as photovoltaic or wind generation when there is not enough demand. The verdox process is electrical, quick to run up (or down), and could serve as a useful purpose for the unused electricity.

    CO2 is a saleable commodity right now, but if we start doing carbon capture we will have too much to use, so the proposal is to pump it underground where heat and pressure will turn it into solid carbonates.

    It’s the first and probably most difficult breakthrough needed for carbon capture, turning back the clock on greenhouse gasses. (Yes, we will have to capture a lot.)

    It’s unlikely that this type of process will work for Neon, but only note that gas extraction technology gets better over time.

  3. Header pic shows argon-florescent light not neon.

    I was in a bar tonight and they have “neon” lights in the windows plugging the basic brew. Looking closely I noticed they have a similar diameter tube and diffusing with I assume LEDs inside. Looks like the real thing even up close. No big transformer box on the back was another giveaway. Some places have window beer signs that are HV cold cathode still.

    Between safety and the war we may be seeing less neon.

    1. Found my answer “Deepultraviolet(DUV)lithography,thecurrentworkhorse of the patterning tools, uses an electrical discharge in neon or krypton mixed with halogen gases like fluorine to produce UV light at 193 nm and 248 nm;”
      (ref: plug into your search engine of choice ‘ “EUV Lithography adds to increasing hydrogen demand at leading-edge fabs” filetype:pdf’ ).

      It is used for the bread and butter silicon chips, not the bleeding edge one.

      1. Probably also bleeding edge ones as all layers doesn’t need EUV. EUV masks are extremely expensive and with a limited lifespan so everything like metal routing layers are probably DUV, unless there’s some weird quirk that makes it impossible.

        1. Normally in chip manufacturing a given fab only has a few lithographs “at node”, and then they get progressively larger from there. Mainly since larger nodes have better yield and the lithographs are cheaper. But most chips do not need a lot of layers at node either.

          So a hypothetical 22nm fab might offer you 4-6 layers at 22nm, then 2-4 at 28-32 nm, 2-4 at 32-48 nm, and a couple at 100+ nm for making the large interconnects and such. Most fabs even have a few lithographs at 1+ µm, typically used for making pads and such and these are sometimes maskless, effectively an overgrown extremely high res projector.

          Though, how many layers one can make at a given node will vary depending on the fab, and what nodes a fab has on offer also varies. And since not all chips will use the same number of each, then a fab can start to schedule production to try to make as much use of their machines as possible.

          This is also a reason for why a chip tends to be specifically catered to a specific fab, since it isn’t just about the smallest layer. If fab A offers you 4 layers at a given node and fab B only offers 2, and your chip needs 3, then only fab A can make it.

          1. You are totally wrong and have shown you know absolutely zero about the process!! Anyone reading the above comment please ignore every last thing as it is virtually all incorrect!

          2. Billy

            I were simplifying for ease of readability.
            But the content provided is very generalized, the exact numbers varies greatly depending on what fab one goes to.

            Some fabs can offer one 10+ layers at node, while others don’t.
            Then there is of course all the other manufacturing processes as well, I don’t know of any fab that offers all types of etching or depositioning methods available in the market.

            And for the more cutting edge fabs, then a chip design is more or less specifically designed for a specific fab in mind. But this isn’t typically the case for larger nodes that are a couple of generations older.

            Though, I am curious to what details you consider to be wrong?

  4. Kinda makes me want to start extracting Neon at home, then carefully collect and store it over the next 20 yrs, ya know as a hobby. Maybe by then I’d control most of the world supply of Ne. Similar to Helium and CERN…

  5. Back in the 80’s I apprenticed at a Neon sign shop and learned how to make neon tubes. At that time, neon was a leftover byproduct of oxygen distillation and the sign shops could get large flasks of it for free, including free shipping. Apparently the distillers were not allowed to just release it. Ditto for Argon and Xenon.

    1. That seems profoundly silly since those gases were already in the air to start with.

      But desalination plants have special places to return the salt back from whence it came. Maybe it’s a concentration thing

      1. Would be pretty funny if earlier oxygen distillation plants had an eerie glow due to just releasing their noble gases on site.

        Then again, I agree, we should keep the nobility in check, no matter how inert they may act

        1. Unfortunately this JS seems to work terribly in Firefox.
          Know from a friend that it basically occupies 2-2.5 CPU cores while barely producing a few kB/s of throughput…

          And the LOIC doesn’t support HTTPs

  6. Look at the data for Australia and you will notice that parts of it are the safest and most secure places in the world, yet they have almost no manufacturing. What on earth is going on there?

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.