Micron is a US memory manufacturer.
Mikron is a Russian semi fab which makes several products from credit card chips to 90nm CPUs using SPARC and Elbrus2K architecture.
lancelot wrote:Micron =/= Mikron.
Micron is a US memory manufacturer.
Mikron is a Russian semi fab.
But as pointed out earlier, Russia has already instructed to build two types of Lithography equipment for production of chips 130nm and 28nm. So your suggestion isn't being taken into consideration by the Ministry of Industry and Trade - they know far better than you or I about what is needed and in demand in Russia.
Especially when Russia get pushed into a box where they cannot sell to the whole world market while those countries can basically continue trading.
sepheronx, magnumcromagnon and kvs like this post
sepheronx, GarryB and magnumcromagnon like this post
lancelot wrote:A 28nm fab would not be that expensive. It would cost like $2-4 billion USD. That is well within the reach of Russia.
It is when you get to 5nm or lower with EUV that the price of the fab goes up by like an order of magnitude.
At that point you cannot justify that investment unless you have a huge market to sell to.
The problem is with the current sanctions Russia cannot buy the equipment to make a 28nm fab. Neither new nor used.
Not that there are used tools around given all the factories which can manufacture at 28nm are probably near full capacity right now.
Russia bought the tools for Angstrem when AMD (now Globalfoundries) in Dresden switched to 28nm production. AMD had to get rid of their older 130/90nm tools. So they sold them for cheap including a license for the process technology to Russia. I think that was a good deal. Even if it wasn't followed through quickly enough and they took forever to build the facilities and put the equipment in them. Supposedly Angstrem got the equipment and licenses for the process from AMD for like $500 million USD back when AMD were in a tough financial condition around a decade ago.
The Chinese are still working on making a 28nm production line with 100% Chinese tools. So even they cannot do it right now.
Even when the Chinese can do it they will have issues scaling production and getting the tools to work reliably. Realistically at best this will happen in like two years. There are rumors the SMIC fab at Beijing will have 100% Chinese tools and it is going to use 300mm wafers. Still unknown which design rules it will use though. Might not even be 28nm. For reference the fab is supposed to cost $7.6 billion USD and produce 35,000 wafers of 300mm diameter a month. That is like twice the cost of other similar fabs and probably reflects the fact the factory was built more recently (i.e. inflation) and the extra premium of having to use Chinese tools. You cannot expect the early production tools to be cheaper than more established ones with a long production run.
Existing Chinese semi fabs at 28nm or better will be using Western tools and will be vulnerable to sanctions.
GunshipDemocracy likes this post
magnumcromagnon, Rodion_Romanovic and LMFS like this post
sepheronx, magnumcromagnon, kvs and GunshipDemocracy like this post
sepheronx, magnumcromagnon, GunshipDemocracy, Rodion_Romanovic, Sprut-B, gc3762 and rigoletto like this post
kvs wrote:Russia does not have to start from scratch. It already has companies that produce tools for wafer processing. The all-important
practical experience is not absent in Russia. The time is ripe for Russia to get in gear and produce EUV IC manufacturing equipment
itself. It can use China to contract production volume which it cannot have itself in the initial stages. In a few years that need
will abate.
I really do not see a technological barrier for high resolution lithography in Russia. It is the same story as with all the other sectors
that became dominated by the west in the 1990s. After western sanctions spasm of 2014 this distortion is being successfully removed.
GunshipDemocracy wrote:kvs wrote:Russia does not have to start from scratch. It already has companies that produce tools for wafer processing. The all-important
practical experience is not absent in Russia. The time is ripe for Russia to get in gear and produce EUV IC manufacturing equipment
itself. It can use China to contract production volume which it cannot have itself in the initial stages. In a few years that need
will abate.
I really do not see a technological barrier for high resolution lithography in Russia. It is the same story as with all the other sectors
that became dominated by the west in the 1990s. After western sanctions spasm of 2014 this distortion is being successfully removed.
I guess truths is somewhat between urra we got all and oh no we are in the middle of big ass. AFAIK There are parts of the whole microelectronics lifecycle, the question is how fast you can add missing elements i mean time vs. money. How cooperation with China and eventually late India can help here?
GunshipDemocracy likes this post
sepheronx, kvs and GunshipDemocracy like this post
magnumcromagnon, kvs and GunshipDemocracy like this post
sepheronx, magnumcromagnon, GunshipDemocracy and LMFS like this post
In Russia, the development of domestic lithographic scanners has begun — their production will begin in 2025-2026
10.02.2022 [16:28], Gennady Detinich
At the end of last yearThe Ministry of Industry and Trade has allocated 5.7 billion rubles for the development of domestic lithographic scanners. Recently it became known that the Zelenograd Nanotechnology Center (ZNTC) has started project work on this order. Learn more about the center's project resource plans zelenograd.ru Anatoly Kovalev, General Director of the ZSTC, said.
The Center has signed two contracts with the Ministry of Industry and Trade: one for the development of scanners with a topology level of up to 350 nm, the second — up to 130 nm (the interview refers to steppers, although this is the old name for scanners, when photomasks were mechanically shifted to project onto a new section of the silicon substrate). As a result, as planned, fully domestic photolithographs will replace imported ones, in particular, those currently used in Zelenograd factories.
Serial production of 350 nm photolithographs is expected to start in 2025. A ready-made semiconductor laser of domestic production will be used for installation. It is possible that in the future lasers of Zelenograd companies may also be used for this purpose. According to the contract, the scanner design and its prototype should be ready by the end of 2024, including all documentation for the launch of mass production of the installation.
From the tender documentation, it follows that the 350-nm scanner will be a 3.5-ton installation with dimensions of 2 × 2.6 × 2.5 m and a control complex with dimensions of 2 × 0.8 × 1.6 m. The unit is focused on processing 150-and 200-mm silicon substrates.
Development of the 130-nm scanner will be completed about a year later. For this scanner, a domestic 193-nm laser will be created from scratch in order to get away from imported components, in particular, from the lasers of the American company Cyber, which are now widely used in Russian semiconductor factories. For this reason, the completion of the 130-nm scanner project will take a little longer.
The project of a 130-nm scanner with the ability to process up to hundreds of 200-mm plates per hour will be put into mass production at the end of 2026. In the future, the plant will be upgraded to produce chips with lower technological standards up to 65 nm. Reducing the size of the topology, among other things, will be achieved by double exposure. The main plate size for this scanner will be 200 mm substrate with optional use of 150 mm substrates.
According to the developers, the vast majority of orders for the production of chips in the world lies in the range from 250 to 65 nm, so there is no point in chasing tens and units of nanometers. The main thing for Russia in this project is to avoid possible sanctions on the supply of industrial semiconductor equipment. Own production of scanners can protect the country's electronics industry.
The serial scanner manufacturer has not yet been selected. This may be ZNTC, Zelenograd companies, or Belarusian partners represented by Planar. In the USSR, lithographic scanners, which, by the way, still continue to work in Russian factories, were produced by the Minsk plant "Planar". This practice may extend to the release of a new generation of scanners.
sepheronx, GarryB, magnumcromagnon, Rodion_Romanovic, thegopnik, LMFS, gc3762 and like this post
LMFS likes this post
sepheronx, Werewolf, magnumcromagnon and kvs like this post
sepheronx, magnumcromagnon and LMFS like this post
LMFS likes this post
lancelot wrote:Like I said this will be hard to do because most Chinese foundries and other semi fabs use Western chip making tools and in some cases supplies and they can be sanctioned themselves. At this point it is still too early for the Chinese semi industry to assert itself.
sepheronx, magnumcromagnon and LMFS like this post
LMFS likes this post
sepheronx, dino00, magnumcromagnon, kvs, LMFS and lancelot like this post
Technology that will destroy the Dutch company ASML: development of a mask-free photolithograph for 28 nm standards has begun in Russia
2 days ago
4,6 thousand reads
65% of all modern processors are produced on photolithographic scanners of the Dutch company ASML. It costs billions of dollars to develop such
technologies from scratch. Therefore, well-known manufacturers such as Intel and AMD preferred to invest in Dutch technologies. However, Russian
scientists intend to "break" the monopoly of ASMLithography. We'll tell you exactly how.
What Russian scientists came up with
The cost of the lithograph is above $300 million. The smaller the technical process (and now 7 nm, 6 nm are already used), the more expensive it is.
To work, the photolithograph needs masks, so-called templates, which are used to transfer the drawing. This template is made for each new generation,
that is, for core i7 3770 — one mask, and for core i7 4770-another.
The Moscow Institute of Electronic Technology (MIET) is developing the concept of an X-ray lithograph, for which masks (as a separate part) are
not needed. The entire drawing, which is a digital image, is projected onto a micro-optical system( MOEMS), essentially a micro-screen. At the
same time, the equipment supports a process speed of 28 nm or less, up to 10.
Theory or reality
MIET, based on Russian developments, received a state contract worth 670 million rubles for the creation of technology and models of masks,
technical specifications and technical and economic justification of the first sample of the lithograph. All work is expected to be completed in
November this year. That is, the creation of a prototype may begin in early 2023. Industrial production is projected to start in about 5 years.
The project will use theoretical developments and equipment of many research institutes and design bureaus. In particular, the Zelenograd
company "ESTO", TNK "Zelenograd", SIC "Kurchatov Institute", branch of the Institute of Applied Physics of the Russian Academy of Sciences
in Nizhny Novgorod, Institute of Spectroscopy of the Russian Academy of Sciences in Troitsk.
It is noted that such lithographic scanners will be several times cheaper than those that use radiation with a wavelength of 193 nm (deep ultraviolet).
And they, in turn, are several times cheaper than the most modern ones that use hard ultraviolet light.
And this is not taking into account that the speed of operation of maskless multipath scanners will be several times higher. We wrote more about
the technology and device here.
If Russian scientists successfully complete the development, the Dutch company will face an inglorious end.
Economic prospects
Research on the creation of a maskless photolithograph has been conducted since 2010. So, success is quite likely. There are no analogues of the
Russian technology. It can really change the idea of manufacturing microprocessors, as it reduces the cost of the entire process.
In the current conditions, large-scale production of chips is economically profitable. A maskless lithograph will allow you to create popular narrow-profile
and specialized processors. Very optimistic forecasts. How real do you think they are?
sepheronx, GarryB, magnumcromagnon and LMFS like this post