TR Industry, Science and Technology

[Oublious]

Can anyone please explain why Turkey got into the semiconductor game so late?

Also, why did Aselsan choose to make a DUV machine instead of an EUV one?

This video explains why we did enter chip industry late, it is Turkish.

 

[byzero]

Peki bu fabrikayı kim inşa edecek ve hangi eğitim yönetimi bu fabrikayla işbirliği yaparak çalışacak

The Ministry of Industry and Technology wants to establish a domestic chip factory with an investment of 5 billion dollars. Until 2030, the work of companies that produce solar panels such as Tübitak, ASELSAN, Kalyon, institutions such as the Presidency of Defense Industries, etc. continues. For example, Kalyon is working on semiconductor wafer production, Tübitak is working on the 65 nm chip production process, ASELSAN is working on the duv lithography device. The facility to be opened in Ankara in 2028 will be a medium-sized production facility.

 

[Saithan]

I don't feel I got an answer to the question. Just a bunch of names of stakeholders who could be part of it.

If the investment is going to contribute and we rise up as a chip producing nation we need to have solid plans for cooperation with universities and tech companies using these chips for domestic products

 

[Zafer]

Sanırım TÜBİTAK Yital'den bahsediyorsunuz ama İstanbul'da değil, Gebze'de, laboratuvar ortamında 250nm çip üretebiliyor, Katar'daki bir üniversiteyle işbirliği yaparak 65nm çipler üzerinde çalışıyor, tabii ki Katar'ın hediye ettiği 40 milyon dolarlık makinalarla seri üretim değil de prototip üretim yapacak bir tesis.

I don't know how my post ends up being in Turkish , anyway Gebze is nextdoor to Istanbul and is like an extension of the city, so. I am talking about Qatar partnership for 65 nm chips. I would say it should be for a rather small capacity investment but still beyond prototyping. If Aselsan can pull of a DUV machine that would be promising rather than buying foreign made machines as this effort can be upgraded to finer processes over time.

 

[byzero]

Yazım Türkçe olarak nasıl sonuçlanıyor Bilmiyorum :emin değilim: , her neyse Gebze İstanbul'un hemen yanı başında ve şehrin bir uzantısı gibi, yani. 65 nm çipler için Katar paylaşımından bahsediyorum. Oldukça küçük kapasitede bir yatırım için yine de prototiplemenin ötesinde olmalı. Aselsan bir DUV makinesi üretebilirse bu, yabancı yapımı makineler satın almaktan daha umut verici olur çünkü bu çabayla daha ince ayarlara yükseltilebilir.

There is a supply shortage of lithography devices in the market, China even collects second-hand devices. I think it is very difficult to set up a chip factory without a domestic lithography device. ASELSAN will prepare a duv lithography device that can produce 90 nanometer chips for the facility to be opened in 2028, then it will be the turn of the device that can produce 65 and 45 nanometer chips.

 

[Zafer]

There is a supply shortage of lithography devices in the market, China even collects second-hand devices. I think it is very difficult to set up a chip factory without a domestic lithography device. ASELSAN will prepare a duv lithography device that can produce 90 nanometer chips for the facility to be opened in 2028, then it will be the turn of the device that can produce 65 and 45 nanometer chips.

That's because ASML will not supply spare parts and maintenance to China anymore. Türkiye did not have a chip-making ecosystems other than for low production runs. Once we start building that echo system we will know if the west will cooperate or not, but we need to prepare for the worst.

 

[what]

For perspective: there are two companies at best that lead in the lithography space.

 

[mohammadtd]

That's because ASML will not supply spare parts and maintenance to China anymore. Türkiye did not have a chip-making ecosystems other than for low production runs. Once we start building that echo system we will know if the west will cooperate or not, but we need to prepare for the worst.

Excellent preparation. However , would you please shed light on front running companies in the industry. Thank you in advance.

 

[moz68k]

SMIC (China) is pushing ASML DUV systems to their absolute limits (5nm) with multipatterning. I wonder what manufacturer of DUV machines we'll be using before (if ever?) ASELSAN produces something. Probably Canon. We're so far behind that these factories will only serve our military unfortunately.

EUV relies on research from the 90s by US taxpayer funded labs (LLNL, Sandia). That's why the US has so much power over:

1. ASML (Netherlands, EUV systems)
2. Zeiss (Germany, makes the molecularly perfect mirrors in EUV systems)
3. IMEC (Belgium, fundamental IC and VLSI research partner of ASML)
4. TSMC (Taiwan, does the hard work of process development at scale, which even Intel [RIP] has consistently failed to catch up to)

Japan (Canon, Nikon) wasn't given access to this research, and ASML, IMEC and Zeiss were the only ones brave enough to realize it. Now ASML has new High-NA EUV systems that will push the boundaries of photolithography even further. Canon claims that nanoimprint lithography is viable, but this is just marketing for the foreseeable future.

Modern ICs are so advanced and in demand that wars will soon be fought over them. Brainless idealist techbro Sam Altman (who'll be the end of us) wanted 36 new factories from TSMC, and was promptly brushed off because the cost of even a single one is unfathomable. While AI is overhyped to some degree, our demand for compute will grow exponentially. We're too late, but I suppose something (literal early 2000s tech) is better than nothing. I am eternally saddened by how poorly we invested in our nation's development.

 

[Zafer]

SMIC (China) is pushing ASML DUV systems to their absolute limits (5nm) with multipatterning

They are reporting 4nm now.

 

[Chocopie]

Now ASML has new High-NA EUV systems that will push the boundaries of photolithography even further.

ASML's TWINSCAN EXE:5000 High-NA EUV lithography machine (called "tool") is mindblowing technology. It‘s massive and expensive as hell.

165t weight
5-6 months installation time by 250 engineers followed by months of calibration
$380-400 million per system

Intel is the 1st to introduce High-NA EUV (they bought the most of ~10 machines produced so far), 2nd TSMC, 3rd Samsung, 4th SK Hynix.

Installation clip at Intel‘s Hillsboro fab:

ASML’s marvelous promo clip: