Chinese semiconductor industry


tokenanalyst

Senior Member
Registered Member

China reports robust semiconductor industry growth despite US sanctions​


Although some of its key semiconductor industry players were put on the Entity List by the US government since 2019, China managed to register double-digit growth in the revenues of all of its semiconductor sub-sectors in 2020, pushing up global market shares to challenge Taiwan and the US, according to several recent reports.

If the robust growth continues, China's semiconductor industry will grow rapidly in the next decade and by 2030 will become a world champion in semiconductor manufacturing, seizing a 24% global market share, as previously predicted by the Semiconductor Industry Association (SIA).

According to an SIA blog, the semiconductor industry in China grew 30.6% in 2020 to reach $39.8 billion in total sales, representing robust annual growth of 36%, 23%, 32%, and 23% for the fabless, IDM, foundry, and OSAT sectors, respectively.

China registered 9% of the global semiconductor market share in 2020, surpassing that of Taiwan for the second year in a row. It is likely to overtake Japan and Europe soon, as the global market shares of the semiconductor industry in the two major economies both stood at 10%.

China's fabless semiconductor sub-segment ranked third after the US and Taiwan, taking 16% of the global market share, up from 10% in 2015, even though it was subject to tightened chip export control restrictions.

In terms of wafer capacity increase, China accounted for 26% of the worldwide total, according to IC Insights and SIA, while its top 3 outsourced semiconductor assembly and testing (OSAT) companies collectively garnered more than 35% of the global market share.

China has already announced US$26 billion in new planned funding for 28 additional fab construction projects in 2021, mostly focusing on mature processing nodes.

Private sector investors, supported by government subsidies and favorable tax incentives, have poured money into China's semiconductor industry. Some consumer electronics and home appliance OEMs, such as Xiaomi and Alibaba, are designing chips in-house and having them manufactured by local foundry makers such as SMIC or Hua Hong, an effort aligned to Beijing's call to seek IC sovereignty and self-sufficiency.

SIA said nearly 15,000 semiconductor startups were established in 2020 alone, in which many are IC design companies specializing in GPU, EDA, FPGA, AI computing and other high-end chip design. China now has more than 350,000 businesses operating in its semiconductor sector, 80% of which were registered within the last 5 years, and 30% registered within the past 12 months, according to Global Times, quoting data from Chinese online business information provider Tianyancha.

In addition to copious funding provided by China's central government, provincial governments, and municipal governments, startups in China also got financial backing from local companies such as Huawei, which has also aggressively invested in 56 Chinese semiconductor supply chain companies via a "Hubble Technology Investment Fund" to foster its own semicon ecosystem, according to the Wall Street Journal.

Please, Log in or Register to view URLs content!
 

Weaasel

Junior Member
Registered Member
Asking for a friend. In a synchrotron, the charged particles being accelerated to emit EM radiation are normally and initially widely dispersed and scattered, right? Therefore those that are accelerated to the extent of emitting EUV and x-ray photons will not do so coherently to produce a single or multiple concentrated beams... I suppose that steady state microbunching is a methodology of accelerating a very close concentration and collection of charged particles to emit highly concentrated beams of radiation that one aims for... How exactly is that done? Anyway, how exactly, basically speaking does SSMB work?
 

FairAndUnbiased

Major
Registered Member
Asking for a friend. In a synchrotron, the charged particles being accelerated to emit EM radiation are normally and initially widely dispersed and scattered, right? Therefore those that are accelerated to the extent of emitting EUV and x-ray photons will not do so coherently to produce a single or multiple concentrated beams... I suppose that steady state microbunching is a methodology of accelerating a very close concentration and collection of charged particles to emit highly concentrated beams of radiation that one aims for... How exactly is that done? Anyway, how exactly, basically speaking does SSMB work?
No, it is the opposite. Synchrotron radiation is not an isotropic source, it is actually one of the most highly focused, highly collimated sources of light possible.

Please, Log in or Register to view URLs content!

Please, Log in or Register to view URLs content!

LPP is an essentially isotropic source.

Please, Log in or Register to view URLs content!

Large, natural synchrotron sources like black hole jets can be seen from literally across the universe where isotropic sources (like stars and galaxies) are too faint to even be resolved individually from the same galaxy cluster.
 

Phead128

Captain
Staff member
Moderator - World Affairs
  • Is it hard to produce light on the EUV spectrum using a synchrotron?
  • Is the light on the EUV spectrum produced using a synchrotron powerful enough to cut patterns into a chip?
  • Is the difficulty building the synchrotron, or is the difficulty to modify synchrotron to produce EUV light?
    • I see Shanghai ZhangJiang has a synchrotron facility already, would it be hard to modify an existing facility to produce light on EUV spectrum for lithography purposes?
 

tokenanalyst

Senior Member
Registered Member
Asking for a friend. In a synchrotron, the charged particles being accelerated to emit EM radiation are normally and initially widely dispersed and scattered, right? Therefore those that are accelerated to the extent of emitting EUV and x-ray photons will not do so coherently to produce a single or multiple concentrated beams... I suppose that steady state microbunching is a methodology of accelerating a very close concentration and collection of charged particles to emit highly concentrated beams of radiation that one aims for... How exactly is that done? Anyway, how exactly, basically speaking does SSMB work?
1642197002616.png

With microbunching the ultra bright accelerated light produced become more pulse laser like which is perfect for EUV patterning.

This article from physorg explain it perfectly.

The most modern light sources for research are based on
Please, Log in or Register to view URLs content!
. These are large facilities in which electrons are accelerated to almost the speed of light, and then emit light pulses of a special character. In storage-ring-based synchrotron radiation sources, the
Please, Log in or Register to view URLs content!
travel in the ring for billions of revolutions, then generate a rapid succession of very bright light pulses in the deflecting magnets. In contrast, the electron bunches in free-electron lasers (FELs) are accelerated linearly and then emit a single super-bright flash of laser-like light. Storage ring sources as well as FEL sources have facilitated advances in many fields in recent years, from deep insights into biological and medical questions to materials research, technology development, and quantum physics.

Now, a Sino-German team has shown that a pattern of pulses can be generated in a synchrotron radiation source that combines the advantages of both systems. The synchrotron source delivers short, intense microbunches of electrons that produce radiation pulses having a laser-like character (as with FELs), but which can also follow each other closely in sequence (as with synchrotron light sources).
Please, Log in or Register to view URLs content!
 

FairAndUnbiased

Major
Registered Member
  • Is it hard to produce light on the EUV spectrum using a synchrotron?
  • Is the light on the EUV spectrum produced using a synchrotron powerful enough to cut patterns into a chip?
  • Is the difficulty building the synchrotron, or is the difficulty to modify synchrotron to produce EUV light?
    • I see Shanghai ZhangJiang has a synchrotron facility already, would it be hard to modify an existing facility to produce light on EUV spectrum for lithography purposes?
1. It is trivial to tune synchrotron radiation to arbitrary wavelengths. It is why it is one of the only high intensity sources of soft x-rays, terahertz and other wavelengths with almost no other sources.

synchrotron2.jpg


2. Yes, synchrotron radiation is literally one of the most brilliant sources of light in the universe. In addition is it's collimated light, like a laser, while LPP produced EUV is ironically mostly isotropic like a 19th century light bulb. There's not much directionality because all it is, is basically superheating the Sn plasma to extremely high ionization states. It's essentially a thermal emission process.

3. The difficulty is the synchrotron is large, fixed and takes up an entire building with numerous recurring costs like liquid He coolant for the steering magnets, vacuum pumps, etc. Another difficulty is transmitting the light from inside the synchrotron to the outside when EUV is absorbed and/or scattered by everything. However these are engineering problems that can be solved by i.e. differential pumping, new materials, etc.

Otherwise the problem is that simply has no motivation to look into this because it bought Cyber, was first, and got it's competitors banned. But nobody says first one is the correct one. Why would the first one be necessarily the best one?
 

gelgoog

Brigadier
Registered Member
A lot of people worked on x-ray lithography at least in the 1980s if not earlier. They had trouble getting a proper light source and synchrotrons were one of the trial light sources. Well that was litany of failure. EUV litho is basically soft x-rays all over again.
 

gelgoog

Brigadier
Registered Member
I am not sure this real or not?

The info in that robot's infodump is not new. The video is a jumbled up mess of several previous news which came like two years ago. It talks about two things an i-Line machine which uses a mercury UV lamp light source produced by the Institute of Optoelectronics which can make 22nm details. This machine can only be used for low rate production since the light source has low intensity. This means it requires a much longer exposure time than what is typically used in production environments and would not be cost effective for mass manufacturing. The video also talks about the SMEE DUV machine which does ArF immersion lithography using an excimer laser but there are more recent news which are not talked about in that video.
 

Top