This time it is not a virtual one! In terms of the EUV light source required by the lithography machine, Continental may really be ahead. The SSMB (steady-state microbeam) light source scheme proposed by Tsinghua University can achieve the output of high-power and high-quality EUV extreme ultraviolet light, and the power of the light source can easily exceed 1kW, which is more than twice that of ASML's upcoming 2nm lithography machine, and there is room for further improvement.
The key is that the light quality output by the SSMB light source is surprisingly good, and through different configurations, it can play a variety of frequencies of narrowband light, in addition to the commonly used 13.5 nm extreme ultraviolet light, SSMB light source from terahertz microwave to 6 nm soft X-ray can be output, which may further unlock the ability of the lithography machine, and even trigger a new chip technology revolution.
Why can SSMB light source be so good? That's a long story. As the saying goes, there is no harm without contrast, let's first take a look at what light source the famous ASML company uses. The shorter the wavelength of the light source of the lithography machine, the higher the resolution, the finer the engraving, and the smaller the chip manufacturing process. At present, ASML's most advanced EUV lithography machine uses a laser plasma light source (LPP-EUV), specifically a tin steam light source, which can output 13.5nm of extreme ultraviolet light, thereby producing high-end chips with 5nm or even smaller processes.
This tin steam light source is very complex, it will continue to emit a small tin metal droplet, and then use a high-power infrared laser beam to hit this droplet, instantly evaporate it into tin steam (plasma), and then aim the laser or electron beam at the tin vapor, excite the atoms in it, these excited atoms will jump back, will radiate 13.5nm extreme ultraviolet light.
The tin steam light source has achieved great success in the lithography machine, but it is not without shortcomings, the more significant is that the power is lower and it is not easy to further improve. The greater the power of the light source, the higher the efficiency of manufacturing chips. At present, ASML is trying to achieve the level of 500W light source, which is quite difficult. If you want to further increase the power of the laser plasma light source, it is quite limited.
Another major disadvantage of this light source is poor coherence and low monochromaticity, that is, the wavelength of the generated extreme ultraviolet light is not too concentrated around 13.5nm. This disadvantage makes it necessary to use multiple mirrors with multi-layer coatings to reflect the beam multiple times to purify the energy spectrum and obtain a purer 13.5nm extreme ultraviolet light.
These two major shortcomings actually limit the ability of LPP-EUV light sources, and according to the current demand for chip manufacturing process and production capacity, sooner or later it will reach the limit of this light source. So people think of other types of light sources, often mentioned synchrotron radiation light source and free electron laser.
Synchrotron radiation is produced by accelerators. When charged particles in large accelerators are deflected by a magnetic field, synchrotron radiation (SR) is generated along the tangent direction of the motion, while energy is lost. This synchrotron radiation was once considered "harmful" because it increased energy loss and limited further increases in particle energy. But scientists soon discovered that synchrotron radiation light had many advantages.
For example, high brightness, synchrotron radiation light has high radiation power and power density, and the X-ray brightness of the third-generation synchrotron radiation source is hundreds of millions of times that of the X-ray machine. Others include wide band, narrow pulse, high polarization, high purity, etc., which are powerful tools for scientific research. Since the SR light source can achieve high brightness, isn't it ideal to use it as a lithography machine light source?
There is indeed such a prospect, and even some netizens have envisioned a future synchrotron radiation lithography machine: using the storage ring of a large electron accelerator to draw out multiple synchrotron radiation beams at different locations and carry out lithographic production at the same time. Although large accelerators have a large footprint, which can be hundreds of meters in length, and the investment is high, taking advantage of their high brightness can greatly improve production efficiency.
This kind of "big factory 1 tow N" lithography factory is very desirable, but in fact, there are many roadblocks. Although the brightness of the SR synchrotron radiation source is high, the coherence is poor. The word coherence is not so easy to understand, in layman's terms, the electron beam group in the accelerator storage ring is not concentrated enough, and the synchrotron radiation light emitted cannot form a resultant force, resulting in the average light source power is not as high as expected.
Another light source, free electron lasers (FEL), can achieve higher peak brightness and greater coherence than synchrotron light. The free electron laser uses a linear accelerator to accelerate electrons to close to the speed of light, and then passes through a row of "oscillators" composed of torsional magnets, and the electrons twist around regularly inside, producing synchrotron radiation, which in turn acts with the electron beam itself, causing the electron beam to converge, producing coherent radiation light with high brightness and very concentrated wavelength, and the brightness is 8 to 10 orders of magnitude higher than synchrotron radiation.
Free electron lasers
But the average power of this light source is limited by the linear accelerator. The linear accelerator cannot produce enough electron beam pulses in a short time, so that although the peak brightness of the free electron laser is high, the average power is relatively low because the number of pulses per unit time is not enough, which is also not conducive to lithography.
So can it not only produce narrow band radiation light with high peak brightness, but also increase the number of pulses per unit time? Scientists at Tsinghua University came up with a brilliant solution: combining the synchrotron radiation of the accelerator with the free electron laser, a magical effect was achieved, achieving the output of EUV extreme ultraviolet light with high average power and excellent monochromaticity! This is the SSMB steady-state microbeam light source, how is it realized?
As mentioned earlier, the problem of synchrotron radiation light sources is that the electron beam group in the accelerator storage ring is not concentrated enough, and the beam group is relatively long, in the order of millimeters to centimeters, which affects the quality of the output beam and restricts the improvement of brightness and power. Then the solution to this problem is, of course, to shorten the length of the bundle, and the means used is to move the set of free electron lasers to the ring electron accelerator.
In a straight segment of the accelerator storage ring, add a modulated laser beam to it. In order for the electron beam group in the accelerator to interact with the laser, a row of torsion magnets is added to cause the trajectory of the electrons to twist. After laser modulation, the electron beam cluster rotates back again, and the beam size is amazingly shortened, a process called "micro-beaming".
If the length of the electron beam group is reduced to 3nm, the output of 13.5nm extreme ultraviolet light can be achieved, and the brightness can reach 10 billion times that of the synchrotron radiation source! Because the accelerator storage ring can achieve continuous replenishment of energy, and compared with the free electron laser with insufficient number of pulses, there are a large number of electron beam groups in the storage ring, which can continuously produce beams, which is also called SSMB "steady-state micro-beam".
The Tsinghua University team has carried out a verification test on the accelerator storage ring of the Helmholtz Berlin Center in Germany, using a 1064nm infrared laser to modulate the electron beam group, and detected the strong coherent radiation output at the incident laser frequency and its higher harmonic frequency, verifying the feasibility of the microbeam theory.
The Tsinghua University team designed the SSMB light source scheme accordingly, the accelerator storage ring circumference is 100~150 meters, the electron beam energy is greater than 400 megaelectron volts, and the beam group length is compressed to 3nm, thereby generating radiation power greater than 1 kW at the 13.5nm ultraviolet wavelength, which easily exceeds twice that of the ASML tin vapor light source. Moreover, the quality of this extreme ultraviolet light is far better than that of the traditional tin steam light source, the monochromaticity is better, and the requirements for the mirror are lower, which can fully meet the needs of large-scale chip production.
What is even more exciting is that SSMB steady-state micro-beam can not only produce 13.5nm level of extreme ultraviolet light, but also emit longer or shorter wavelength beams by adjusting parameters, such as 6nm soft X-rays, thereby creating a super strong lithography machine with a higher resolution, realizing a chip process smaller than 1nm and 2nm! Of course, due to the special nature of X-rays, the penetration and absorption rate are relatively large, and it is necessary to break through the problem of mirror manufacturing first.
In short, if the Tsinghua SSMB light source solution is realized, Continental lithography technology is likely to achieve curve overtaking, surpass the seemingly unshakable ASML in one fell swoop, and achieve a leading position in the chip manufacturing industry. At present, the steady-state micro-beam (SSMB) extreme ultraviolet light source project of Tsinghua University has been declared as a national major scientific and technological infrastructure, and is planned to be built in Xiong'an New Area.
However, it should be pointed out that the device is only a light source, not a complete lithography machine like the Internet. Even if the problem of extreme ultraviolet light source is solved through the SSMB scheme, it is necessary to further open up the upstream and downstream chains to create a real new generation of EUV lithography machine, and look forward to this day coming as soon as possible!