1

I've heard about using particle accelerators to shoot particles a specific distance into a wafer to split it and it got me thinking, can we do the same with sounds? Link to a page talking about the particle acceleration method: https://www.theatlantic.com/technology/archive/2012/10/cutting-silicon-with-a-proton-accelerator-instead-of-a-saw/264176/

Would it be possible to generate a high frequency standing wave (or pulse) inside the wafer in such as way that the crystal cracks and forms a "perfect" flat surface?

I'm hesitant to this being possible because of the massive size difference between the crystalline atomic structure and "common" ultrasonic waves. But since we are talking about a solid here one should be able to use extremely high frequencies, right?

Once a layer in the crystal has cracked, would this not create a stress point and the crack would continue along this layer, giving you very flat surface even if the force was not applied completely uniformly?

And any possible attenuation could be used as an advantage by applying the wave directly to the surface you want to cut, only exposing a small part of the wafer to the strong waves, possibly even eliminating the need for some sort of standing wave configuration to make a precise cut.

Anyone knows if this is possible or even being done?

Beacon of Wierd
  • 79
  • "I've heard about using particle accelerators to shoot particles a specific distance into a wafer to split it" -> sources? – wcc Jul 07 '19 at 21:33
  • @wcc I added a link to a news article in the question itself. – Beacon of Wierd Jul 07 '19 at 21:54
  • I think you need to specify how you would like to cut the wafer. Do you want to dice the wafer into finely spaced rows like it was shown in the linked video? In any case, using sound seems highly impractical. Longitudinal speed of sound in silicon is ~9000 m/s, so if you make a 80 MHz wave, the wavelength is still ~100 um, and this is the finest resolution you can achieve, which is not impressive at all. – wcc Jul 07 '19 at 22:05
  • The idea was to just dice it one piece at a time, so you use a combination of waves to get a clean cut at just one location, then you make your next cut by either phase-shifting the waves or mechanically moving pieces around so you can get very thin slices. The question is more if the act of dicing using sound/waves would create clean cuts or just lots of disjoint cracks, and if it would be possible to precisely control where the cut happens. – Beacon of Wierd Jul 07 '19 at 22:44
  • @wcc also, the maximum frequency seem to be in the GHz range for solids, someone is talking about it here: https://physics.stackexchange.com/questions/23418/is-there-an-upper-frequency-limit-to-ultrasound So 80 MHz seems quite small and simply increasing the frequency could give much thinner slices. – Beacon of Wierd Jul 08 '19 at 00:37
  • I only used the 80 MHz because that is a common frequency used to generate sound waves in certain crystals with piezoelectric transducer (see acousto-optic modulators). I doubt if it would be easy to find a transducer on the market that can generate ~GHz sound waves. – wcc Jul 08 '19 at 02:12
  • @wcc Oh, ok. But how about shifting where the cutting occurs? Instead of trying to make a standing wave in the whole crystal and cut multiple places at once (where all the anti-nodes are) wouldn't it be possible to simply make one cut using a pulse or combination of frequencies and then shift that maximum by the desired thickness and make another cut? That would be sort of independent upon wavelength (except the pulse/maximum has to be localized). – Beacon of Wierd Jul 08 '19 at 02:34
  • To make a clean cut, the stress applied by acoustic power must be well-aligned along a crystal cleavage plane. You don't need the stress to be applied along the entire wafer - once a crack forms, it will propagate along the cleavage plane, and this is how the rudimentary scribe-and-break method works. To adjust the location of the stress, maybe you can have a phased array of transducers so that the acoustic wave focuses at one particular location, create a crack, adjust the phases, and then re-focus to a different location. But this still seems very impractical. – wcc Jul 08 '19 at 17:19
  • @wcc Cool, so it's possible then :D Might seem impractical, but it could be faster and cleaner than cutting it with a saw, and you don't lose any material. So it might be more economical :) – Beacon of Wierd Jul 08 '19 at 18:45
  • buying a high power IR laser may be cheaper than building a phased array of transducers.... – wcc Jul 08 '19 at 19:45

0 Answers0