Growing A Gallium-Arsenide Laser Directly On Silicon

As great as silicon is for semiconductor applications, it has one weakness in that using it for lasers isn’t very practical. Never say never though, as it turns out that you can now grow lasers directly on the silicon material. The most optimal material for solid-state lasers in photonics is gallium-arsenide (GaAs), but due to the misalignment of the crystal lattice between the compound (group III-V) semiconductor and silicon (IV) generally separate dies would be produced and (very carefully) aligned or grafted onto the silicon die.

Naturally, it’s far easier and cheaper if a GaAs laser can be grown directly on the silicon die, which is what researchers from IMEC now have done (preprint). Using standard processes and materials, GaAs lasers were grown on industry-standard 300 mm silicon wafers. The trick was to accept the lattice mismatch and instead focus on confining the resulting flaws through a layer of silicon dioxide on top of the wafer. In this layer trenches are created (see top image), which means that when the GaAs is deposited it only contacts the Si inside these grooves, thus limiting the effect of the mismatch and confining it to within these trenches.

There are still a few issues to resolve before this technique can be prepared for mass-production, of course. The produced lasers work at 1,020 nm, which is a shorter wavelength than typically used, and there are still some durability issues due to the manufacturing process that have to be addressed.

2 thoughts on “Growing A Gallium-Arsenide Laser Directly On Silicon

  1. How can you even do reaseach on semiconductor manufacturing those times? All of the equipment is at fabs which are guarded by heavily-armed private security. They’re not going to let some university professor derp around tinkering with their $5b machine when there are 3 years of NVIDIA, AMD or Samsung chips on backorder.

    Also, it’s not like it’s 1950s anymore and you just add a bit of germanium or boron to a hot pot with molten silicon and build a working IC at university.

    1. According to the article, this was done by IMEC, a world-leading semiconductor research company, presumably in their own state of the art fabs. They have 12,000 m2 of cleanroom. It’s not some “university professor derp”. All of which you could have found out if you googled them.

      NOTE TO ED: I’d not heard of IMEC; maybe link to their Wikipedia page?

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.