User login


You are here

IBM Airgap Microprocessors enabled by self assembly (Video)

Teng Li's picture

An earlier post by Xiaohu Liu reported IBM's latest progress in microprocessors. IBM has figured out how to control and perfect the self assembly process to create trillions of tiny, nano-sized holes across a chip, which speed electrons that flow across wires inside the chip and reduce the power consumed by 15 percent.The following short video may help us understand a little bit more about the new technology. More videos, audio and images on this are available here (free, but registration needed)




Teng Li's picture

From the video, it seems (at least to me) the self-assembly process occurs in the thin cap layer above interconnects, while the airgaps are still formed through conventional photolithographical etching. Please correct me if I misunderstand.

Hi Teng,

   You can refer to the link:

The copper interconnects are patterned by lithography, but the airgaps are formed by self-assembly. But I am wondering how the hole comes with vacuum. Is it possible that the pressure outside crash  the hole(hollow pipe)?



Ting Tsui's picture

It is "vacuum" because the next processing step is dielectric(s) deposition usually perform at 300-400 C under low pressure (a few Torr). Once deposited the next layer, it will cap/seal the environment inside the gaps. It is not truly vacuum (10e-10 T). The pressure inside should be in a few Torr range. This goes the same with the nm openings in porous low-k. The nm openings should be in mT also. The espect ratio (height/width) of the air gap is pretty high. Don't think there will be any mechanical issues.

 It is quite important to keep the moisture out of these systems. This will causes copper corrosion and the leakage. What I am interested is the electromigration performances of these copper lines since there is no constraint to prevent metal extrusion.

Hi Ting,

   Got it. Thanks.

   The vacuum air gaps play as dielectric(insulator) between interconnectors. If the moisture comes in, will it change the dielectric properites greatly? Or, the moisture just results in "popcorn"?

    For most cases, the failure(crack, delamination, comes from the mismatch. Now the interconnectors are surrounded by vacuum, which means the mismatch disappears
. Is it possible that the vacuum improves the mechanical properties? 

Subscribe to Comments for "IBM Airgap Microprocessors enabled by self assembly (Video)"

Recent comments

More comments


Subscribe to Syndicate