User login


You are here

IBM Airgap Microprocessor

Xiao Hu Liu's picture

From IBM Press Room announcement on May 3, 2007: IBM Airgap Microprocessor -- This microprocessor cross section shows empty space in between the chip's wiring. Wires are usually insulated with a glass-like material. IBM has integrated self-assembly techniques, long confined to laboratories, with its manufacturing lines to create a test version of its latest microprocessors that use vacuum gaps to insulate the miles of nano-scale wire that connect hundreds of millions of transistors. The breakthrough reduces electrical interference, raises processor performance, and lowers energy consumption.

Taxonomy upgrade extras: 


Zhigang Suo's picture

Xiaohu: Thank you very much for this image. I saw an announcement on this advance in yesterday's New York Times, and was going to ask you about it. I must say, after reading the article in NYT, I still did not have a clear picture of the process. The phrase "self-assembly" will clearly raise curiosity in many mechanicians. Could you point us to a technical paper where this process is described?

Xiao Hu Liu's picture

Zhigang: You may be the first professor who read the announcement. But Honghui, who invited me to give a seminar yesterday afternoon at Mechanical Engineering at City University of New York, may be the first professor who saw the picture and heard the news from me. :) For more information regarding to the airgap processor, you can visit IBM webpage and the links on that page.

Henry Tan's picture

IBM Brings Nature to Computer Chip Manufacturing

Made in IBM Labs: 10 Chip Breakthroughs in 10 Years

Ting Tsui's picture


 That is a very nice pictures and surprised that IBM did not put it on their site.

I was looking at the gap structure closely. They seems to have columns of

dielectrics within the "tunnels" rather than one continuous opening.  Am I correct?

 Very interesting design structure.




Xiao Hu Liu's picture


The picture and announcement are from IBM Press Room. For your convenience, I have added the link where you can find more information.

Just noticed that you have moved. Congratulations on your new job, and wish you all the best.

Best Regards,


Jie-Hua Zhao's picture

Zhigang, Xiao Hu, and Henry:

      Thank you for bring us to the attention to IBM's newest invention. The airgap concept was around for while when the Industry was struggling to identify better low-k materials. This is the first time we see an announcement that the concept will be put into production.

      Since I am more concerned of reliability issues, I want to ask a question about the moisture issue.

      In packaging reliability tests, there is one called moisture sensitivity test. The packaged devices are exposed to moisture environment say 85C/85% relative moisture for 1 week. Then put the packages into solder reflow oven and heat them to 260C within minutes. If there is trapped moisture inside the device, the moisture will "pop-corn" the device. (ref. see e.g., A. A. O. Tay, and K. Y. Goh, IEEE-TRAN on device and materials reliability, vol 3, no.4, pp144-151, 2002)

     In the IBM new airgap device, the airgap is in the order of 20nm. I am not sure if the pop-corn effect will happen with this small crack length. Is there anyone in this forum wants to do a fracture mechanics calculation to estimate the risk of pop-corn effect?


Two quick questions:

 1. from your picture it appears the air gaps were partially populated between the upper level fat lines. Is there a scale limit to the min. gap? or the air gaps can be placed even at M1, which would be ideal from design perspective.

 2. How much improvement in the effective K can the air gap buy us? It's better be huge to offset the manufacturability/reliability issues.

As for the pop-corning in airgap, my understanding is that the current ultra low K CDO material has max pore size of a few nanometers already, so the question boils down to whether there exists a critical pore or gap size for pop corning to occur?


Ting Tsui's picture

I think we will have electrical fail first before it pops. The leakage current between line is very sensitive to the moisture absorbed. It takes much less water to kill the device electrically.

Good point. But hopefully the MOB (moisture and oxidation barrier) would do a good job to block the moisture out.

Zhigang Suo's picture

Here is another piece of news (rumor?) concerning IBM:  150,000 US layoffs.

Ting Tsui's picture

I read somewhere about it too.

Here is what I did before layoff or thought going to layoff.

1) complete health checkup and dental works (while insurance still paying for it)

2) load up all of the medications (while insurance still paying for it)

3) load up the 401k (try to get the matching fund)

4) Sell granted in-the-money stock options... Some companies do not allow you to carry options when you leave. If they walk you out the same day layoff happen, you can lose everything.

5) Pack and clean your office before they walk you out. You do not want the company to help you pack and go through personal items in your office.


Subscribe to Comments for "IBM Airgap Microprocessor"

Recent comments

More comments


Subscribe to Syndicate