Receive the Blog by Email

Your email:

Omniprobe's FIB and SEM Blog

Current Articles | RSS Feed RSS Feed

Failure Analysis (FA) Tools for the Semiconductor Industry

by Cheryl Hartfield

Failure Analysis ToolsetEvery now and then, new technology is introduced that truly advances the state of the art of failure analysis.  Focused ion beam (FIB) microscopes and in situ lift-out (INLO) methods are one such example for the semiconductor industry.  Traditionally, FA tools involve hardware and software.  However, recent trends are proving communication tools can be equally valuable.  A June 2010 Sparkfun web posting resulted in a rapid identification of the true identity of a counterfeit IC.   This reveals that social media has the potential to be a powerful tool for the semiconductor failure analysis community.  Perhaps it is not surprising that counterfeit devices have reached social media.  Counterfeit devices have been a hot topic in the IC industry in the last few years, and this is a focus topic in the 36th International Symposium on Test and Failure Analysis (ISTFA).  The Sparkfun discussion is an interesting example of the power of social media to achieve business results, highlights the insidiousness of counterfeit devices (as well as the complexity of counterfeit confirmation), and reinforces why it is NEVER okay to let a chip shortage drive purchases from unqualified sellers.

Digital based discussions to solve technical issues are not new.  The Microscopy Listserver, sponsored in part by the Microscopy Society of America (MSA) and the Microscopy.Com SysOp, has been widely leveraged by microscopists for many years now. There are numerous subscribers from over 40 countries on 6 continents. While roughly 10-20 messages are posted and circulated daily on a variety of topics, the listserver is email-based without the ability to post and share photos.  Today's social media packs a powerful punch in the combination of live discussion, timely postings, visual media, and RSS subscriptions to control traffic flow.  Photos alone allowed the the Sparkfun community fan base to identify the counterfeit Atmel ATMega382 IC as an On Semiconductor NCP5318.

This is a stellar example of the effectiveness of "social media" in solving real technical problems when users and companies converse, unencumbered by the political correctness that would usually have prevented the posting of the original news item to start with.  Unlike the microscopists participating in the Microscopy Listserver, failure analysts have not extensively engaged the social community for problem solving, with the exception of the networking functions and user groups at the annual ISTFA conference held every November.  There is an Electronic Device Failure Analysis Society (EDFAS) forum for discussion available in the format of an EDFAS LinkedIn group , which is relatively young and starting to gain some momentum.  Perhaps one social media challenge that limits discussion within the failure analysis community is a culture of "failure aversion", promoted by upper management within the IC manufacturers.  No manufacturer likes to admit to having problems, much less discuss them openly, and the topic of failure analysis by its very nature is weighted towards this end.   However, with appropriate guidelines, this powerful tool of open discussion could be leveraged for failure analysis.  Possible guidelines include:

  • Anonymity exists, both for the person and for the company.  Any photos cannot show anything allowing a part to be identified (which would also reveal the company), UNLESS such information has already been publicly disclosed by the company.
  • Discussion is centered on FA method development and not on product failures.
  • Original posts have management's release approval.
  • A dedicated moderator exists to prevent "flaming" by a disgruntled employee or competitor

Social media enables easy commentary that today completely bypasses the traditional company approval process for information release.  Perhaps it is time for companies to formulate social media guidelines and training for their employees.  Resistance is futile, as the vast array of platforms for discussion (such as LinkedIn, Facebook, Twitter, and comments to blog posts) guarantee that discussions will happen.  If users and companies embrace and leverage the discussion, amazing things can happen.  User feedback to posts on this blog and others are highly encouraged to continue the advancement of topics that are valuable for FIB and SEM analysts, and the industries that they work in.

Image by Andrew Lopes

Comments

I believe that this type of forum (social networking and/or discussion groups/blogs) have the potential to be significant aids in developing and growing the Failure Analysis community. I am, frankly, amazed that in a very short time, there are already 280+ members of the EDFAS LinkedIn group. This has the potential to both increase EDFAS membership and to share best practices outside the confines of annual conferences. 
 
I agree that no proprietary information should be posted as this would have legal and ethical implications, but a frank discussion of generic problems (e.g. locating defects in 32 nm technologies, delayering areogel interlevel dielectrics, setting up to probe deep submicron devices and the artifacts involved in contact resistance and maintaining contact integrity, etc. should all be fair game. 
 
Also, there could be a potential for communicating with equipment developers and vendors for a free and frank discussion of needs and capability limitations of current and proposed instrumentation. 
 
Anonymity might or might not be needed -- perhaps just a blog name (superanlyst or something like that) would suffice. I agree that the horse has left the proverbial barn as far as social networking goes -- it's a done deal and employers need to accept that, given that proprietary information should remain sacrosanct in any discussions -- but, then, most of us signed non-disclosure agreements when we joined up, so we are all legally and ethically bound to keep confidential information in-house already.
Posted @ Thursday, September 09, 2010 12:48 PM by Richard J. Ross
Post Comment
Name
 *
Email
 *
Website (optional)
Comment
 *

Allowed tags: <a> link, <b> bold, <i> italics