Home > Nanotechnology Columns > UAlbany College of Nanoscale Science and Engineering > Reliability in the nanoworld
 |
James Lloyd CNSE Senior Research Scientist UAlbany College of Nanoscale Science and Engineering |
Abstract:
Every new technology is met with challenges. One of the major challenges associated with the introduction anything new is one that is often overlooked by the population, unless of course it affects them directly, is reliability. If things are going well, reliability is never thought about, but when things are not, there is often hell to pay. Such is the nature of the beast.
July 29th, 2011
Reliability in the nanoworld
Every new technology is met with challenges. One of the major challenges associated with the introduction anything new is one that is often overlooked by the population, unless of course it affects them directly, is reliability. If things are going well, reliability is never thought about, but when things are not, there is often hell to pay. Such is the nature of the beast.
Because it is not generally perceived as "sexy," reliability work is done behind the scenes, but the effort is highly non-trivial. It has been argued that effort to meet reliability goals is one of, if not the most expensive line item in the development process.
To evaluate a process or product for long term reliability an in depth understanding of the physics of the failure process is required. A major practical problem is how to evaluate a product that must last 10 to 25 years in the field in a short time so that modifications to the process or design can be made. This requires what has been termed "accelerated testing" where mechanisms that would be responsible for device failure can be reproduced in the laboratory in a matter of days or weeks that would represent many years under field use conditions. One cannot make predictions from accelerated tests unless the physics of the failure mechanisms being accelerated are very well understood.
The microelectronics industry, which has been doing "nano" for several decades now, has been remarkably successful in solving these problems, but as in any new endeavor, Mother Nature has ways of surprising us. Almost 50 years ago, the industry was in for a big surprise in the previously unheard of phenomenon of electromigration. When very small "thin film" conductors were first made, they didn't last very long. Instead of many years of operation, the first integrated circuits lasted only weeks before it was found that the "wires" conducting electricity in the chips were breaking. The reason this was happening was that conducting electrons would collide with imperfections in the metal and push the metal atoms in the direction of the current flow. This had never been seen before because you could never get enough current in a normal wire before Joule heating would cause the wire to fuse. But, because of the small dimensions and the ability of the relatively massive Si chip to conduct the heat away, much higher current densities were used in microelectronic thin devices than were ever possible in macroscopic wiring. Whereas a current density of ~104 A/cm2 would be enough to fuse house wiring, thin film conductors on single crystal chips could easily sustain current densities of over 106 A/cm2 without immediately melting. However, at these high densities, electromigration became important, producing failures in a very short time. This is just one of the problems encountered and had to be dealt with as things became very small.
The further reduction of dimensions in what we have become calling "nanoelectronics" has posed new challenges, not unlike those we had seen earlier, but with the need for new clever solutions. Electromigration is still one of the issues, but now with wiring pitches (the measurement of the wiring density, expressed as the sum of a line width and the spacing between lines) of 25 nm planned in near future applications, the challenges are even greater. At the finest dimensions, the metal surface and other interfaces become more important. Interfaces are pathways for rapid mass transport, have unusual mechanical and electrical properties and contribute to instability in a number of ways. In addition, the introduction of new materials has led to considerable research work. All this has to be understood, characterized and incorporated into the design rules.
The descriptions above were for problems associated with electronics which are fairly well understood, if not completely in check. As we enter into other areas of the very small, new unanticipated reliability issues have become the norm and Mother Nature has not given up in finding ways to make life interesting. Many new failure modes in other fields like nanoelectromechanical systems (NEMS) systems or in nanobiology will need to be addressed. At CNSE we are building a center for fundamental research into the reliability of very small things. In order to fund this research, we are operating in an unusual paradigm. We provide reliability services to industry and the proceeds from these services will be applied towards fundamental studies of the physics of failure mechanisms for any and all applications we can anticipate. In this manner we hope to be able to maintain a steady stream of reliability funding without depending on the vagaries of the traditional funding sources.
Reliability science is always on the edge of what we know about materials properties. Often, our ability to deal with reliability issues is limited by our fundamental understanding of the physics of those materials. Probing into these issues is compellingly interesting and fun. When I've spoken to children in grammar schools about what I do, I tell them "I break things." "Who would like to do that?" The response is always enthusiastic. "Me…Me...Me"
Dr.Lloyd's Bio
Bookmark:
 |
||
 |
||
| The latest news from around the world, FREE | ||
 |
||
|
|
||
| Premium Products | ||
 |
||
|
Only the news you want to read!
Learn More |
||
 |
||
|
Full-service, expert consulting
Learn More |
||
|
|
||