I have never liked to hear the colloquial word “chips” as a shorthand term for integrated circuits, except when it is used between engineers who actually know what an IC really is. When it is used by the general public, journalists, and those who wouldn’t known an IC chip from a potato chip, I think it has a demeaning, diminutive implication that trivializes what it takes to designs and build an IC. After all, you can make basic potato chips at home, but producing ICs requires sophisticated tools, smart engineering, precision materials, and fab facilities which cost billions of dollars, all while pushing the boundaries of physics on that road map which we casually refer to as Moore’s “law.”
So imagine my surprise when I saw a report on work done by researchers at the University of Wisconsin where they used wood chips – actually cellulose nanofibril (CNF), a flexible, biodegradable material made from wood – as the scaffolding or substrate for an IC. They coated the CNF with epoxy to act as a moisture barrier and address the need for surface smoothness and controlling thermal expansion. On top of this they placed a layer of silicon or gallium arsenide.
The team has built and tested some impressive GHz-range GaAs ICs, and they report on the performance in detail. There’s a very readable summary of the project in the university press release “A new kind of wood chip: collaboration could lead to biodegradable computer chips” or you can read the lengthy, detailed paper in Nature , “High-performance green flexible electronics based on biodegradable cellulose nanofibril paper ” (the details of how they did this are actually quite interesting).
“Why even bother doing this?” is an obvious question, of course, given that both Si- and GaAs-based technologies are mature, well understood, and so darn good. The primary rationale given by the researchers has a lot do with going green and the “b-word”: biodegradable. They cite impressively scary numbers on waste and trash due to electronics and ICs, and maintain that their CNF-based ICs are environmentally friendly; this is in sharp contrast to the known hazards posed by GaAs and the general headaches of electronic discards. Project leader Jack Ma notes, “Now the chips are so safe you can put them in the forest and fungus will degrade it. They become as safe as fertilizer.” I’ll have to take him at his word for that, of course.
How valid is the argument that these wood chips are a better environmental deal than silicon or gallium arsenide? I’m skeptical, for sure. Most the electronic waste we generate – and it is a lot – is not due to the IC die themselves, but to the PC board, various metals, wire leads, plastics, epoxies, product enclosures, and the IC packaging. The die is a very small part of the story, so any comparison of “electronics” waste and ICs seems to be confusing two very different things.
My cynical side says that the focus on the green aspects is largely a play get the right kinds of attention and those critical follow-on grants. Further, RF ICs tend to be mostly analog and small-ish (whether using Si or GaAs processes) compared to the all-digital processor and memory ICs, so the potential savings would be further reduced.
However, there were two aspects of these wood-chip based ICs that do seem to be quite interesting and have potential. Due to the extreme thinness of the die, the researchers say the resultant ICs are both flexible and transparent. There are lots of applications that could conceivably benefit from those attributes, and there could be radically different product designs that could leverage one or both of these characteristics.
What’s the realistic future for CNF-based devices? I certainly don’t know. It’s appropriate to be skeptical, and it’s easy to be cynical. One thing we know is that there something almost “magical” about the electrical, mechanical, and thermal properties of silicon as a base material (and to a lesser extent, that of gallium arsenide).
Still, you never know what the next big breakthrough will be, and our ability to see it coming or predict it is pretty poor. Since these CNF devices do use silicon and GaAs as their top layers, perhaps they can provide the performance needed while still using many of the tools and techniques which are already available. At the same time, a high-performance, consistent IC in volume takes more than a good die; the leads, packing, and subtle mechanical, thermal, and materials issues all have major impact on performance.
What’s your guess or prediction on the future of these ICs which use CNF as a substrate? Of course, until we know more, we should revert to the cliché and say, “let the chips fall where they may!”