Here is a guest blog from my colleague, Lee Teschler, Executive Editor, Design World magazine
3D-printing techniques can be optimized to create integrated circuits. The possible result: Printed electronics that make the IoT ubiquitous and put an end to the semiconductor industry as it exits today.
A paper presented at the prestigious IEEE International Solid States Circuits Conference in 2011 caused more than a little stir in the electronics industry. Researchers from the Belgian nanotech research center Imec described how they inkjet-printed a 3,381-transistor logic circuit made out of plastic. The number of transistors is noteworthy because it slightly exceeds the number of transistors in the Intel 4004, the world’s first microprocessor released in the early 1970s. The Belgian device occupied 2 cm2 and operated at a sizzling 6 Hz.
You might think most technologists would regard this work as a laboratory curiosity, seeing as state-of-the-art processors squeeze billions of transistors into the same space and run at speeds measured in gigahertz. But the trend of developments in plastic circuits has captured the imagination of some who see printed electronics as what’s called an exponential technology. Moreover, it could be one of the technologies that help realize the IoT.
Exponential technologies get their name because they experience exponential growth. Exponential growth is often exemplified by the classic chessboard legend: A vizier at the court of a great sultan won a bet with an inexperienced prince. As a prize, the vizier asked for a chessboard and a grain of wheat on the first square, two grains on the second, and so forth, with the number of grains doubling on each square. Thus the amount of grain on the chessboard squares grew exponentially. On the last square, the prince could only have fulfilled the prize by giving the vizier enough grain to cover the earth several inches deep.
Similarly, some technologists think the dimensions of plastic printed electronics will shrink in an exponential manner. The argument goes something like this: Today’s printed plastic electronics perform roughly as well as the state-of-the-art electronics of 40 years ago. If plastic electronics technology advances at an exponential rate, in ten years we may see plastic circuits operating at levels considered state-of-the-art just 20 years before, in 2004. If exponential growth continues, the year 2034 could see plastic printed circuits performing at levels equaling that of silicon circuits minted 10 years before in 2024.
Even better, plastic electronics cost much less to build than their silicon counterparts because their manufacturing processes don’t involve high vacuums or high temperatures. Most plastic circuit fabrication employs straight-forward printing through inkjetting and processing with relatively unexotic liquids.
“Plastic circuits could end up having perhaps 1/1,000th the cost-per-area of the equivalent silicon device,” says Dr. Janusz Bryzek, founder of the Trillion Sensor movement and a co-founder of nine Silicon Valley MEMS companies. If Bryzek is right, printed transistors could be 1,000 times cheaper than those fabbed conventionally in silicon. Thus he sees plastic printed transistors as a key enabling technology for the wearable electronics and other innovations that are an important part of the Internet of Things movement.
An even more important aspect of plastic printed electronics is the impact it could have on the semiconductor industry. Bryzek points out that super-cheap plastic electronics could eliminate the need to fabricate ever-smaller devices as a means of getting chip costs down. He thinks printable transistor technology may make it possible for almost any firm to print transistors with adequate performance on a 3D printer. If that’s the case, there may be less need for traditional semiconductor companies. The result could be a massive upheaval among chip makers, he thinks.
You might suppose it would be tough to come up with a 3D printing approach that could duplicate the micron-level features seen in today’s integrated circuits. But there are already developments on this front that look promising. For example, researchers at IBM’s Zurich labs reconfigured an atomic force microscope so it can create 3D patterns having a nanometer-scale resolution in organic material. One use for the patterned organic material is as a mask for creating circuits.
The IBM researchers basically created the device by adding a heater on the tip of an AFM probe. The heated tip desorbs material from a thin film of either an organic molecular glass or a polyphthalamide polymer (PPA). The glass or PPA material evaporates when the heated probe tip hits it, so there is no pile-up of excess organic material as the tip moves. The patterned material can then serve as a mask for lithographic methods of producing circuits on the substrate below. Alternatively, the 3D shape can serve as a mold for the replication of its features.
The device produces line widths having an accuracy of 10 nm, but it can make holes with a depth accurate to one nanometer. Initially, IBM wants to use the device to fab tunneling field-effect transistors (FETs) in III-V and graphene materials via a lithographic transfer method. Here, once the AFM device has created its patterns, the substrate goes through process that thins the entire patterned layer to produce holes where the pattern has been made. Then standard techniques etch through the holes and deposit materials in conventional ways.
IBM licensed its device to a Zurich start up called SwissLitho AG. The commercial version is called the NanoFrazor. It is said to outperform conventional electron-beam lithography equipment used in semiconductor manufacturing but costs much less – around $500,000 compared to e-beams which can run from $1.5 million to as much as $30 million.
SwissLitho says it is developing several enhancements of the technology, including machines carrying arrays of tips to speed up the patterning process. Also, it plans on creating models able to make patterns on a wider variety of substrate materials.