Communications of the ACM
Communications of the ACM
The Desktop Fab
Future chips, displays, and micromachines will be fabricated by printing. A version of the future fab will sit on your desktop, will be programmable by the user, and fabricate the things such as chips and micromachines we normally associate with billion-dollar fabs. This vision of the future has its basis in new chemistries, printing technologies, and design tools.
Why does a laptop cost more than a newspaper? The answer has to do not principally with the raw cost of materials but rather with our inexperience in manipulating atoms and molecules with great precision in parallel and at very low cost. A newspaper is made using continuous roll-to-roll fabrication with web-press speeds as fast as 1200 meters per minute, resulting in over one trillion impressions per year for the printing industry. Computer chips, on the other hand, are fabricated one wafer at a time. Presently it takes in excess of two weeks (seven days a week, 24 hours a day) of processing to create a typical computer chip like a Pentium. A new set of chemistries, printing technologies, and new design tools will allow logic and micromachines to be fabricated in several seconds by direct printing without the use of photolithography. In addition the time required from design phase to actual fabrication might similarly be cut from several weeks to as short as minutes.
Our group, The Nanomedia Group, at MIT's Media Laboratory is presently working toward enabling such a desktop fab for printing out both functional logic and micromachines. To date we have demonstrated the ability to print early versions of both logic elements [1] and micromachines [2] on printers that can fit on a desktop. The technology is based on new chemistries composed of small clusters (on order of 100 atoms) of an inorganic semiconductor, which are further covered with a cap of organic molecules that allow the clusters to be suspended in a liquid suitable for direct printing. When the clusters are printed onto a surface such as plastic or glass the organic molecules are removed leaving only the semiconductor, thus forming logical devices and/or micromachines without the requirement of either large and expensive vacuum deposition machinery or photolithography. Using special printing processes we have demonstrated printing resolutions of about 200nm, similar to the resolution of present top-of-the-line chips.
What are the implications of such a technology? The obvious implication is to be able to make logic vastly less expensively and in vastly larger quantities than what exists now. The least expensive logic circuits presently on the market cost on the order of tens of cents. As an example where price drives the ability to enable a market there is a great interest in being able to close the so-called e-commerce loop. The scenario is as follows: You are finished with a box of cereal. Ideally, you would like to throw it away and have another box appear on your doorstep the next morning. In order to realize this scenario the idea is to replace the UPC bar code on all packages with an RFID (Radio Frequency ID) tag that allows a computer to inventory an entire room by wirelessly pinging a small chip in the package. Considering the average cost of packaging for an item in the supermarket is approximately three to five cents, most estimates are that one requires a chip with a cost of about one cent. Such a cost is enabled by finding a new way to fabricate chips such as fabrication by printing.
Another area of importance is volume of chips. In order to put an IP address on every product produced as described here, one needs a means of manufacturing logic at this vast scale. Presently chips are typically made in the tens-of-millions up to appromimately one billion units in the case of smart cards. As mentioned previously, printing is a technology that allows trillions of impressions per year, a number commensurate with the number of packages and products produced.
Imagine if you could download different designs for logic and micromachines modules, put them together to form some completely new invention or design, and then print out a working version of it on the desktop.
These aspects of cost and volume, as important as they may be, may not be the most profound implications for this technology. As described earlier, a version of this technology would allow an atom fabricator to sit on the desktop...In other words a desktop fab. Imagine if you could download different designs for logic and micromachines modules, put them together to form some completely new invention or design, and then print out a working version of it on the desktop.
Computers and the Internet have had such a tremendous impact precisely because they allow a large number of people to be involved directly in the creative process of programming, manipulation, and the design of bits. Unfortunately, with the limited exception of desktop printers and people who have access to billion-dollar fabs, most people have no way to be involved in the creative process of designing and building with atoms.
The key to invention (as well as genius) is the iteration time through ideas. It is difficult to predict the impact of such a technology but if computers are any guide in the world of bits, what an interesting world it might be if we put a desktop fabricator on every desktop.
Without libraries what have we? We have no past and no future.
Ray Bradbury, science fiction writer
References
1. Ridley, B.A., Nivi, B., and Jacobson, J. All-inorganic field effect transistors fabricated by printing. Science 286, 746 (1999); www.media.mit.edu/nanomedia
2. Fuller, S. and Jacobson, J. Nanoparticle MEMS. In Proceedings of the IEEE MEMS conference. (Tokyo, Japan, Jan. 2000).
Figures
Figure. In the KidRoom2 experience, children enter a fantasy bedroom where the bed becomes a raft and the walls turn into a raging river. The exhibit was created by interactive entertainment developers Nearlife, Inc. ©2001 by Nearlife, Inc.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2001 ACM, Inc.
No entries found