In the 1970s, PARC, a Xerox (XRX) company (formerly the Palo Alto Research Center), helped usher in an era of modern computing, as it built a graphical mouse-based operating system atop the foundations laid by Douglas Englebart's NLS. The PARC Alto would inspire both Steve Jobs and Bill Gates to develop the earliest consumer graphical operating systems, which continue to influence interface design today.
In the 2010s, PARC might again transform the world through technology. This time, the big step forward is "chiplets," a minuscule form of printed electronic circuitry that can be created with modified laser printers, another PARC innovation from the 1970s. Each little chiplet is no larger than a grain of sand:
A New York Times feature published earlier this week highlights the potential this advance has in the field of printed flexible electronics, which PARC estimates will explode from a $1 billion global industry today to one worth $45 billion by 2016. Some possible applications:
- Flexible and nearly unbreakable smartphones
- Pressure-sensitive robotic skin
- Sensor-enabled electronic (but disposable) medical bandages
- Large digital screens you can roll up, fold, and put in your pocket
- Low-cost durable circuit boards for a wide range of uses
These are just some of the ways in which printed circuitry might be used. As The New York Times feature points out, PARC's development represents a radical departure from the current chip-making model, in which a large wafer of silicon is processed, cut up into individual chips, and then reassembled into the various parts of a circuit board. Times writer John Markoff says: "The emerging printing technology poses a heretical idea: Rather than squeezing more transistors into the same small space, why not smear the transistors across a much larger surface?" The PARC laser-printing format (it's not exactly a laser printer, but it's quite similar) proposes to place sand-sized chiplets by the thousands onto the surface of flexible substrates, and these chiplets can be assembled to perform any of the diverse jobs undertaken in today's dizzying array of electronics by an increasingly specialized array of components.
The opportunity to upend chip-making is clear. Where does manufacturing come in?
This is a 3-D printer, one of the largest Stratasys (SSYS -2.50%) models available. It can print just about anything you can think of (within the limits of its print area), including a scale-model Aston Martin. One thing it can't do -- the single most important thing it can't do -- is print electronic circuitry. It's not designed for that.
Stratasys could print a plastic-based robot with articulated joints. Other printers, particularly the largest variants from 3D Systems (DDD -2.99%) and most of ExOne's (XONE) small product line, can print in metal, rubber, ceramic, and an increasingly diverse range of materials -- enough variety to create a functional robot. These companies may not even be necessary, as iRobot (IRBT -3.55%) recently filed a patent for a self-assembling 3-D printing robot manufacturing machine. All of these companies present some sort of opportunity to create a 3-D printed robot, but none of them can provide the brains. That's where Xerox's chiplets come in.
The greatest obstacle to 3-D printing's takeover of the manufacturing process (besides its slow pace of operation and difficulty in cost-efficiently scaling to mass production) is the fact that there's been no easy way to produce electronics. You can't print out a smartphone, because all the important components will still need to be plugged in. You can't print out a Kindle, or a new router, or a Roomba, or what have you. All you could ever print was the external casing, and that means nothing without circuitry to make things work. The need for external components won't be eliminated (you'd still need batteries, for example), but by enabling users of 3-D printers to create the circuitry for their custom devices, PARC could open up a whole new wave of manufacturing.
Manufacturing processes today take a number of diverse components and place them together just so to make a functional device. iRobot has hit on the next logical step -- 3-D printed devices with completely automated assembly -- but it still needs circuitry. A PARC chiplet printer installed next to a 3-D printer, with the requisite automation technologies, could finally make 3-D printing a viable manufacturing format for complex electronics. Since nearly everything we use now comes with some kind of embedded circuitry, the inability to replicate the electronic functionality of mass-manufactured goods in a 3-D printer would have remained a stumbling block even after the speed issue might be solved. Now, it's speed that remains the primary obstacle. Circuitry is within reach.
This is one step closer to my dream for 2025: distributed, small-scale, on-demand manufacturing. The 3-D printer simply doesn't make sense as the next laser printer. It's far more complex and requires far more technical expertise, and its output is less readily useful in entry-level consumer machines. However, when manufacturing becomes easier to do in high-tech automated warehouses stuffed with top-notch 3-D printers than it is in massive factories, the shift will be on. We've got a few years before that prediction's set to come true, and PARC has already pushed it in the right direction. Let's see if the 3-D printing companies can solve the speed problem in the meantime.