Printed Paper Is Cheap, Cheerful, And Ubiquitous. It's The Bedrock Of A Billion-Dollar Global Industry. And If MIT's Joe Jacobson Can Work Out A Few Little Details, It's Over.
Let's talk about something really dull.
Ink.
"It's everywhere!" Joe Jacobson exclaims, waving his arm to include his office, the MIT campus outside, and the world beyond. "It's in books, in magazines, on T-shirts, on wallpaper, on software boxes - it's even on the keys of your laptop!"
Yes indeed, ink permeates modern life.
But suppose there's a way to replace ink as we know it. An invention that revolutionizes publishing and threatens the global paper industry. Something so basic yet so flexible, it changes everything from books and newspapers to wallpaper and package design.
Now suppose this is your idea. You wrote the patents; you have sponsors waiting to market it. And you're 31 years old.
That's why Joe Jacobson is excited.
Long before the first primitive dot matrix printers started cranking out phenomenally ugly text with a resolution marginally better than braille, it's been a basic tenet of printing that words and pictures can be broken into dots. The corollary is this: to get higher quality, you use smaller dots - a process that digital technology makes trivial. In fact, quality is no longer an issue; at 2,540 dpi - dots per inch - the digital heritage of the text you're now reading is invisible, even under a magnifying glass.
Most of the rest of the printing business has been tougher to modernize. Paper is made by pulping trees in a manner the ancient Egyptians would have understood. Even the most modern offset press drags sheets under rollers and marks them using methods that haven't changed in a century. As for the ritual of trucking around books, newspapers, and magazines, it's so inefficient that half the copies of a typical paperback are thrown away, unsold.
Obviously electronic storage and distribution are the answer. Obviously! That's why for more than 20 years people have been proclaiming that print is dead.
Of course, it isn't - because paper has entrenched advantages. There's tangibility: pretty much everyone prefers paper to screens. You can carry it around; it's compact, it's convenient and doesn't break when you drop it. It doesn't need outside power. It's cheap and totally reliable. In other words, it's everything a laptop computer is not.
But what if we could build a display so thin, it would literally be a sheet of paper? Something that mixes data storage and portability in a simple, rugged package. Something that looks good, feels good, is infinitely reusable. Something you can pipe text and images into, from any digital source - the Net, the local library, your computer.
Now suppose we bind a few hundred sheets of this clever paper to create an electronic book. Unlike a laptop, it wouldn't need memory, just a little CPU powered by a couple of AA cells in the spine. Unlike a screen, you could actually remember where your place was, and instantly flip to anything you needed.You could dump it in your backpack, take it to the beach, read it on the bus going home - and anytime you like you could instantly erase the contents, then load in something new.
The idea of an e-book has been around since the late 1970s, when researchers at Xerox PARC got on the case. Their prototype used millions of little magnetic particles, black on one side and white on the other, loosely embedded in the surface of a soft sheet of rubber. Electric charges made the particles flip, creating an effect like pixels on a video screen. Clunky and cumbersome, it went nowhere.
Unaware of his distinguished predecessors, Jacobson, a physics PhD doing postdoc work in quantum mechanics, stumbled onto the idea in 1993. In classic hard-science fashion he was looking for something really challenging. "A book that typeset itself sounded difficult enough to be interesting," he remembers.
He spent a day sketching possibilities. "First I had a strange, crazy idea of fluid particles being pumped around in tubes. There would be a little gate determining whether a black or white particle went into a long tube." Then a variation on the PARC idea hit him: why not use reversible particles, each encapsulated in its own transparent shell? If the spheres were small enough - around 30 microns, like grains of laser toner - they could be strewn across a sheet of paper and glued into its fibers. Millions of them.
That left one overwhelming problem: how to control the tiny "pixels" without a rigid screen or a tangle of wires. "Two known materials are both transparent and conductive," says Jacobson. "There's indium tin oxide. And there's a new class of polymers, a type of vinyl. We normally think of vinyl as an insulator, but if you dope it with a metal, it conducts electricity." So there was the answer: Prime a sheet of paper with millions of tiny, two-toned particles. Coat it with a grid made up of thousands of lines of flexible, transparent, conductive ink. Apply carefully controlled electric charges and - in theory - watch as text, images, or anything else you like snaps into focus.
Rotating microspheres the size of toner particles? Conductive traces printed with invisible ink? All sandwiched in a structure not much thicker than, say, the cover of this magazine? As Jacobson says, it's certainly difficult enough to be interesting.
Consider one basic problem: to display text at 150 dpi (about twice as sharp as a typical video monitor, but half the resolution of a cheap laser printer) requires a grid with 150 vertical traces per inch and 150 horizontal traces per inch, in separate layers with an insulator in between. On a microchip or a screen, it's easy. On a sheet of paper, it's daunting.
Or how about this: Jacobson insists his special power-conducting ink will be robust enough to handle daily use. OK. But what happens when hundreds of e-pages are bound together? Now we have hundreds of thousands of delicate traces, all converging in the spine and its cheap little CPU.
Solution? Give each page its own onboard microprocessor built up out of ... ink. "A silicon wafer is too thick and too expensive to integrate into paper," Jacobson says, "but we can print insulating material, conductors, and resistors, plus two more things that we believe no one has printed before - p material and n material (the basic components of a transistor). We've already printed the first pn junction, a diode."
When each page carries its own computer, you have "intelligent paper," and the possibilities cascade. Air tickets that self-validate when you write a PIN on them. Packages that flash a message when you pick them up. Notepads that automatically keep a copy of your memos.
Or, as Jacobson envisions, an all-purpose electronic book. There can be formats for every occasion, from pocket-size to tabloid, kidproof, weatherproof. And a couple of AA cells will run it - once charged, the microspheres need virtually no power to maintain their display.
But why make a book at all? Why not just a good screen?
A one-word answer: navigation. "When you turn a page," Jacobson says, "you don't lose the previous page." Our sense of touch, he says, maps to the brain, making things easier to remember and relocate. And the act of turning pages helps a reader stay oriented in a long text, or even a long argument. How's your laptop at that?
Fluid solutions
All that leaves just one little problem: a working prototype. To see the progress so far, I'm following Jacobson through an open-plan clutter of desks, video screens, and cable runs, down into the basement of the lab's famous Bauhaus-style building.
Suddenly, we're in a hardware world. "In the past," Jacobson explains, "the lab emphasized software - integrating computers with sensory input, that kind of thing. Four years ago, it started getting more fundamental when Neil Gershenfeld, who's a physicist, started a section called the Physics and Media Group, doing research ranging from desktop nuclear-magnetic resonance imaging to quantum computing." Gershenfeld brought in Jacobson - he was doing a Stanford postdoc fellowship - in September 1995 after hearing his ideas for electronic ink.
We walk into a windowless room about 20 feet square. There are lab benches, oscilloscopes, tensor lamps, test meters, and industrial gray steel shelves cluttered with parts boxes and reference manuals. Atop one bank of shelves sit a couple of pairs of sneakers. Chris Turner, 22, and Barrett Comiskey, 21, are waiting for us.
Comiskey, a math student with a youthful face and a severe haircut, is working on the display. "Math is a flexible degree at MIT. I was doing machining and programming when Joe came in looking for someone to fabricate black-and-white particles using vacuum deposition. I've been working on that ever since, for a really long time. It's been ..." He thinks for a moment and shakes his head in dismay. "A whole year."
Didn't he feel unprepared, tackling such a challenge in a field he knew little about? "The area I 'm researching is so narrow, there's hardly anyone who knows anything about it. So really I was at the same level as someone with a PhD in chemistry."
The process he's focusing on now is electrophoresis - moving particles through a fluid under the influence of an electric charge. It's a standard technique, used for things like analyzing proteins and depositing oxides in vacuum tubes.
Comiskey shows me a flexible rectangle of clear plastic, with copper edges and a wire emerging from each side. When voltage is applied, millions of microscopic white spheres migrate through the dark blue liquid. When they move up, they become visible, and the "screen" turns white; when they sink, the dark blue fluid hides them. "This display is very high-contrast," says Comiskey, "and stable in both states, so the particles stay where they're put. It's also ultra low-power."
Comiskey's goal right now is to shrink that process by encapsulating bunches of tiny white spheres in individual fluid-filled shells, millions of which could be spread across the pages of an electronic book - tiny pixels that would turn colors when a change is applied. It could work, but it's far from clear whether that's the best solution. "We're pursuing different ideas in parallel," he says. "There are chemical issues to resolve. There are viscosity issues, density issues - but there's no fundamental obstacle."
Chris Turner is devising the circuits that would control Comiskey's particles. He shows me a little component board. "Previously, the electrophoretic display required 200 or 300 volts. By going to smaller particles we can get it down to 10 volts or so. The next stage is to reduce components like this to circuits that can be printed on the page."
He sounds casually confident. But how do you do that on a mass-scale, without billion-dollar, "clean room" facilities?
"It's untested," he admits. "Chips are made in clean rooms - a pure environment. Very little research has been done on other approaches. But we think we know how."
Creative paranoia
Back upstairs, I sit in on a lunch meeting with a dozen students. We gather around a long, Media Lab-gray Formica-topped table, talking about ways to dye the particles.
Jacobson sits at one end, wearing beat-up shoes, old corduroy pants, and a wrinkled navy Ralph Lauren shirt with the cuffs unbuttoned. He rests an ankle over his knee and discusses science while consuming a sandwich with lip-smacking gusto. He's a mix of contradictions: the vigorous, earthy style of a jock with an intensely analytical, wildly creative mind. He's amiable and good-humored but also cautious, verging on paranoid. When I first arrived, he asked me to sign an email message, in which I had already agreed to his terms of confidentiality. Later, after the lunch meeting, I ask for a copy of a presentation he had made on his laptop. He carefully removes some of the illustrations before handing it over.
When I ask him where he's from, he says, "Bawun and bred in Bawstun," with a cheerful grin. But his office, five miles across the river in Cambridge, looks as if he still hasn't moved in. Cardboard boxes, software packages, and some heavyweight reference books are scattered around. There's a Cannondale mountain bike fitted with lightweight tires against the wall, a squash racket, and a hanger with a couple of spare shirts. An IBM ThinkPad sits on the table, a PowerMac tower lurks underneath. Family snapshots are tucked around the edge of the window, looking as temporary as the rest.
It's odd to think that this office belongs to someone sitting atop an idea that could be worth billions. Jacobson looks uncomfortable when I ask about his financial prospects. He hesitates. "This is a fun project," he says finally, as if that answers my question.
The digital ink project is being funded in part by the Media Lab's Things That Think consortium, whose 41 companies - including Compaq and Microsoft - pay up to US$150,000 a year. Other backing comes from the lab's News in the Future group. Participating companies have the right to use any patents assigned to MIT without further royalties. Anyone else has to pay. And given that electronic paper could replace a large proportion of books and newspapers as they exist today - not to mention a lot of printers and laptops - those payments could be huge.
By long-standing policy, MIT owns all intellectual property created by its employees - when Jacobson writes his patents, they are assigned to the institute. But MIT policy also specifies that royalty income is usually split into thirds among the inventor, the school, and the Media Lab; the lab always passes its share on to the inventor.
How about his students?
Comiskey has coauthored one patent with Jacobson, but he sounds bored by the subject. "None of us thinks about money," says Comiskey, whose own ambition is to set up a community-based engineering cooperative. "None of us is driven by that at all." At this particular moment he's focused on a microscope, peering at tiny little spheres. "I'll have some crazy little problem," he says, "I'll mess with it for a week, and then I'll solve it. Who knows how many problems there will be? I only see the big picture when visitors come in. Joe talks about that, I don't."
Turner's no better. "There's so much that has to be done before it's commercially viable," he says. "In any case, the chemistry stuff is more patentable than the electronics that I do." He really does sound indifferent.
Go wireless
Jacobson may be reluctant to reveal details, but he loves talking up larger concepts. One obvious enhancement could be photographic half-tones, which might be simulated using an alternating current to flip the microspheres very rapidly, with unequal on and off cycles. A better option might be to mix different sizes of particles, with each responding differently to varying charges.
With its onboard computer memory, an e-book could also accept user input. "You should be able to write notes in the margin," Jacobson says. "You'd do this electronically, with a stylus - and the notes should stay with that particular text when you save it to a disk. Another intriguing thought is to create books by drawing material from different places on the Web, on narrow topics where books may not exist. The genetics of Arabian horses, for instance - something like that. A sufficiently intelligent crawler would create your own book automatically."
Ultimately, he wants the book to be able to capture new text out of the air. "We want to make radio paper," he says. "Print a complete radio circuit on the paper itself so that it can receive news via FM-sideband transmissions, like a pager network."
Don't radio receivers need inductance - some kind of coil? How do you print that?
"We ... have a way." He smiles enigmatically.
There are other possibilities for displaying graphics and pictures. In one system developed by Jacobson's students - for a slightly different concept, reusable paper - a single dark dye is used to display monochromic images. "With a thermal print head we have a way to print today's newspaper onto a piece of paper. When you've finished reading it, you pass it back through the printer, and make a new newspaper. Everything is reused."
Toner manufacturers and paper companies aren't going to be too happy, but Jacobson makes it seem inevitable. A similar dye-based technique can create poster-size output. "We had a variation of this up and running, using an eight-inch print head from an industrial label printer, which takes several passes at the image, creating it in stripes. Then a cooling head erases the paper. The ink should support from 10,000 to 100,000 reversals. We're not sure, because we haven't been able to change it enough times to wear it out."
All these processes are monochromic, but naturally there are ideas for color. Color would enable you to put a different fine-art reproduction on your wall every day. You should download it from an online database. With a more sophisticated control system, you could have a huge wall screen displaying animated images. Taken to its extreme, you could paper an entire room with moving images.
But Jacobson isn't satisfied with reversible- dye solutions. "Candidly, I hope they don't reach the marketplace," he says, "because I regard them as intermediate technologies. We can make a sheet measuring 11 inches by 14 for under 20 cents. But this is only preferable so long as a purely electronic display is much more expensive. What if an electronic page costs just a few dollars? That's the route we really want to take."
He points to his ThinkPad. "This screen draws 2.5 watts and has an OEM price of $1,000," he says. "It's expensive because a liquid crystal won't retain its polarization unless you latch it with a transistor, and whenever you have a million transistors on such a large substrate, the cost is always going to be at least $400. Metal-insulator technology (MIM) could cut this in half, but it will always be several hundred dollars."
Jacobson's goal is a display that will slash power consumption by a factor of 100, and to do that he doesn't see a plausible alternative to digital ink and paper. Minolta, Sharp, and Xerox are funding research into improved, cheaper displays, "but they're all flat panels using two pieces of glass, plastic, or elastome," he shrugs dismissively.
Resolution is still a problem. At 150 dpi, elegant type can't be smaller than 12 points in size, and anything below 10 points looks ugly and is hard to read. Theoretically, electronic ink particles can be small enough to allow 1,000 dpi. The problem would be printing 1,000 transparent lines per inch - meaning pages with millions of pixels, requiring heavyweight computing power.
That's one reason Mark Hatch, who looks into the future for Avery Dennison (those enterprising folks who make so many different kinds of labels) isn't particularly worried - for now. "Digital paper will replace computer screens," he says, "but pads of paper won't vanish." Hatch cheerfully admits that paper's traditional storage function - keeping information in file cabinets - is already "dead meat." But for lovely posters and packaging - not to mention glossy magazines - he thinks paper's still-unsurpassed look and feel will buy it at least another 20 years. "As a presentation tool," Hatch says, "150 dpi hasn't got a prayer."
But for a billion-dollar global industry, that's hardly a vote of confidence. And what Hatch calls paper's "transportation" aspect - newspapers, trade magazines, old media's heart - is clearly ripe for the picking.
Jacobson thinks a working e-book is just a couple of years away, with a retail price of maybe $400. If he and his basement colleagues can do that - let alone extend the concept to everything from wallpaper to FedEx envelopes - it could launch an upheaval that will complete the revolution we've already witnessed with desktop publishing, laser printing, and, most recently, the Net.
Dull, obvious ink. "It's a little bit silly," Jacobson says. "Every day we get a newspaper delivered; we read a column or two and then have to get the paper recycled. Since no one likes to read a paper on a monitor screen, new inventions are needed."
He pauses thoughtfully. "You know, newspapers are a $50 billion market ..."
