The New Alchemy

, who dubbed the new material wellstone in novels like The Collapsium (Del Rey, 2000). Now McCarthy has written his first nonfiction book about programmable matter, Hacking Matter: Levitating Chairs, Quantum Mirages and the Infinite Weirdness of Programmable Atoms. Associate Editor Stephen Cass talked to him about this bleeding-edge technology and how McCarthy himself is helping to transform science fiction into science fact.

What is programmable matter?

Programmable matter is fundamentally a solid-state technology—something that can change its optical, physical, magnetic, or electrical behavior without any moving parts except for electrons or photons. In that sense, there are certain things now that already qualify as programmable matter, like an LCD [liquid-crystal display] screen. This is an assembly of devices, but you can also look at it as carefully arranged material that has the interesting property of changing color under electrical stimulation. By adjusting quantum dots instead of pixels, you can make artificial atoms and adjust a lot more than just the color of the material.

What are quantum dots and how do you use them to make artificial atoms?

A natural atom is a particular means for confining electrons—the positively charged nucleus gathers electrons around it and doesn't let them escape. By confining the electrons, you force them to behave as standing waves. And those standing waves are responsible for nearly all the chemical, electrical, and optical properties that we associate with atoms.

But you don't have to have an atomic nucleus to get that sort of behavior out of electrons; you just have to confine them in a small space. There are a lot of ways to do this. One way is to use the standard techniques of semiconductor chip design to create junctions that will herd electrons into an area of choice, known as a quantum dot. Once confined, the electrons will form a structure known as an artificial atom. With artificial atoms, unlike natural atoms, there is no reason why you can't pump electrons in and out and change their characteristics dynamically, making them programmable.

But if these programmable atoms are buried in a semiconductor substrate, how do they interact with anything? How do you make the entire material behave like it's made out of, say, gold?

With programmable atoms in a substrate, what you are really doing is creating controlled impurities—dopant atoms—so the properties of your semiconductors are going to be very important in determining the final properties of the programmable substance. You can get a very high level of doping with a properly designed quantum dot array and overwhelm the normal behavior of the semiconductor. You can never ignore the fact that the semiconductor is there, but you can change its properties almost beyond recognition.

So would you have to combine different types of artificial atoms to end up with a material whose net behavior is like that of gold?

Probably. An artificial atom of gold— pseudo-gold—is almost certainly going to be a lot larger than an atom of natural gold. One consequence of this is that its absorption and reflection spectrum will be redshifted, because the electrons are less tightly bound so they will be at lower energies. So even if you could somehow have atoms of pseudo-gold without any substrate, they'd be a different color. To get something that looked like gold, you'd have to produce something that was a little bit different in order to mimic the optical characteristics that you want.

All the quantum dots so far developed in laboratories exist as structures at or near the surface of the semiconductor. How do you go from a programmable film of artificial atoms to a bulk programmable material?

It's true that a major limiting factor is that you can only affect the surface properties of the semiconductor by doping it with artificial atoms. But my business partner, Gary Snyder, and I thought up a solution. Imagine a thin, but very long, semiconductor that has a two-dimensional array of closely spaced quantum dots on its surface. If I form that into a cylinder, I've created a fiber, and a fiber has the property that it's mostly surface.

If I weave these fibers together into a bulk material, I can control the quantum dots on the surface of every fiber and thereby control the bulk properties of the material. I think this idea hadn't occurred to people working with quantum dots because they hadn't envisioned many material science applications.

How do you control all those quantum dots on all those fibers?

That is the manufacturing challenge, the primary engineering difficulty that separates this from today's reality. In some of our back-of-the-envelope designs, we only need a thousand—or fewer—control wires running through the middle of the fiber: you can control the properties of an artificial atom with a large amount of precision using as few as four to 16 electrodes. The atoms are addressed in repeating groups of around 20 to 64, and each group is sent the same signal.

So, for example, I'll be able to tell one atom in a group to emulate helium and the adjacent atom to emulate iron, producing a sort of repeating checkerboard pattern over the entire surface of the fiber. The fibers interact to produce a bulk doping of helium and iron. Of course, I still have an interconnect problem at the ends of the fiber, where I have to bring these control wires out and interface them with a larger regular chip that provides the power and control signals. That involves a large number of connections. But a lot of things we do today involve a large number of connections.

Could you make programmable matter stronger than the semiconductor substrate, so that it could be used as a structural material for buildings or vehicles?

The binding energy between two artificial atoms is going to be limited by the binding energy of the substrate. An analogy is that if you build something out of Lego bricks, you're inherently limited by the bond strength between two Lego bricks, which is strong in compression but very weak in tension. But if you put an iron bead in the middle of every Lego brick and apply a magnetic field, you can keep your Lego structure always in compression and stop it from being pulled apart.

Similarly, silicon is actually a pretty tough material. Particularly in compression, it's inherently a lot stronger than some of the normal building materials we use today. If you can generate artificial atoms with the right magnetic properties, you could keep silicon under compression and make it stronger in tension. Conversely, a substrate that is weak in compression could be kept under tension. So that is a potential way to create stronger materials.

Beyond changing bulk physical properties, could you do any chemistry with programmable matter?

You can certainly perform chemistry. However, because artificial atoms tend to be larger, the electrons form very weak chemical bonds. You wouldn't be able to, say, chemically dissociate water atoms because the energy involved is too large. But there's an awful lot that you can accomplish with weak chemistry—our bodies make use of very weak chemical bonds for things like receptor binding.

How long before programmable matter gets into everyday use?
The path that we're on now is going to take a very long time because bringing quantum dots out into the real world, so that they operate at room temperature in a normal atmosphere, is not the focus of quantum dot research right now. But when I talk about programmable matter, the reaction tends to be very strong and favorable. I've gotten a lot of inquiries from the venture capital community and government agencies, who want to know what's involved in bringing this stuff to fruition.

I think that you could produce something like a wellstone fiber within 10 to 15 years. From there—if we look at the transistor, and the amount of time it took from the first transistors until the time when we were packing them onto microchips for 10 cents apiece—I think that within the lifetimes of most of your readers, we'll start to see applications in the real world.

How did you come to first use programmable matter in fiction—and then get involved in trying to direct the real field of quantum dot research toward material science applications?

The physicists involved in this research tend to be extremely conservative by nature. To a large extent, the people most closely involved in the field are the least willing to speculate about its implications, so you need an outsider to come in, and I think the researchers recognize that. The idea of programmable matter occurred to me because I'm a science fiction writer more than anything else. I was looking for an idea that was outrageous on the surface and yet plausible when you understood the details. I think it took the hubris of a science fiction writer to come up with the idea, and I think it took the general nature of an engineer to envision how it might actually work. But, of course, there is still a long way to go.