We now look back with pity on old computer printers, with their glacial bizz-buzz and annoying perforated-edge paper. A decade from now people will surely look back in pity on the things we call printers today. Three-dimensional printers, capable of producing entire objects, are already coming down in price, and new types of printers can output electronic circuit boards or even entire functional circuits. Now researchers have a printer that outputs silicon chips.
The device, created by materials scientist Masahiro Furusawa of Seiko Epson Corporation in Japan and his colleagues, squirts out polysilane, a polymer of silicon and hydrogen. Once laid down, it can be baked at kitchen-oven-cleaning temperatures to drive out the hydrogen and leave behind crystalline silicon. The technique provides an alternative to the conventional process for producing microchips, which requires refining, stenciling and etching the silicon—a sequence that is both complicated and wasteful: 99 percent of the original silicon is thrown away. The silicon printer is still fussy and does not produce chips with as fine a level of detail. Nevertheless, it might lower the cost of low-resolution silicon devices such as display circuitry and solar cells.
Gizmo users of the future will also look back in pity on what flows through those circuits. Right now circuits convey signals by the presence or absence of electrons, ignoring the particles’ spin (the quantum analogue of rotation). The emerging technology of spin-based electronics, or “spintronics,” seeks to make use of this wasted information. So far its only application has been in hard disks. Researchers have also sought to exploit it in logic gates for computational processing, not least because it takes less energy to flip an electron’s spin than to change its direction of motion. The hang-up is that detecting and manipulating spin requires magnets, which are tricky to integrate into silicon circuits.
Physicist Hanan Dery of the University of California, San Diego, and his colleagues have now come up with a workable logic-gate design. A combination of fixed magnets and voltage levels steers electrons based on their spin and the desired gate function. The team outlined a proposal for a spintronics computer based on the gate, but it remains to be seen whether it will work in practice.