杏吧原创

A new take on a centuries-old printing technique could shrink silicon chips and lead to advances in ultra-high-density computer storage.

Computer chips are made by a process called photolithography, in which intricate patterns are etched into silicon wafers at the nanoscale to mark the areas where the insulators, metal tracks and substrates that form the chips are to go. But as the size of electronic components shrinks, this technique is approaching its useful limit: it becomes too costly and difficult to go smaller.

Now 鈥榮 team at the University of Wisconsin-Madison thinks a new variation on the original lithographic technique could be the answer.

In Germany, late in the 18th century, devised a printing system in which a master image was created in oily ink on a block of limestone 鈥� whence the term 鈥渓ithography鈥�, after the Greek for 鈥渟tone writing鈥�. Using a combination of the oily inks and watery fixing solutions, which naturally repel each other, Senefelder could create a printing plate on which ink lines were sharply defined, making it possible to reproduce multiple exact copies.

Nanoprinting

Nealey鈥檚 team鈥檚 new technique 鈥� dubbed molecular transfer printing (MTP) 鈥� also involves the transfer of ink from a master to a replica, and they have shown that it can be used to duplicate one costly master silicon chip 20 times.

They took a master chip made using conventional photolithography and covered its surface with a class of compounds known as block copolymers. The copolymers grow in a crystal-like manner, using the slight surface relief caused by the photo-etching of the master chip to guide them. When two copolymers were used, they self-assembled to form crystal patterns that replicated those of the etched silicon chip beneath.

The polymer mix also included two conductive polymer inks 鈥� each attracted to just one of the copolymers 鈥� that pooled in the regions formed by their respective copolymer, mimicking the oil-and-water repulsion used in traditional lithography.

The team then heated the chip to 250聽掳C for a few hours to lock the pattern in place as a copolymer film before they pressed a second blank silicon chip onto the top. When the chips were heated to 160聽掳C for a further 24 hours, ink molecules diffused from the polymers and covalently bonded to the replica silicon chip鈥檚 surface to reproduce the pattern there. Finally, the copolymer film between the master and replica chip was dissolved away, leaving two identically patterned chips.

Fine detail

The technique is similar to another, called , in which the pattern is carved using a photolithographic procedure onto a master chip, which is then used to cast a soft resin stamp. The stamp is covered in ink and used to replicate the pattern. But the topographic features on the stamp are fragile, which means their size can鈥檛 be smaller than a few hundred nanometres, says Nealey 鈥� and that鈥檚 too big in the world of silicon chip design.

But such fragility isn鈥檛 a problem for the new technique, because the ink patterns are defined chemically rather than topographically, says Nealey. In fact the block copolymer method actually allows printing at a 鈥渟ignificantly higher resolution鈥� than the master copy, he says.

For example, the team created a master chip using a conventional etching technique to create a pattern of parallel lines 120 nanometres apart, but as the copolymers grew, the crystals they formed naturally switched from one copolymer to the other, to form lines separated by just 30 nanometres.

That means the technique could find applications in computer storage, where dense arrays of identically repeating punch-card-like patterns are required, claims Nealey.

Scale models

Using the technique, Nealey鈥檚 team has made features on a silicon wafer that have a half-pitch 鈥� half the distance between identical features 鈥� of just 15聽nanometres. The computer industry is currently trying to create commercial products with a half-pitch of 22聽nanometres.

鈥淲e will extend MTP to smaller-feature dimensions 鈥� less than 10聽nanometres 鈥� by using [different] polymers,鈥� says Nealey.

鈥淚t isn鈥檛 every day that a new idea for a lithographic method comes around, and this one does have potential to be applied commercially,鈥� says at the Massachusetts Institute of Technology.

It鈥檚 too early to tell how useful the new technique will be in practice, he says. 鈥淏ut there is a real need for new ideas in the field of nanofabrication, and this approach takes the field in an innovative new direction.鈥�

Journal reference: