LIGA: early 1980's
LIGA is a process that was developed in the early 1980's by W. Ehrfeld at the West German IMT in Karlsruhe. It is an acronym standing for the stages [informative image] in the overall process: LIthogafie Galvanoformung Abformung. LIGA was one of the first major techniques to allow high aspect ratio structures (in other words, very skinny and tall) with lateral dimensions below one micron in size (~100's of nm) to be manufactured on-demand in a research environment (you still need a synchrotron radiation source, however).
But in a loose sense, LIGA is a deep X-ray + electrochemistry version of the late 18th century process of lithography (writing in stone), in which a thin coating of wax, grease, or ink on stone or metal is scribed with a pick, and the unprotected areas are etched with acid. When the wax is removed by heating, you are left with a patterned piece of material, suitable as a work of art, or as a replica master stamp that can be used to print almost unlimited negative duplicates of the original hand-sketched pattern. While photolithography has been used extensively in integrated circuit manufacture for almost 50 years, nanostructured stamp lithography techniques to pattern silicon and other device-relevant materials have recently been improved enough to offer a realistic alternative to purely radiation-based patterning and formation methods.
Whereas LIGA greatly assisted research in micromachining and MEMS, leading elements of nanotechnology will likely take advantage of the various new forms of nano-contact printing.
Soft Lithography: mid-1990's
Soft Lithography is a process developed at Harvard by researchers in the laboratories of chemist George M. Whitesides. Some of the most immediately useful applications of the micro-contact printing process [informative images therein] have been in microfluidics and NEMS (nano version of MEMS). Instead of having a resist layer that is exposed by radiation, physical abrasion, or molecular impact, a few-nanometer thick self-assembled monolayer is the "ink" that is stamped onto the patterned surface before etching. The height dimension of the molecular resist layer allows for moderately high aspect ratio structures to be made down to the 20-25 nm range (and lower in certain circumstances), limited to restrictions in patterning accuracy by chemical etching effects and the magnetic optical elements used in electron-beam lithography (which must be used at least once to make the nano stamps in the first place).
The stamps are used the same way a potato stamp is used if you ever made one in kindergarten (just press them onto something), and the molecular impressions they leave on surfaces can be used to seed crystal growth, bind strands of DNA for bio-analysis, or protect part of a surface from an etchant.
Another way of patterning the surface with monolayer ink that doesn't require the nano stamps is to use a tiny stylus to write it onto a surface serially. Dip-pen nanolithography is an example of this motif [many informative images within], where the tip of an atomic force microscope on gold acts somewhat like an inkwell pen on paper. The areas that have not been inked can be etched to leave terraces, mesas, and embossed structures that can later be used as stamps for printing.
Good beginners' references for Soft Lithography processing are: Xia Y. and Whitesides G. M., "Soft Lithography," Angew. Chem. Int. Ed. 1998, 37, 550-575 and Jackman R. J. and Whitesides, G. M., Chemtech 1999, 5, 18.
The recent report in Nature of silicon-quartz nanoimprinting was so important that Slashdot reported it twice. Chou and co-workers at Princeton [somewhat misleading image therein] are using a process called LADI (Laser-Assisted Direct Imprint) where a quartz "stamp" is made first by e-beam lithography, photolithography or soft lithography. That quartz structure is pressed directly against a flat silicon wafer, and then, due to the optical transparency of quartz, a high energy light pulse (20 ns, 308 nm ultraviolet XeCl laser, 1.6 mJ cm-2) melts the first few nanometers of the silicon. The molten silicon flows into voids in the quartz stamp and then cools, all in less than 250 nanoseconds. The simplicity and utility of the LADI scheme [informative image therein] is extremely impressive. There are no "inks", resists, or etchants to speak of, for this step (this would only be one step in an overall process), and the fidelity of the pattern transfer over large areas is very encouraging. The quartz stamp can also be used many times without apparent degradation.
If you have online access to Nature, the article is here, otherwise the reference is: Chou, S. Y., Keimel, C. and Gu, J. Nature 2002, 417, 835 - 837.
The principal investigator is reported by the BBC as saying:
But now Professor Stephen Chou of Princeton University says he has a way of stamping out chips with a die which could keep Moore's Law in operation for decades and maybe even beat it.
This statement is somewhat misleading in that LADI as a front-end processing technology for silicon is a divergent technology with respect to the ITRS roadmap, so the viability of its integration into production is unproven. The ability to grow a high-quality field oxide layer over the LADI structures, so that they may be insulated from the many layers of interconnect wiring directly above them, must be demonstrated. And while this technique has shown excellent structure-forming capabilities, the demands of solid-state electronics at the nanometer scale are sensitive to small perturbations in crystalline and even atomic order, so it remains to be seen what effects the rapid melting-cooling process will have on a doped transistor channel. Finally, the semiconductor industry has many other technological challenges facing it besides the ability to make small structures reliably and economically, so this advance alone cannot make up for the shortcomings in our current knowledge.
Much of the speculation that follows big discoveries in science, especially in this day and age of quasi-realistic corporate forecasting, is colored by a degree of optimism that is commonly encountered in the parents of newborns (she'll grow up to be Chief Justice of the Supreme Court!). But there is a good reason to expect that there will be technological advances in the near future that would seem totally alien to us now. The possible end of the ITRS roadmap brings with it the exciting prospect of uncertainty, and interest in new and unorthodox ways of approaching existing engineering problems has never been greater.