Optical projection lithography--enabling nanolithography: photolithography has played the central role in microlithography for decades. Now it is the choice for manufacturing integrated circuits at the nanometre regime.
Invented by Alois Senefelder in Bohemia in 1798, lithography has quickly become one of the most popular printing techniques. Today's photolithography uses the same basic principle and has become the main technology used in computer chips and printed circuits manufacturing.
The silicon chip industry has followed the technology roadmap that was initially published in 1965 by a physical chemist, Gordon Moore. Moore predicted that by 1975, a computer chip would contain 65,000 circuits. Computer chips with a resolution of 90 nm are fabricated today using excimer lasers with wavelengths at 193 nm.
There are quite a few technologies available for integrated circuits fabrication. Electron beam lithography (EBL), extreme ultraviolet lithography (EUVL), ion beam lithography (IBL), X-ray lithography (XRL), and nanoimprint lithography (NIL) have been developed during the years, but they did not make it to mass production. By far, the most commonly used technology is optical projection lithography. All of the above technologies use the top/bottom approach. However, a new technique has gained attention in the past few years. Self-assembly of molecules following certain patterns makes use of the bottom/up approach and the size of the features can go to as low as a few nanometres.
The materials used in optical projection lithography consist of a resist and the substrate, usually silicon wafer. The resist is a complex mixture of polymers and photoactive components. The photolithographic process involves a few steps, depicted in the diagram from Figure 1.
[FIGURE 1 OMITTED]
In the first step, a pattern from a mask is transferred through a complex projection optical system to the photoresist coated on top of the silicon wafer. Photochemistry that occurs in the exposed regions will induce a difference in solubility between the exposed and unexposed regions. This may be only one step in the manufacturing process, but it is the most important one, and the amount of time and money invested in new materials with better performances is impressive. If the exposed regions become soluble, they are washed away during the developing step, in which case one obtains a positive image. If the exposed regions become insoluble, the ones washed away are the unexposed regions in which case one obtains a negative image. The next steps in the fabrication process are etching mad stripping.
Photoresists have evolved at the same time as the wavelengths used in photolithography. From novolac-based resists used at 436 and 365 nm to phenolic based resists at 248 nm, poly(methylmethacrylate) resists at 193 nm, and fluorinated polymers at 157 nm, the evolution had to take into account the absorbance of the polymer at the wavelength of irradiation, as well as its stability under the tough etching conditions. The source of radiation also evolved, from UV lamps to [F.sub.2] excimer lasers. The continuous race towards shorter wavelengths has a scientific explanation, since the features size (R) is given by the formula below:
R = [k.sub.1] [lambda]/NA
where [k.sub.1] is a parameter in the range 0.4-1.0, [lambda] is the wavelength of radiation and NA is the numerical aperture. Smaller features size can be obtained either by decreasing [k.sub.1] and [lambda] or by increasing NA.
The technique currently used in optical projection lithography has a Canadian connection in that it makes use of a concept developed at the beginning of the 1980s by Ito, Frechet and Wilson, namely chemical amplification (CA). Frechet was at the time professor in the department of chemistry at the University of Ottawa and collaborated with researchers from IBM in the field of photolithography. The idea is simple and versatile. A photoacid generator (PAG) is introduced in the photoresist, during exposure to radiation the PAG is photolyzed and produces [H.sup.+] and other photoproducts. Upon heating, [H.sup.+] will catalyze the deprotection of a t-Boc or t-Bu protected carboxylic pendant group, releasing isobutene, C[O.sub.2] in the case of t-Boc groups, and recovering the proton, thus inducing a solubility of the polymer in the exposed regions (Scheme 1). The proton can then catalyze another deprotection and the chain will continue. Washing with a basic solvent will remove the material from the exposed regions of the photoresist, leaving alternating lines and spaces.
As the race towards smaller features continues, new obstacles and challenges have to be overcome. Economic factors dictate which technology is used. Since the equipment for computer chip manufacturing costs on the order of hundreds of millions of dollars, the industry shows an understandable inertia when it comes to making radical changes. Changing the wavelength in optical lithography is easier than adopting a completely new technique. The pressure moves then on to the chemists who have to design and synthesize new polymers that are transparent at the wavelength used, that can withstand tough etching conditions, stick to the substrate, and show a clear discrimination in terms of solubility between the exposed and unexposed regions. There are also difficulties presented by the optical components. They must be adapted as well, and in the case of 157 nm photolithography, this was the main obstacle that led the industry to hunt for another solution for the 60 nm and 45 nm nodes. At 157 nm, the absorbance of the materials designed so far has been too high and the exposure has to be done either under vacuum or under a nitrogen atmosphere, increasing the cost of the final product. Fluorinated copolymers emerged as one solution for photoresists at 157 nm, but after Intel announced in 2003 that it stopped the 157 nm project, the attention focused mainly on 193 nm. The feature size obtainable with 157 nm is 45 nm, only 25 percent less than with immersion 193 nm photolithography--and it just does not justify the costs.
Immersion lithography at 193 nm appears to be the next technique to be used for integrated circuits with 60 nm spatial resolution. Features as small as 30 nm have been obtained in a laboratory with optical lithography, however, there is a long way from the lab bench to industrial scale. The competition is open for all the lithographic techniques after the 193 nm generation, and at this time no one has gained a strong advantage over the others. By 2010, a new technology should replace 193 nm immersion lithography and at this moment it is not clear which one will win the race. New materials with exciting properties at the nanometre scale are being developed and may replace the classic polymer-based photoresists. Building structures from the bottom up appears to be a solution, polymers deposited through spin coating will be replaced by self-assembling structures with sizes on the order of a few nanometres. Molecular circuits will probably be the ultimate goal, but before that there is still a long road ahead with exciting and challenging obstacles.
Marius Ivan, MCIC, is a PhD candidate in the department of chemistry at the University of Ottawa, under the supervision of J. C. (Tito) Scaiano, FCIC. His research is in the field of photochemistry.
|Printer friendly Cite/link Email Feedback|
|Comment:||Optical projection lithography--enabling nanolithography: photolithography has played the central role in microlithography for decades.|
|Publication:||Canadian Chemical News|
|Date:||Oct 1, 2005|
|Previous Article:||Little goes a long way: block copolymers on their way to nanodevices.|
|Next Article:||Geared up for nano: people, ideas, and infrastructure--the essentials of a world-class environment for nanoscience and nanotechnology research at...|
|Principles of Lithography, 2d ed.|
|Handbook of photomask manufacturing technology.|
|Microlithography; science and technology, 2d ed.|