157 nm OPTICS: MEETING THE CRITICAL CHALLENGE OF CONTAMINATION

by James Doty, Ph.D., Alpine Research Optics, 3180 Sterling Circle, Boulder, CO 80301, 303-444-3420, FAX 303-444-1686, E-mail AROcorp@AROcorp.com, http://www.optics.org/arocorp/

Background

The demand for deep UV optics is growing rapidly, particularly at the 157 nm fluorine excimer laser wavelength. The most important application is obviously microlithography, where chip manufacturers need shorter wavelengths to produce higher density memory and processor chips. Indeed, although 193 nm ArF laser based systems represent the current state-of-the art in chip production, the 157 nm laser is now widely regarded as the inevitable source of choice in next generation microlithography steppers. There are also emerging materials processing applications at 157 nm that rely on the high photon energy to ablate "difficult" materials. For example, this is the only readily available laser wavelength that can be used to micromachine teflon.

Unfortunately, surface contamination presents unique problems at 157 nm, posing challenges for optics manufacturers and end users alike. The problem is that most chemical species demonstrate extremely high absorption at this wavelength. Indeed the term "vacuum UV" arose because of the need to remove air from the optical path in deep UV systems. In fact, absorption is so strong at 157 nm that even a monolayer of surface contamination (oil, water or even oxygen) can cause significant losses - up to 15% per surface. A microlithography system can contain up to 100 individual optical surfaces, so that a loss of only 2% per surface reduces total system transmission by 87%. In addition to this reduced throughput, surface absorption may also lower the lifetime of smaller optics that experience higher fluences. This increases costs to the end user in three ways: the cost of the replacement optic, the cost of downtime, and the potential costs caused by introducing more contamination when opening the system for optics replacement.

The end result is that both manufacturers and end users are now faced with eliminating surface contamination at the monolayer level. But fortunately, by following rigorous protocols and practices this problem can be effectively circumvented without pushing the effective cost of the optics to unacceptable levels.

Manufacturing Challenges

The main goals for manufacturers are to produce high quality (low transmission loss) beam delivery optics with high yields at market enabling costs. There are really three process stages from the manufacturers point of view - fabrication, cleaning and shipping. Contamination must be addressed at each stage.

The fluorite materials used as both substrates and in coatings are chosen for their low losses at 157 nm. Thus, transmission losses are primarily caused by contamination and/or scatter on the surface of the substrate and the outer surface of the coating. The former is the most critical, since there is no way to remove contamination trapped under a coating by any cleaning method.

Hydrocarbons, oxygen and moisture are the typical contaminants present when coating optics. At longer wavelengths, some manufacturers rely on storing the substrates under dry nitrogen. However, this is not sufficient at 157 nm, because, while dry nitrogen is moisture-free, it may contain hydrocarbons as well as particulates that could lead to surface scatter losses. The optics must be stored under ultra-pure nitrogen.

It does not take long exposure to the ambient atmosphere to produce a monolayer of surface contamination, so the substrates are always re-cleaned immediately prior to coating. At these levels, heating the substrate is not sufficient, and manufacturers such as Alpine Research Optics have developed proprietary, in situ cleaning protocols based on the reactive cleaning method described later in this text. The end result is a typical surface transmission of = 99% per coated surface. The goal is then to maintain this transmission even after shipping to the end user, storage on site, and final installation in the beam delivery system.

End User Perspective

In a microlithography system, the entire beam delivery path is a closed, clean environment. The system is constructed of materials that outgas very slowly; it is important to note that at this level of contamination, all materials outgas to some degree. To maintain clean optics, the system is continuously flushed with a fresh supply of ultra-pure inert gas. Thus, once the optics are in use, they are well-protected from contamination. However, shipping and storage are another matter.

The optics industry is already investigating improved shipping methods for 193 nm optics to avoid contamination during transit. But at 157 nm, the problem is so critical that these are unlikely to offer a complete solution. For instance, the most extreme approach for avoiding contamination during shipping would be to place the optics in a sealed, stainless steel container, filled with ultra-pure inert gas. The optics would remain unopened until installation. Clearly this is not a cost-effective approach, with shipping costing more than the optics themselves.

But fortunately, if the correct protocols are followed, small amounts of surface contamination can be completely removed from most 157 nm optics. It therefore makes more sense to ship the optics in conventional plastic packaging (see Figure 1), and to remove any contamination just prior to installation. Alternatively, some end users clean the optics immediately after receipt, and store them in a closed cabinet whose environment mimics the pristine microlithography system.

So what are the correct cleaning methods? The most effective protocol is a two-stage process. The first stage is to perform a methanol wipe, using high quality lens tissue and nanograde ethanol. The second stage is to perform some type of reactive cleaning. This is carried out in a sealed container continuously flushed with ultrapure pure oxygen. The optic is irradiated with either a deep UV laser or light from a deep UV discharge lamp. The reactive combination of energetic photons and oxygen removes most types of surface contamination. (Although ozone probably plays some role in this cleaning, the mechanism is still not fully understood). The oxidized and vaporized contamination is then flushed away by the gas flow.

However, it is very important to note that reactive cleaning is not a panacea or one-step process, and must only be used after careful methanol cleaning. For example, any silicon containing oils or molecules can be removed by methanol. But if reactive cleaning is used on its own, the silicon materials will be transformed into hard deposits of silicon dioxide which absorb at and scatter 157 nm radiation, and which are impossible to remove without damaging the surface.

The effectiveness of this two-stage cleaning method is clearly illustrated in Figure 2. This shows spectrophotometer plots of the deep UV transmission for the same optic after two types of cleaning: methanol wipe only, and methanol wipe followed by reactive cleaning.

Although this two-stage method will remove virtually all types of surface contamination, it should also be used in conjunction with a rigorous program of contamination minimization. An ounce of prevention is worth a pound of cure.

Conclusion

The extremely high absorption of most materials in the deep UV presents many practical barriers to cost effectively utilizing these wavelengths. However, the tremendous benefits offered by deep UV processing, in terms of producing smaller devices and features, has spurred tremendous efforts aimed at overcoming these limitations. The techniques already developed for fabricating and handling 157 nm optics clearly illustrates the progress that has already been made.