|
May 1999 |
Applications Challenge |
||||||
|
|
|
Producing optics that can withstand prolonged exposure to intense UV radiation has presented a number of challenges to manufacturers. Over the past few years, intensive research and testing have yielded fabrication methods enabling a significant increase in the lifetime of excimer laser components. However, the different output characteristics of excimer and solid-state lasers result in somewhat different damage mechanisms. It is not yet clear if the techniques developed for excimer laser optics will be applicable in the solid-state regime, because of their different damage mechanisms. This article will review what has been learned about long-life excimer laser components and discuss the probable direction of future work on solid-state UV laser optics. Excimer lasers are pulsed sources that produce light at the UV wavelengths of 351, 308, 248, 193, and 157 nm. The most powerful excimer lasers produce up to about 200W of total power, with repetition rates in the 100-Hz to 1-kHz range, and pulse lengths of tens of nanoseconds. The output beam from an excimer laser is rectangular, with a smallest dimension of at least 10 mm. The intensity profile is fairly homogeneous, and free of hot spots (localized areas of high intensity). Because its wavefront quality is far from diffraction-limited, an entire excimer laser beam cannot be easily concentrated into a small spot. This fact keeps the peak fluences normally encountered in a subsequent optical train fairly low, generally in the 0.1- to 5-J/cm2 range. Because of these moderate fluences, the failure mechanism in optical components used with excimer lasers is not usually catastrophic surface damage. In fact, the actual mechanism of damage is often impossible to clearly identify. A gradual decline in performance (e.g., reflectivity for a mirror) is observed as the total accumulated pulse count increases. The component typically ceases to perform adequately before catastrophic failure occurs. With no single damage event, it has become standard in both the microlithography and PRK industries to define a 5% decrease in reflectivity as failure for an excimer laser mirror. At the present time, it is possible to produce excimer optics that can withstand well over one billion pulses before failing.
Damage can also occur when residual polishing compound left on the part from fabrication absorbs UV laser light and heats up. But, it is very difficult to entirely remove all traces of polishing compound from a surface. This material adheres to the part's surface and can also work its way into microfractures and surface defects. In these latter instances, it becomes virtually impossible to remove through cleaning. Alpine Research Optics (ARO) has been producing high-damage-threshold UV laser optics for eight years. In that time, we have found no single magic bullet that eliminates laser damage. Rather, incremental lifetime improvements have come by combining a number of individual techniques. For example, our approach to minimizing SSD is to use a sequence of successively finer grinding and polishing steps. Each step removes sufficient material to eliminate any damage caused by the previous step. Another important change has been to utilize polishing compounds other than the commonly used ceria (CeO2). Ceria absorbs strongly at short wavelengths, and any residual ceria on a part can quickly lead to damage. The only two materials that make suitable substrates for the fabrication of transmissive optical elements in the deep UV are fused silica, which transmits down to about 190 nm, and calcium fluoride (CaF2), which extends down to around 130 nm. One of the most basic lifetime-limiting factors with both of these substances is bulk material absorption. Absorption leads to heating, which can directly damage a component. Absorption is a much more significant problem for fused silica than CaF2, because the latter has both a significantly higher coefficient of thermal conductivity as well as a lower thermal coefficient of refractive index. The result is that CaF2 heats up more slowly than fused silica, and experiences a smaller optical change when it does heat up. Another bulk material problem that occurs with extended exposure to deep UV light, especially at wavelengths of 193 nm and below, is color-center formation. Color centers are lattice defects that exhibit increased absorption. For fused silica, prolonged exposure to intense deep-UV light can also produce a phenomenon known as densification. This is an actual physical shrinkage, accompanied by an increase in the index of refraction, in the region of exposure. While densification does not necessarily damage a component, it can significantly change its optical properties, thus rendering it useless in a system. Unfortunately, the processing of calcium fluoride presents several unique challenges. CaF2 is anisotropic, hygroscopic, and very prone to chipping and fracturing. During polishing, small particles can break loose from the substrate edge and be dragged over the surface, producing scratches. Eliminating these scratches requires lengthening the polishing time. Unfortunately, the longer the polishing time, the more difficult it is to hold a given flatness or surface figure. Thus, fabricating CaF2 components requires a very delicate balancing act between the various process parameters (polishing time, spindle speed, spindle pressure, etc.) in order to simultaneously meet both shape and surface quality specifications. In this connection, we have found that the use of pad polishing, as opposed to conventional pitch laps, can speed up the process, making it easier to achieve tight surface quality specifications. Probably the most important step we have taken, however, is simply to environmentally isolate the polishing area for CaF2, completely preventing any cross-contamination from our fused silica polishing procedures. Considerable development has also been achieved on the optical coatings, especially for use at 193 nm. The main problem at this wavelength is the limited number of adequately transmissive materials. We have found that various fluoride materials produce thin films with the best combination of optical and mechanical characteristics. We have also determined that conventional E-beam evaporation is the most effective deposition method for these materials. Ion-assisted deposition, which can create high-density, high-damage-threshold coatings at visible and IR wavelengths, sometimes produces oxides that are absorptive in the UV. Solid-state UV lasers typically function by frequency-multiplying the fundamental wavelength of a laser crystal, usually Nd:YAG or Nd:YVO4, which both have a fundamental of 1.064 mm. This leads to multiplied output at the third (355 nm), fourth (266 nm), and fifth (213 nm) harmonics. Both CW and pulsed operation are possible, with pulse repetition rates usually in the 1-kHz to 10-kHz range. The total UV pulsed power available from solid-state lasers is much smaller than from excimer lasers, with a typical maximum of about 5W. However, they can have pulse lengths of just a few nanoseconds, so that the peak power can be quite high. Furthermore, solid-state lasers tend to have inhomogeneous beams, exhibiting hot spots of very high fluence. Finally, solid-state lasers can produce excellent mode quality (TEM00), meaning that the beams can be easily concentrated into small focused spots. As a result, solid-state laser optics can routinely expect to experience peak fluences in the 10 to 50 mJ/cm2 range, a value much higher than for excimer laser optics. Also, the higher repetition rates of solid-state lasers mean that their optics can quickly accumulate a very large total pulse count. Figures 2 and 3 show differences in damage mechanisms:
These unique characteristics lead to different damage mechanisms from those experienced with excimer lasers. The high peak fluences can literally blast a small hole in an optical coating or surface. For a typical beam diameter of about 1 mm, damage sites of 100 to 200 mm in size generally occur. Several of these may accumulate before the optic becomes unusable. Under certain circumstances, optics for solid-state lasers may also exhibit a gradual performance deterioration. This happens when dust or another contaminant settles on the optics surface. A high-fluence pulse may ablate this contaminant, leaving a very small damaged area in its wake. This damaged area, which is usually in the 10-mm size range, then scatters incident light, thus reducing the efficiency of the optic. A buildup of such damage sites over time will cause an appreciable reduction in component performance. Extensive testing is now being performed by organizations such as Lawrence Livermore National Lab (as part of the National Ignition Facility program) to determine the effect of various process techniques on UV solid-state laser damage resistance. One promising approach appears to be the use of superpolishing methods, where the optic is typically fully immersed in the polishing slurry (a mixture of water and abrasives). As the polishing progresses, the slurry is gradually thinned until it becomes essentially just water. Superpolishing results in optics with very low surface roughness, which may raise the peak fluence levels that the part can withstand. In recent years, tremendous progress has been made in extending the lifetime for excimer laser optical components. Now, the emergence of solid-state UV lasers, with much higher peak powers and repetition rates, has raised the performance bar once again. Through cooperation among sophisticated testing facilities, laser and component manufacturers, and systems integrators, ARO confidently expects to meet the need for long-lifetime optical components for these lasers.
For more information, please contact David Collier or Wayne Pantley at Alpine Research Optics, 3180 Sterling Circle, Boulder, CO 80301; (303) 444-3420, Fax (303) 444-1686, E-mail: AROcorp@AROcorp.com
|
