Lifetime testing of optical materials is of particular importance for UV laser applications. One straightforward approach to lifetime testing involves irradiating the material with a laser, then measuring the change in transmission with a spectrometer. Unfortunately, this method does not always give a correct result. Some optical degradation is permanent; other damage is temporary. If, for example, we study the transmission of an irradiated sample of fused silica during UV exposure, the induced absorption (centered at 215 nm) is clearly visible (see figure 1). After the laser is switched off, however, the defect centers that cause this absorption relax into a nonabsorbing state with a time constant of about 120 s. A spectrometer would incorrectly judge this type of material as radiation hard.
Figure 1. The absorption centers in fused silica show transient behavior.
The better method is to measure the decrease of power transmitted through a sample during irradiation. For UV-grade fused silica, this would reveal a linear decrease of laser transmission over time. Based on the fluence and the required transmission, we can calculate the effective lifetime of an optical element.
In general, optical systems cover a wide range of fluence levels, so this assessment has to be done for all elements and operating conditions. It is nontrivial. A misjudgment in the fluence level by a factor of two, for example, will introduce at least a 4X error in lifetime estimation as a result of the nonlinear behavior of induced absorption caused by a two-photon process.
For measurements performed over a wide range of fluence levels (varying energy density and number of pulses), a phenomenological model allows the extrapolation of the results from a short time test to a much greater number of pulses:
αinduced = a X εb X P
where ε is the energy density, P is the number of pulses, and a and b are material parameters.
Figure 2. Different types of fused silica show saturation or sudden absorption transition.
But even this model cannot be applied to all irradiation conditions. If we plot the increase of induced absorption (transmission loss) against the number of pulses (time), we see that the generation of defect centers results in a decrease of molecular solved hydrogen in the glass matrix (see figure 2). Surprisingly, one type of material shows a saturation behavior; the other, a sudden increase of induced absorption. Depending on the system design, this variation could result in the premature failure or unexpectedly long-term survival of an optical system. Therefore, a physical understanding of the type of material used is required. If this is not possible, some kind of universal dose should be applied, but this necessitates understanding the dependency of fluence on the laser-induced damage.
The semiconductor industry has focused a large amount of R&D on the long-term stability of fused silica and calcium fluoride for UV applications. The main suppliers now understand the damage behavior of their materials and how it can be controlled during the production process. The damage properties of these "excimer-grade" materials are documented quite well.
As a user, keep in mind that if you leave the well-known area of laser wavelength, pulse width, fluence, or lifetime, you may find unpleasant surprises. For example, consider microlithography and high-energy UV lasers. There is an excimer-grade material that fails to perform under the extreme fluence experienced by high-energy laser optics but performs well for microlithography. Another material will pass for a high-energy laser optic but fails to meet the requirements of microlithography due to an effect only present at very low energy densities. If component lifetime is critical to your application, be sure to perform laser damage testing under conditions close to the operating conditions. oe
Bodo Kühn is a physicist in the R&D department of Heraeus Quarzglas GmbH, Hanau, Germany.