Damage of optical components due to laser irradiation reduces reliability and limits durability. Calcium fluoride (CaF2) is commonly used for deep UV laser optics because it shows a very low tendency of color center formation as, compared to other UV-X optical materials. Here, we report on the exterior damage of CaF2 UV-X optics due to radiation with high pulse-energy densities (80 mJ/cm2) from an ArF laser. At such high energy densities, damage occurs on the external resonator side. The damage is generated by a partial alteration of the CaF2 substrate to crystalline CaCO3 (calcite). The decomposition of CaF2 is mainly driven by photochemical processes in the presence of water vapor, which are induced by the UV-laser light and the elevated temperature within the beam profile.
© 2009 Optical Society of America
The reliability and durability of optical instruments is limited by alteration or damage of the optical components and optics due to laser irradiation [1–7]. Thus, the influence of UV light on optical components is of great relevance for industrial applications such as lithography, metrology and medicine [8, 9]. In particular, deep ultra-violet (DUV) lasers damage calcium fluoride (CaF2) windows after relatively short exposure times . The damage appears as a white film on the exterior surface in the area of the beam profile. The composition and origin of this white material is so far unclear. A better understanding of the damage processes is needed to guide the development of improved UV optics with increased lifetime.
During the last few years, applications of deep ultra-violet (DUV) and vacuum ultra-violet (VUV) radiation have widely expanded. The current modes of laser operation in this optical range conform to the lithography sector, where the optical components are operating under well defined and closed working conditions at low energy densities [10, 11]. In contrast, the operating conditions in medical applications, photochemistry, micro-material machining, metrology, or life sciences are less well defined [5, 7, 12–14]. In these ‘nonlithography’ applications, the optics must sustain higher energy densities in poorly defined environments. Such demanding applications imply mechanical, thermal and chemical stress on the optical material, which result in lower product life times. In the worst case, the maintenance costs of optical components amount to 40 % of the total costs within tube durability. Thus, lifetime considerations for the optics must include endurance data of optic materials and thin films in ambient conditions.
Several experiments have been carried out to address this topic. Recent progress in the fabrication of plasma-assisted optical thin films and thick coatings for laser applications and their characteristics are summarized in [15, 16]. The importance of light scattering on multilayer coatings is highlighted in . The dose-rate alteration (densification) of fused silica caused by an Excimer laser (ArF, 193 nm) was investigated to understand the effects of light fluence, pulse count and material grade . The investigation of damage threshold for UV light and the behavior of dielectric thin films show dependence on substrate, thickness and layer material [19–21]. Light scattering on multilayer coatings is also very important . The choice of polishing grades affects the quality of coated components. A better polishing grade increases the UV-laser (193 nm and KrF, 248 nm) damage threshold of CaF2 [12, 22]. Thin magnesium fluoride (MgF2) films deposited on CaF2  by electron-beam evaporation provide a very high UV-laser damage threshold of 9 J/cm-1 (focused 248 nm laser). Uncoated CaF2 outcoupling optics showed damage of the external part of the laser resonator when exposed to energy densities of more than 50 mJ/cm2 on the transmitting area. A similar damage of the high reflective mirror, however, never has been reported. A degradation of fused silica and CaF2 with fluxes in the range from 0.2 to 4 mJ/cm2 was reported in . The chemical details of the damaging process remain unclear. In the following we show that the alteration of CaF2 into CaCO3 plays a major role in the damage mechanism.
2. Materials and methods
To determine the lifetime of an outcoupling mirror, we have performed marathon laser-irradiation tests for different optic substrates and thin film coatings. The samples were exposed to UV light from a water-cooled argon fluoride excimer laser (ArF, 193 nm, photon energy 6.42 eV, Coherent, Munich, Germany) with a solid-state pulse power switch, which has been tailored to purpose. The maximum laser pulse energy of 90 mJ was reached with the upper limit of the charging voltage of 1500 V with a maximal possible pulse frequency of 2 kHz and pulse duration of about 12 ns (FWHM). With a beam profile of 10 × 3.5 mm2, an energy density of more than 240 mJ/cm2 could be achieved. All measurements were carried out in an energy-stabilized mode at 1 kHz in continuous operation, where the high voltage was regulated to an output pulse energy of 28 mJ (80 mJ/cm2). The pulse energy was measured externally by a power meter (PM 150–50, Coherent, USA). A stream of filtered laser gas minimized the deposition of dust that was generated from electrode abrasion due to the high voltage discharge on the resonator internal side.
The outcoupling test optics (diameter 38.1 mm, 5 mm thickness) were loaded into the mirror mount for each experiment and sealed against the atmosphere by an O-ring. The optical mount was designed to allow for the access to the sample for direct temperature measurements on the mirror at distances of 1 mm, 3 mm, and 6 mm from the laser beam. The generation of oxygen radicals and ozone in front of the optic was minimized by flushing the external optical path with high-purity nitrogen. All components were made from stainless steel. Contamination was prevented by carefully handling components and optics in a clean environment. The dielectric reflection coatings (>98 % for the high reflective mirror and 25 % for the outcoupling mirror) were on the resonator’s internal side with contact to the fluoric laser gas. One vendor provided all the polished CaF2 substrates. The optical coatings on the substrates were fabricated by different manufacturers. Due to the variety of coating producers, the coatings differed in layer design and production.
Various state-of-the-art optics were tested until 250 million laser pulses were reached or damage occurred. In a second set of experiments, the measurements were repeated with improved optics, which were produced with improved manufacturing parameters for thin film deposition, layer design, and crystal polishing. These mirrors were stepwise tested until damage occurred. After each step, the outcoupler was inspected optically and the test was continued if there was no sign of alteration otherwise the damage was investigated.
We used a confocal Raman microscope alpha300 R (WITec, Ulm, Germany) for Raman measurements. The system used a standard 100× microscope objective (Air, NA = 0.90, working distance 0.26 mm, Nikon, Düsseldorf, Germany) for diffraction limited focusing and collecting the scattered light. A frequency doubled Nd:YAG (532 nm) laser was used for excitation. The scattered light was analysed with a lens based spectrometer, equipped with an 1800 lines/mm grating.
Energy dispersive X-ray spectroscopy (EDX), integrated in a transmission electron microscope (TEM/STEM JEM-3010, JEOL) was used to characterize elements on top and below the damaged mirror surface to clarify the chemical modification of the bulk material. The EDX measurements were supported by time-of-flight secondary ion mass spectrometry (ToF-SIMS; IONTOF TOF.SIMS5-300). Transmission Electron Microscopy (TEM CM20, Phillips) was used to characterize the structure of subsurface damaged regions. A focused ion beam (FIB, NVision, ZEISS) technique was used for TEM sample preparation and subsurface 3D volume reconstruction.
3. Results and discussion
3.1 Damage formation
Various UV transparent substrates were tested prior to the experiments: barium fluoride (BaF2), UV grade fused silica, lithium fluoride (LiF), samarium fluoride (SmF3), MgF2 and CaF2. All crystals showed color centers after a short exposure time, with the only exception of CaF2. Thus, we focused on CaF2 as a substrate material because it showed the lowest tendency of color center formation [15, 24, 25].
All tests showed similar temperature characteristics. Directly after starting the irradiation test, the temperature increased and stabilized within a few hours at about 70 °C (1 mm from laser beam boundary), 65 °C (3 mm) and 55 °C (6 mm). The laser tube temperature was in the range of 40 °C to 45 °C, depending on the discharge voltage for laser power control. The window temperature inside the beam profile could not be measured. It was estimated to be much higher than 150 °C. Shortly before damage occurred, the temperature on the exterior surface rapidly increased to 150 °C (1 mm) within some million pulses (ca. 1h).
For all samples, initial external surface damage occurred at the outer boundary of the transmitted laser beam profile of 3.5 × 10 mm2. Once damage occurred, the damaged area grew rapidly towards the center of the beam profile. The morphology and size of the damage was similar on all samples. There was only a slight variation in number of laser pulses until damage occurred. In the following, we will discuss the chemical alteration process of two typical samples: one uncoated and one with a protective MgF2 coating.
3.2 Raman analysis
The damaged area of a mirror without protection layer was characterized with the confocal Raman microscope. Figure 1(a) shows an optical micrograph of the damage area, whose surface size was 3.5 × 10 mm2. The damage appears as a white contamination on the surface. Figure 1(b) shows a detail of the damage boundary.
Spectra measured with a 10 s integration time on the damaged and undamaged material are shown in Fig. 2. Within the spectra, five intense sharp peaks are visible together with fluorescence in a range of 150–500 cm-1. One peak at 330 cm-1 was present in all spectra, the others (155 cm-1, 284 cm-1, 713 cm-1 and 1089 cm-1) were only found in the damaged areas.
These peaks correspond to CaF2 (330 cm-1) and calcite (crystalline CaCO3) [26, 27]. Table 1 lists the peak assignment and the corresponding mode symmetry and origin. There is a very good agreement of the measured peaks with the reference data. In the damaged areas, some fluorescence with minor peaks occurred in the spectra, which is typical of organic contaminants.
The images in Figs. 3(a) and 3(b) show the lateral distribution of calcium fluoride and crystalline calcium carbonate (calcite). The images consist of 160 spectra per line and 80 lines per image with a size of 18.6 × 8.6 μm2. Each spectrum was integrated for 1 s. The spectral images were filtered by summing up the counts in a selected interval around a Raman peak as indicated in Fig. 2 (top).
The lateral intensity distribution of CaF2 lattice mode at 330 cm-1 (filter window: 315 – 355 cm-1) shows a higher intensity in the undamaged area and a decreased intensity in the damaged part. Simultaneously, the intensity of the CO3 symmetric stretch peak (1089 cm-1, filter window 1080 – 1110 cm-1), is higher in the damaged area. The size of the CaCO3 (calcite) particles could be resolved to a few micrometers. The calcite crystals were surrounded by organic material (fluorescence at the lower wave number ranges) appearing as dark blue in Fig. 3(a) and bright blue in Fig. 3(b).
Vertical slices were measured for subsurface investigation. Figure 4(a) shows the optical image of a damaged area. The black line indicates the deep scan location were a slice was acquired perpendicular to the surface. The total scanning height was 8 μm starting at 3 μm above the surface with a length of 15 μm. 40 lines with 150 points were acquired; each spectrum was integrated for 1.5 s. Figure 4(b) shows the sum filtered CaF2 lattice mode image (330 cm-1) and Fig. 4(c) a map of the CO3 symmetric stretch mode (1089 cm-1). The image indicates that CaCO3 occurs in a depth 1 μm beneath the surface (considering the different indexes of refraction). Additionally, the CaCO3 has grown to a height about 0.5 μm above the surface.
Figure 5 shows a vertical profile of a mirror with an exterior protection layer of MgF2 (<100 nm). The optical image in Fig. 5(a) shows a part of the damaged area, which is covered with debris from the protection layer. The black line indicates the position of a vertical profile. The 25 μm long slice was taken from 3 μm above to 5 μm below the surface. The images consist of 200 points per line and 30 lines per image with an integration time of 1.5 s for each spectrum. A Raman map of the CaF2 lattice mode is shown in Fig. 5(b). Figure 5(c) shows a map of the MgF2 lattice mode; Fig. 5(d) shows the CO3 symmetric stretch and some artifacts. The white arrows indicate calcite crystals. It is evident, that the CaF2 and MgF2 signals are decreased on the altered part. This implies that the protective coating has been removed and that CaF2 was altered to CaCO3.
3.3 SEM /TEM/EDX, and ToF-SIMS analysis
Scanning electron microscope (SEM) images (Fig. 6(a)) of an exterior unprotected mirror show dendric pyramidal crystals grown on the surface. A surface damage with a height of about 1 μm and a depth of about 1 to 2 μm was observed (SEM cross-section, Fig. 6(b)). The crystals nucleated independently and grew in a 3-fold symmetric pyramids. All the trees have the same orientation of the shape, which should be defined by crystallographic orientation of the matrix. Electron diffraction and EDX data show that these pyramids are composed of CaF2. Cross-sectional TEM imaging (Fig. 6(c)) reveals cracks propagating 1–2 μm beneath the surface. The cracks occur although the matrix beneath is perfectly crystalline, as indicated by bending contours (dark curved lines). 100–500 nm sized crystals were observed on dark field images (not shown here), in SEM cross-sectional views as slightly darker areas around the cracks and on EDX maps (Fig. 6(c) insert) as inclusions in the matrix at the edges of cracks (marked as V at Fig. 6(c)). These inclusions either represent slightly rotated grains (as evidenced by electron diffraction) or an oxygen rich phase (as evidenced by EDX).
Using a sequence of images (Fig. 6(b)) a tomographic reconstruction of subsurface volume was made (Fig. 7). The reconstruction revealed a peculiar porous structure with pores having a shape of conical structures or hemi-spheres and the cracks developing from the tops of the conical structures.
ToF-SIMS analysis showed aliphatic and aromatic hydrocarbons, fatty acids, and nitrogen and oxygen-bearing hydrocarbons on the entire sample surface. In the inner part, nitrogen and oxygen-bearing hydrocarbons (in particular C3H8N, C18H40N, C20H44N) were found together with triglycerides. Finger prints can be barred as the origin for the hydrocarbons due to the usage of clean room gloves. A deep profile analysis of the damage area (300 × 300 μm2) showed higher intensities of Ca2OF, CaxOy and CaOH. An elevated CaCO3 intensity could not be observed. This was probably due to the large scanning area and small amount of CaCO3.
The damaged area of high flux, illuminated DUV outcoupling optics was investigated with different analytic methods. The damage started at the outer boundary of the transmitted laser beam and grew towards the center of the beam profile. The exterior side was roughened and small amounts of CaF2 were altered into calcite (crystalline CaCO3). Additionally, calcium oxide was detected, which was likely an intermediate. As a consequence, the transmitted laser power was decreased due to scattering losses and adsorption respectively.
A possible deterioration process is discussed in detail in . Briefly, the alteration of the optics can be attributed to the influence of the high-energy photons flux on defects located at the interface between substrate and coating.
The exchange reaction is supposed to follow:
This assumption is supported by the observation that the temperature on the boundary and in the center of the beam profile is very high (>150 °C) and the UV photon energy (6.42 eV). The very small amount of emerging HF is rapidly removed by the external flushing with nitrogen. Calcium oxide hydrates Ca(OH)2 were probably generated in an exothermic reaction due to presence of water. Minor amounts of water are present on surfaces in ambient conditions. Such a reaction could be partially responsible for the increase in temperature, which was observed shortly before the damage occurred. Together with increased light absorption, such a reaction expedites the damaging process.
The laser-induced damage on the exterior side of outcoupling CaF2 mirrors was investigated. The mirror was subject to a transmitted ArF UV-laser beam with an energy density of 80 mJ/cm2. The irradiation caused the surface to roughen and to partially alter CaF2 into CaCO3. The exchange was driven by temperature and the photo-induced chemistry due to the UV laser light. The alteration started at the edge of the beam profile where the highest temperature gradients occur. Crystalline CaCO3 (calcite) within the damaged area could be observed by confocal Raman spectroscopy. Further surface analysis with TEM/EDX confirmed this result. ToF-SIMS also revealed calcium oxide, which is very likely a precursor for the alteration to CaCO3. Calcium oxide cannot be detected by Raman spectroscopy and thus went undetected. The CaCO3 crystals were embedded in a variety of organic materials, which were identified using surface analytical techniques. High energetic photons flux induces photochemical reactions of imperfections located at the interface between substrate and coating or with contaminants from the environment.
The results indicate that a better protection of the surface is required in order to avoid the alteration of CaF2. Such a protection could be realized by additional buffer layers, improved optical materials, advanced deposition processes and for coating designs.
We gratefully acknowledge financial support by the German Federal Ministry of Education and Research (BMBF) for founding the cooperative project FLUX (FKZ: 13N8937). We thank VDI-TZ, division “Optische Technologie”, for supporting the subproject “Untersuchungen der Dauerbetriebs-festigkeit von beschichteten optischen Bauteilen im Resonator von Excimerlasern bei 193 nm”. We are grateful to Prof. Rettenmeyer (University Jena) for providing TEM/EDX - and to Dr. Horn (Zeiss, NTS) for FIB access. Furthermore we thank Guido Winkler and Elke Tallarek, tascon GmbH for ToF-SIMS analyses.
References and links
1. J. H. Apfel, E. A. Enemark, D. Milam, W. L. Smith, and M. J. Weber, “Effects of barrier layers and surface smoothness on 150-ps, 1.064 μm laser damage of AR coatings on glass,” in (NIST Spec. Pub. 5009, 1977), p. 255.
2. V. Liberman, M. Rothschild, J. H. C. Sedlacek, R. S. Uttaro, A. Grenville, A. K. Bates, and C. Van Peski, “Excimer-laser-induced degradation of fused silica and calcium fluoride for 193-nm lithographic applications,” Opt. Lett. 24, 58–60 (1999). [CrossRef]
4. T. Feigl, J. Heber, A. Gatto, and N. Kaiser, “Optics developments in the VUV - soft X-ray spectral region,” Nucl. Instrum. Methods Phys. Res. A 483, 351–356 (2002). [CrossRef]
5. W. Arens, D. Ristau, J. Ullmann, C. Zaczek, R. Thielsch, N. Kaiser, A. Duparre, O. Apel, K. Mann, R. H. Lauth, H. Bernitzki, J. Ebert, S. Schippel, and H. Heyer, “Properties of fluoride DUV excimer laser optics: influence of the number of dielectric layers,” Proc. SPIE 3902, 250–259 (2000). [CrossRef]
6. M. Marsi, A. Locatelli, M. Trovo, R. P. Walker, M. E. Couprie, D. Garzella, L. Nahon, D. Nutarelli, E. Renault, A. Gatto, N. Kaiser, L. Giannessi, S. Gunster, D. Ristau, M. W. Poole, and A. Taleb-Ibrahimi, “UV/VUV free electron laser oscillators and applications in materials science,” Surf. Rev. Lett. 9, 599–607 (2002). [CrossRef]
7. D. Ristau, W. Arens, S. Bosch, A. Duparre, E. Masetti, D. Jacob, G. Kiriakidis, F. Peiro, E. Quesnel, and A. V. Tikhonravov, “UV-optical and microstructural properties of MgF2-coatings deposited by IBS and PVD processes,” Proc. SPIE 3738, 436–445 (1999). [CrossRef]
8. C. R. Clar, M. J. Dejneka, R. L. Maier, and J. Wang, “Extended lifetime excimer laser optics,” US Patent 7242843 (2007).
9. R. J. Rafac, A. Lukashev, and W. Zhang Kevin, “Method and apparatus for stabilizing optical dielectric coatings,” US Patent 20050008789 (2005).
10. V. Liberman, S. T. Palmacci, D. Hardy, M. Rothschild, and A. Grenville, “Controlled Contamination Studies in 193-nm Immersion Lithography,” Proc. SPIE 5754, 148–153 (2005). [CrossRef]
11. V. Liberman, M. Switkes, M. Rothschild, S. T. Palmacci, J. H. C. Sedlacek, D. E. Hardy, and A. Grenville, “Long-Term 193-nm Laser Irradiation of Thin-Film-Coated CaF2 in the Presence of H2O,” Proc SPIE 5754, 646–654 (2005). [CrossRef]
12. A. Duparre, R. Thielsch, N. Kaiser, S. Jakobs, K. R. Mann, and E. Eva, “Surface finish and optical quality of CaF2 for UV lithography applications,” Proc SPIE 3334, 1048–1054 (1998). [CrossRef]
13. A. Gatto, J. Heber, N. Kaiser, D. Ristau, S. Gunster, J. Kohlhaas, M. Marsi, M. Trovo, and R. P. Walker, “High-performance UV/VUV optics for the Storage Ring FEL at ELETTRA,” Nucl. Instrum. Methods Phys. Res. A 483, 357–362 (2002). [CrossRef]
14. J. Ullmann, M. Mertin, H. Lauth, H. Bernitzki, K. Mann, R., D. Ristau, W. Arens, R. Thielsch, and N. Kaiser, “Coated optics for DUV excimer laser application,” Proc SPIE 3902, 514–527 (2000). [CrossRef]
15. L. Martinu and D. Poltras, “Plasma deposition of optical films and coatings: A review,” J. Vac. Sci. Technol. A 18, 2619–2645 (2000). [CrossRef]
16. L. Pawlowski, “Thick laser coatings: A review,” J. Therm. Spray Technol. 8, 279–295 (1999). [CrossRef]
18. V. Liberman, M. Rothschild, J. H. C. Sedlacek, R. S. Uttaro, and A. Grenville, “Excimer-laser-induced densification of fused silica: laser-fluence and material-grade effects on the scaling law,” J. Non-Cryst. Solids 244, 159–171 (1999). [CrossRef]
19. E. Welsch, K. Ettrich, H. Blaschke, P. Thomsen-Schmidt, D. Schafer, and N. Kaiser, “Investigation of the absorption induced damage in ultraviolet dielectric thin films,” Opt. Eng. 36, 504–514 (1997). [CrossRef]
20. H. Blaschke, W. Arens, D. Ristau, S. Martin, B. Li, and E. Welsch, “Thickness dependence of damage thresholds for 193-nm dielectric mirrors by predamage sensitive photothermal technique,” Proc. SPIE 3902, 242–249 (2000). [CrossRef]
21. K. Ettrich, H. Blaschke, E. Welsch, P. Thomsen-Schmidt, and D. Schaefer, “UV-laser investigation of dielectric thin films,” Proc. SPIE 2714, 426–439 (1996). [CrossRef]
22. H. Johansen and G. Kastner, “Surface quality and laser-damage behaviour of chemo-mechanically polished CaF2 single crystals characterized by scanning electron microscopy,” J. Mater. Sci. 33, 3839–3848 (1998). [CrossRef]
23. M. L. Protopapa, F. De Tomasi, M. R. Perrone, A. Piegari, E. Masetti, D. Ristau, E. Quesnel, and A. Duparre, “Laser damage studies on MgF2 thin films,” J. Vac. Sci. Technol. A 19, 681–688 (2001). [CrossRef]
24. J. Ferre-Borrull, A. Duparre, and E. Quesnel, “Roughness and light scattering of ion-beam-sputtered fluoride coatings for 193 nm,” Appl. Opt. 39, 5854–5864 (2000). [CrossRef]
25. L. Escoubas, A. Gatto, G. Albrand, P. Roche, and M. Commandre, “Solarization of glass substrates during thin-film deposition,” Appl. Opt. 37, 1883–1889 (1998). [CrossRef]
26. N. I. o. A. I. S. a. Technology, “Raman Spectra Database of Minerals and Inorganic Materials (RASMIN)” (National Institute of Advanced Industrial Science and Technology), retrieved http://riodb.ibase.aist.go.jp/rasmin/.
27. Q. Williams, Infrared, Raman and Optical Spectroscopy of Earth Materials, Mineral Physics and Crystallography - A Handbook of Physical Constants (American Geophysical Union Washington, D.C, 1995), Vol. AGU Reference Shelf 2, pp. 291–302.
28. N. Beermann, H. Blaschke, H. Ehlers, D. Ristau, D. Wulff-Molder, S. Jukresch, A. Matern, C. F. Strowitzki, A. Görtler, M. B. Gäbler, and N. Kaiser, “Long Term Tests of Resonator Optics in ArF Excimer Lasers,” in XVII International Symposium on Gas Flow and Chemical Lasers & High Power Lasers 2008, 2008),