Enhanced optical damage resistance of fused silica surfaces using UV laser conditioning and CO2 laser treatment

Transcription

1 Enhanced optical damage resistance of fused silica surfaces using UV laser conditioning and CO2 laser treatment Laurent Lamaignère1, Hervé Bercegol', P. Bouchut2, Annelise During1, Jérôme Néauport',Hervé Piombini3, and Gerard Raze1 1 CEA-Cesta, BP n 2, Le Barp, France 2 CEA-LETI, 17 rue des Martyrs, Grenoble, France 3 CEA-Le Ripault, BP n 16, Monts, France ABSTRACT For high power laser applications like the << Laser Megajoule >> facility under construction in France, laser-induced damage threshold (LIDT) in fused silica is a limitation. CEA has made efforts to improve LIDT at the wavelength of 35 1nm. Polishing and post polishing processes have been optimized. Laser damage sites density was decreased by several orders of magnitude by combining different fabrication steps. In order to further enhance optical laser resistance and to remove damaged sites on fullsize optics, several small-beam raster scanning techniques have been studied and developed to condition fused silica optics. To stop the growth of damage sites, a continuous CO2 laser was used to re-melt them. Laser induced damage tests, performed on instrumented and automated facilities, are reported in order to check and illustrate the effectiveness of these treatments. Damage initiation studies as well as damage growth measurements are presented. Keywords : fused silica, laser damage, polishing process, laser conditioning, laser mitigation 1. INTRODUCTION The operation of high power lasers for fusion research is mainly limited by the destructive interaction between the beam and the optical components in the high energy part of the beams. Megajoule laser (LMJ) facility, under construction at CEA- Cesta (France) will focus about 2 megajoules of energy at the wavelength of 35 1 nm in the center of an experiment chamber. The final optic assembly of these systems is made of large fused silica optics used in transmission. These optics are designed to sustain high energy at 3a of 10 to 16 J/cm2 for 3 ns pulse length. CEA had made several efforts for understanding mechanisms at the origin of laser damage, and then for manufacturing of 3ü) optical components with not much damage sites. The aim of this paper is to present several procedures used in order to enhance the optical laser damage resistance of fused silica surfaces by means oflaser treatments. After a synthetic view of the correlation between polishing phenomena and laser induced damage and the efforts realized in order to reduce the defects number, we will present in the second chapter the studies concerning laser conditioning. In the third part, we will focus on laser mitigation of damage sites created during conditioning. And then, we will recap the several steps used to enhance the optical laser damage resistance of fused silica surfaces. 952 High-Power Laser Ablation V, edited by Claude R. Phipps, Proceedings of SPIE Vol (SPIE, Bellingham, WA, 2004) X/04/$15 doi: /

2 2. interface OF POLISHED GLASS AND LASER 1TDUCED DAMAGE THRESHOLD 2.1 Polishing layer In order to have a good understanding of the physical mechanism of the act of polishing fused silica, we report the reader to the reference [11 by Néauport and coll. A typical polished glass interface was admitted by the scientific community. This interface geometry is depicted on figure 1. polishing layer SSD Bulk optical glass Figure 1: Interface ofa polished glass Subsurface damage (SSD) is probably introduced by the cutting of the material blank and by the fixed abrasive grindings steps. After grinding, the polishing "covers" the SSD with a polishing layer created by a mechano-chemical effect. Bibliographic data report a polishing layer thickness on silica of about 10 to 100 nm and a SSD depth of 1 to 10 or 100 m 2.2 Improvements of polishing and LIDT SSD is supposed to be constituted of deep cracks which could enhance light intensity by several factors of amplitude. Then, SSD is likely to be a major contributor for laser induced damage. Polishing layer contains contaminants coming from polishing compounds and materials used which can induced micro-pits damage type. Then many efforts were made in order to reduce SSD with post processing manufacturing steps and to suppress the polishing layer again with some post processing. To illustrate this work, figure 2 shows that since 1999 about two decades were gained in term of damage density when improving material removal and working on the abrasives and polishing conditions11. The several samples were tested at 3o with the raster scan method on Lutin facility at CEA Cesta (pulse duration of about 2.5 ns) [3] Duringthis test, one square centimeter of the sample is irradiated at fixed fluence. One shot is made by site. The number of damages created is then measured to give access to the damage density at the tested fluence. Then, we can compare firstly the samples each others and secondly, we have an absolute value of the damage density for each component. We observe in this graph the important diminution of the damage density. In order to further enhance optical laser resistance, several small-beam raster scanning techniques have been studied and developed to condition fused silica optics. These methods are described in the next section. Proc. of SPIE Vol

3 R1G < RwO1 i ii )< 5 : U Ruence (J/c&) Figure 2: Damage density as function as fluence (3a, 2.5 ns) for several process since 1 999[h] 3. LASER CONDITIONiNG OF FUSED SILICA SURFACES 3.1 Excimerlaser conditioning : Procedure and parameters Figure 3 shows the excimer laser conditioning system developed at CEA-Le Ripault. The laser source is a commercially high average power (25 W) Lambda Physik excimer (XeF, 351 nm), which can operate at up to 100 Hz with a 16-ns pulse-width. The main feature of the bench is the use of a micro-lens set which permit to shape and to focus the beam in order to achieve high laser fluence (up to 50 J/cm2) with a particular energy distribution in the beam. The spatial beam profile at the sample surface (figure 4, left)) exhibits a flat top profile along the long axis (y) and a quasi-gaussian profile along the short axis (x). The high output power and the flat top spatial beam profile allow to reduce the scan time in comparison with a gaussian beam. For conditioning treatments, the sample is translated continuously along the x-axis and stepped along the y-axis (figure 4, right). After each scan, the fluence is increased step by step up to 45 J/cm2. The beam presents low fluence wings along the x-axis : they can contribute to the conditioning since a line along y-axis sees a pre-scamiing by the wing before the scanning by the peak fluence. For all conditioning treatments, the final scanning speed (346 m/s) is set in such a way that the top 95%ofthe gaussian beam in the x-axis overlaps in 100 laser shots. 954 Proc. of SPIE Vol. 5448

4 Figure 3: The opticai layout of the excnner laser conditioning installation. The main features consist of a set ofmicro-lens in order to shape the beam and various diagnostics which pennit to observe and follow the result ofthe conditioning method used. The second feature of the bench is tle use of three photoelectric cells, coupled with a probe He-Ne laser. These diagnostics allow to "observe" the rear and front surfaces and the bulk of the component before, during and after conditioning. By these means, we can detect the occurrence of surface damage or bulk damage during the scan and we can observe damage growth during additional scans (figure 5, left). We can easily analyze the effect of each scanning method during or after the scan (figure 5, right) and therefore optimize the procedure. level LIvrTi1 ii torlzontal size (tm) WHi.. Vertical site (Rin).. i q( :r:::çi::: " t / ' J \ Figure 4: On the left, the beam profile at sample is reported, the long axis (y) exhibits a flat top beam profile (900 pm) while the short axis (x) is a quasi-gaussian (500 jim at lie?). On the right, the successive laser pulses overlap spatially to achieve uniform conditioning over the entire crystal. Proc. of SPIE Vol

5 &00E06 3 DOE 06 20J 25 J 200E06 30J 3: J 100E06 0.OOE+00 lu/h /.ffll.v( fr/ Detection and evolution of bulk damage Figure 5: On the left, occurrence and evolution of a bulk damage during scanning. On the right, observation by diffusion ofthe conditioning effect on the bulk ofeach method after scanning. 3.2 Nd:Yag laser conditioning : Procedure and parameters The laser source is a commercially Quantel Nd :Yag which can operate at 355 nm up to 20 Hz with a 6.5- ns pulse-width. The method used in this case is the same than previously consisting to a good overlapping of the beam shot after shot. In this case however, the beam is small and gaussian: the diameter is about 95 im at l/e2. The overlap ofthe beam is about 80 to 90%. 3.3 The damage test system The laser damage tests were realized at 355 mn. The damage test facility is a frequency tripled, Nd:YAG laser (Coherent Infinity localization at CEA-Cesta) operating at 10 Hz with a 2.5-ns, 3o pulse. It is capable of delivering up to 30 J/cm2 of 3o light in a im (at l/e2) spot to the sample (at the focus plan of a 5 meters focal lens) [3j In order to observe the occurrence of a damage during the test, a laser probe collinear with the pump beam and a photodetector detects changes in the scatter intensity from the localized damage. The onset of damage is detected as the abrupt increase in scatter signal. For a more precisely detection of the damage initiation, a direct observation with a CCD camera has been integrated to the set-up. The method used for determining the effectiveness of the laser conditioning is the "rasterscan" technique. This procedure is precisely described in reference [3]. In this case, the sample is continuously moved along the "x" axis in order to have a good overlapping of the beam, and step by step along the "y" axis (figure 6). One square centimeter of the sample is irradiated at fixed fluence. One shot is made by site. The number of damages created is then measured to give access to the damage density at the tested fluence 956 Proc. of SPIE Vol. 5448

6 Stc1) I )(.) Lm (X) pm Figure 6: On the left, sketch of the rasterscan procedure ; on the right, post-iiradiation optical observation of the irradiated area. 3.4 Results We present first, on figure 7, the results obtained on fused silica sample conditioned by the excimer laser source. One area is untreated and can be considered as "witness". The two others area have been conditioned up to 40 J/cm2: - for area 2, the sample wa displaced at 2162.tm/s, - forarea 3, the sample ws displaced at 346 pm/s. Finally, we can considered that the area 3 was seven times more irradiated during conditioning than area 2. First, we observe on figure 7 a decrease of damage density after conditioning, at fixed fluence. Secondly, area 3, which has been more conditioned, is the less damaged. The damage density is twice as less than the unconditioned area. The excimer conditioning permits to reduce the laser induced damage of fused silica and the effect is more important with first increasing fluence and secondly with several scans41. Damage density (/cm2) 100 IJcodtioned arca I i.: COUditOfld area ()) C ()nditioflcd arca ( Fluence (JIcm2) Figure 7: Rasterscan test results obtained at 355 nm 2.5 ns (Nd:Yag laser) on fused silica sample after excimer conditioning. The area (1) is unconditioned and the areas (2) and(3) have been conditioned respectively at 2162 and 346 jim/s 40 J/cm2. Proc. of SPIE Vol

7 On figure 8, we report results obtained on fused silica sample conditioned by the Nd:Yag (6 ns) laser source. One area is again untreated. The other has been treated with three pass at 6, 12 and 18 J/cm2. We observe once again a decrease of damage density after damage test. In this case, the Nd:Yag treatment permits to reduce the damage density by a factor 2 [4] Damage density (fern2) Fluence (JIcm2) Figure 8: Rastersca test results obtained at 355 am 2.5 ns (Nd:Yag laser) on fused silica sample after Nd:Yag conditioning 6 ns. The area (1) is unconditioned and the area (2) has been conditioned up to 16 J/cm2. 4. LASER MITIGATION During conditioning, damages are created in transmissive optics on rear surface. Several references indicate the exponential nature of growth: above a fluence threshold, every shot would increase the area of a preexisting damage by a growth factor51. This factor increases exponentially with fluence. For components made of fused silica, damage growth is considered today as the major limitation of their lifetimd61. To enhance this lifetime, one needs to avoid growth of damaged site. For this, we choose to locally refuse silica by means of continuous CO2 laser. This technique is now developed at CEA-LETI and we refer the reader for more details to reference [7]. 4.1 Experimental set-up The experimental bench developed by Bouchut and coll uses a continuous CO laser, 7W, at the wavelength of 10,6.un (this was chosen regarding to the absorbance of fused silica to this wavelength). The spatial mode of emission is a ThM gaussian beam. The laser beam is focused on silica component by means of a commercially Zinc Selenide lens. The beam waist diameter on silica is about 300 tm. The use of a shutter allows for the adjustment of the energy and pulse length of the pulses incident upon the optic. The damage is refused so that it lacks the memory form of its previous state and does not grow under subsequent laser shots. 958 Proc. of SPIE Vol. 5448

8 We have reported on figure 9 a typical optical microscopy observation of a mitigated damage. The laser beam locally melted and evaporated the silica surface, typically producing smooth, gaussian shaped pit (the pit depth depends on both laser power and time irradiation). i.::: jini St I division 20 pin Figure 9: Microscopic observation of a mitigated damage by means of CO2 laser. 4.2 Laser induced damage of mitigated sites To check the effectiveness of this process, we perfonned some experiments on the installation Lutin. About 70 damage sites (between 20 and 200 pm) has been mitigated. The sites then were irradiated first by 200 shots at 8 J/ciri2 and 200 shots more at 10 J/cm2 (2.5us, 355 mu) at 10 Hz. After the first irradiation at 8 JIcm2, 3% of sites grew again; all other sites survived with no growth in the site and no visual change to the site after irradiation. After the second pass at 10 JIcm2,we observed that 13% grew again but on peripheral surface (see figure 10). All of these latter sites were to large to be well mitigated (about 200 nn in diameter). These tests reliably establish the fact that CO2 laser treatment of initiated damage sites can thhibit the growth of those sites under further 355 nm irradiation81. Additional work is in progress at the CEA-LETI in order to achieve a 100% mitigation rate. E2 EE I sö -Ii:S1r =ii Figure 10: lvuic;oscopic observation ofa reinitiated damage after 200 shots at 10 JIcm2. Proc. of SPIE Vol

9 5. CONCLUSION Progress have been made on LIDT at 3co of fused silica components combining several steps during polishing process. To improve again LIDT, laser conditioning is a procedure which can reduce the damage density by a factor 2. In the case of fused silica, we observed a positive effect after excimer laser conditioning and after Nd-Yag (3n) laser conditioning. This final step reveal some damage sites which should be mitigated in order to avoid the damage growth shot after shot. The CO2 laser mitigation method seems to stop the damage growth and then to be a good method to obtain components without damages. BIBLIOGRAPHIE [1] J. Néauport, D. Valla, J. Duchesie, P. Bouchut, L. Lamaignère, J. Bigarré and N. Daurios, "Building high damage threshold surfaces at 351 urn", Optical and design engineering, SPIE vol. 5249, (2003). [2] L. Nevot and P. Croce, "Caractérisation des surfaces par réflexion rasante de rayons X. Application a l'étude du polissage de quelques verres de silicate", Revue Phys. App!. 15, (1980). [3] M. Loiseau, L. Lamaignère, C. Sudre, T. Donval, G. Raze, M. Josse, R. Courchinoux and H. Bercegol, "Automatic damage test benches: from small to large beams, from samples to large aperture components", Optical and design engineering, SPIE vol. 5252, (2003). [41 R.M. Brusasco, B.M. Penetrante, J.E. Peterson, S.M. Mancle, and J.A. Menepace, UV laser conditioning for reduction of 35 1 rim damage initiation in fused silica", Laser Induced Damage In Optical Materials 2001, SPIE vol. 4679, pp [5] G. Raze, J.M. Morchain, M. Loiseau, L. Lamaignère, M. Josse and H. Bercegol, "Parametric study of the growth of damage site on the rear surface of fused silica windows", Laser Induced Damage In Optical Materials 2002, SPIE vol. 4932, pp t61 H. Bercegol, P. Bouchut, L. Lamaignère, B. Le Garrec and G. Raze, "The impact of laser damage on the lifetime of optical components in fusion lasers", Laser Induced Damage In Optical Materials 2003, SPIE vol., pp. [7] P. Bouchut, L. Deirive, D. Decruppe and P. Garrec, "Local refusion of silica by a continuous CO2 laser for the mitigation of the laser damage growth", Optical and design engineering, SPIE vol. 5252, (2003). [8] R.M. Brusasco, B.M. Penetrante, J.A. Butler, and L.W. Hrubesh, "Localized CO2 laser treatment for mitigation of 351 mu damage growth in fused silica", Laser Induced Damage In Optical Materials 2001, SPIE vol. 4679, pp Proc. of SPIE Vol. 5448