Multi-pulse laser-induced damage threshold (LIDT) was experimentally investigated up to ~106 pulses for Cu, Ag mirrors. The surface roughness and the hardness dependence on the LIDT were also examined. The LIDT of OFHC-Cu decreased with the pulse number and was 1.0 J/cm2 at 1.8 × 106 pulses. The expected LIDT of cutting Ag at 107 pulses was the highest; Ag mirror would be one of the best choices for ITER Thomson scattering system. For the roughness and hardness, material dependences of LIDT are discussed with experimental results.
© 2013 OSA
In future fusion experimental reactors including ITER, metallic mirrors will be used for various optical diagnostics: optical spectroscopy systems and laser aided spectroscopy such as an interferometer, a polarimeter, and Thomson scattering diagnostics . Concerning the mirrors for Thomson scattering (TS), they are required to have a high multi-pulse laser-induced damage threshold (LIDT), because they are irradiated with frequently repeated high-power laser pulses . Furthermore, it is necessary for the mirrors to have high durability for sputtering by charge-exchange neutrals and for neutron irradiation [3–5]. If the damages were formed on the surface, the laser transmission mirrors could not be used until they are replaced. Therefore, investigation of multi-pulse LIDT is a critical issue for the laser transmission mirrors . In addition the multi-pulse LIDT is not only important for laser diagnostics mirrors for ITER. For example, in inertial fusion reactors, many mirrors will be necessary to transmit laser pulses to a target. They are subjected to X-ray, burn ions, and laser pulses, and damage threshold of metallic mirror is also an important parameter for the mirrors .
The laser transmission mirrors have a role to transmit laser pulses to plasma. The candidate materials are metals with high reflectivity such as Cu and Ag [7,8]. In previous studies of multi-pulse LIDT for Cu and Ag mirrors, number of laser pulses was limited, and, moreover, the dependence of multi-pulse LIDT on the surface roughness and hardness have not been investigated. The investigated pulse number was up to ~105 pulses for Cu mirrors; it was pointed out that the quality of the mirror surface was not sufficient . In terms of Ag mirrors, experimental multi-pulse LIDT data are limited to 104 pulses and the surface condition of the sample was not defined . Thus, further systematic investigations are necessary to assess the lifetime of metallic mirrors.
In this study, multi-pulse LIDT was experimentally investigated up to ~106 pulses for Cu and ~105 pulses for Ag mirrors. The surface roughness and the surface hardness dependence of multi-pulse LIDT also were examined. It is necessary to maintain its optical properties for more than 107 laser pulses in ITER TS system . Based on experimental data, the laser damage threshold of greater than 107 pulses will be extrapolated.
The measurements for multi-pulse LIDT were conducted with the setup shown in Fig. 1 . The system consists of a pulsed Nd: YAG laser, an attenuator, a beam splitter, and a focus lens with a focal length of 307 mm at 1064 nm. The wavelength of the laser was 1064 nm, and the pulse duration and repetition rate were 5-7 ns (FWHM) and 10 Hz, respectively. The energy of a pulse was ~22 mJ. The special profile laser beam was measured with a beam profiler and had a spatially Gaussian intensity profile; the spot size at metal mirror was 7.0 × 10−3 cm2 (FWHM). The incident angle of the laser beam to metallic mirrors was 12-15 degree from the normal. Damage threshold does not depend on the angle because the polarization of the laser beam was p polarization . The laser beam was focused with the lens. The laser power at the sample position was adjusted with an attenuator. Using the reflected light from the beam splitter, it was confirmed that there was no change in the output of the laser beam.
Following mirror samples were used: oxygen-free high thermal conductivity-Cu (OFHC-Cu), pure Cu, aluminum oxide dispersion strengthened copper (ODSC), and pure Ag. The measured roughness and hardness are shown in Table 1 . The surface roughness of (i) OFHC-Cu mirror, which was manufactured by Kugler Co., was the best (Ra. = roughness < 3 nm). The ODSC mirror was the hardest and had a resistance for plastic deformation . In this study, some rough surfaces were used on purpose. For e.g., the surface of sample (iv) was intentionally roughened compared to the sample (iii) to investigate the influence of the roughness of the surface on the multi-pulse LIDT. Concerning the samples (vi) and (vii), since the surface was very hard, a normal surface treatment method could not be used to sufficiently smoothen the surface. The ODSC sample (v) was processed by elliptic oscillation cutting. We also used them because the comparison with these samples can also reveal the influences of the surface roughness on the multi pulse LIDT. The cutting Cu, ODSC, and Ag mirrors were made by a method of diamond turning. The polished Cu, ODSC, and Ag mirrors were mechanically polished with alumina suspension. For the mirrors surfaces, it should be said that alumina particles was observed by SEM (scanning electron microscope). The alumina particles were not embedded in but just attached on the surface.
2.2 Determination of multi-pulse LIDT
The power of the reflected light was measured with power meter and recorded with an oscilloscope. A typical temporal evolution of the reflected laser power is shown in Fig. 2 . The reflected laser power decreased seriously if the mirror surface was damaged by laser irradiation. In Fig. 2, the laser irradiation was conducted using the sample (i) at the pulse energy of 1.6 J/cm2. After 1.4 × 104 pulses irradiation, the surface was damaged and the output of the specularly reflected laser power was significantly reduced.
When the metal surface is damaged, the signal level of diffusion scattering increases , and the specular reflection from the metallic mirror decreases. In this study, we defined a multi-pulse LIDT at the point where the power of the reflected light became 95% of the initial power because the error of measurement equipment was about 5%.
3. Multi-pulse LIDT measurements
3.1 Validation of the measurement
Figure 3 presents the pulse number dependence of the LIDT for (i) OFHC-Cu. As explained in previous section, the LIDT was determined from the pulse number when the laser power decreased by 5%. From previous study , it was suggested that the pulse number dependence of LIDT can be explained with a power law similar as the fatigue of metals. It was presented that the multi-pulse LIDT can be explained with following equation:13]. In response to the laser irradiation, the surface temperature increase leads to the transient increase in the local stress on the metal. It was recognized that the multi-pulse LIDT was caused by the repetitive stress put on the surface; the above equation has been theoretically explained from the similarity with mechanical fatigue cycles. The same dependence can be seen in Fig. 3. From fitting of data, the LIDT was expressed as FN = (3.8 ± 0.3) × (N)^(−0.083 ± 0.015). To validate these values, we also measured a single shot LIDT experimentally with a conventional method . The measured single LIDT shown in Fig. 3 as a closed square was 3.5 J/cm2. The extrapolated value at single shot from the multi-pulse LIDT was 3.8 ± 0.3 J/cm2, which was consistent with the measured single LIDT.
3.2 LIDT of Cu mirrors
Multi-pulse LIDT measurements of Cu and Ag mirrors under multi-pulse laser irradiation were carried out up to ~106 pulses. The LIDT for the samples (i)-(iii) and (iv) are shown in Fig. 4 . The effects of the multi-pulse LIDT for (i) OFHC - Cu, (ii) cutting Cu, (iii) polished pure Cu1, and (iv) polished pure Cu2 were experimentally investigated up to 1.8 × 106, 1.6 × 105, 5.4 × 103, and 4.4 × 103 pulses, respectively. The sample (i) was irradiated with the highest number of pulses in the Cu mirrors, and the LIDT was 1.0 J/cm2 with 1.8 × 106 pulses. The expected multi-pulse LIDT for 107 pulses was 1.1 J/cm2 for (ii) Cutting Cu, 9.6 × 10−1 J/cm2 for (i) OFHC-Cu, 6.7 × 10−2 J/cm2 for (iii) Polished Cu2, and 6.4 × 10−2 J/cm2 for (iv) Polished Cu1 in descending order.
Among the Cu samples, the LIDTs for the samples (i) and (ii) can be explained by the model. For other samples, i.e., the samples (iii) and (iv), the LIDT suddenly dropped around 102-103 pulses and could not be explained by the model. In terms of the samples (i) and (ii), the surface condition was good, and the LIDT was on a line on a log-log plot. The slope of the sample (ii) was more moderate than that in . It was mentioned in  that the surface remains scratches after the surface processes and the surface quality was not so high. Thus, the difference in the slope is probably caused by the difference in the condition of the surface. The correlation coefficient was 0.89 for the sample (i) and was 0.81 for the sample (ii), indicating that the pulse number dependences of LIDT for the samples (i) and (ii) can be well explained with the power law. On the other hand, concerning the samples (iii) and (iv), the data was not on a single line on the plot, and the LIDT significantly decreased when the pulse number exceeded 102-103.
From the measured LIDTs for pure Cu mirrors, it can be said that the LIDT increases as decreasing the surface roughness. From the comparison between the sample (iii) and (iv), the LIDT increased by approximately 40% when the roughness decreased from 15 to 30 to 8-10 nm. In addition, the sample (i), which had minimum roughness of less than 3 nm, had the best LIDT among pure Cu samples. Furthermore, focusing on the difference of the samples (ii) and (iii), the material and the roughness were the same, but the method of processing the surface and the hardness of the surface were different. It is likely that the difference was caused by the difference in the method of processing the surface. The alumina particles attached on the surface might have decreased the LIDT for the samples (iii) and (iv), especially when the pulse number was higher than 102-103 pulses.
Concerning the hardness, it is likely that the LIDT increases with the hardness, because, in general, it is harder for dislocations to move and initiate plastic deformation when the hardness is high . On the same material, the difference in the hardness is caused by the work hardening during the polishing or cutting processes, while ODSC originally has higher hardness than pure Cu. The LIDT may also be organized from the view point of hardness. From the comparison between samples (i)-(iii) as excluding the high roughness sample (iv), the multi-pulse LIDT increases with the hardness. It is thought that the roughness of the surface, the surface processing method, and the hardness of the surface are important factors for the LIDT.
3.3 LIDT of ODSC mirrors
In general, ODSC samples have a high hardness and durability for plastic deformation. Figure 5 presents the multi-pulse LIDTs as a function of the pulse number for the samples (v), (vi), and (vii). The multi-pulse LIDTs were measured up to (v) 4.0 × 104, (vi) 2.0 × 104, and (vii) 3.7 × 103 pulses. In terms of ODSC mirrors, the roughness of the surface was deeply related to LIDT similar to pure Cu samples.
Interestingly, the LIDT for ODSC samples revealed a sharp decline when the pulse number exceeded 103-104. The expected multi-pulse LIDT for 107 pulses was 2.0 × 10−1 J/cm2 for (v), 1.6 × 10−2 J/cm2 for (vi), and 6.3 × 10−3 J/cm2 for (vii). Extrapolating the data to single shot, it is expected all the samples had relatively high LIDT, say higher than 2 J/cm2. Even for a cutting sample, the single shot LIDT is expected to be higher than 3 J/cm2, and the value would be better than pure Cu samples. The LIDT is higher than pure Cu samples for small pulse numbers, but it decreased drastically when the number of pulses becomes higher, for e.g., >103-104.
For pure cupper mirrors, it is likely that the inflection point was caused by the attached alumina particles. The alumina particles attached on the surface might have decreased the LIDT for the samples (iii) and (iv), especially when the pulse number was higher than 102-103 pulses. On the other hand, since on ODSC, all the samples had an inflection point, the different mechanism might have existed. At present, the reason to cause the sudden drop in LIDT for ODSC at high pulse number was not fully understood. ODSC have better mechanical property compared with pure cupper  and stable thermally. Further, the optical reflectivity of ODSC at 1064 nm was comparable to that of pure copper. Since the roughness and hardness of the sample (v) were not inferior to those of Cu mirrors, the results indicated that some other important factors that were not taken into account exist.
It is likely that mechanical and thermal properties were stable during the laser irradiation, and the thermal conductivity was comparable to pure Cu; it is thought that the property of bulk ODSC did not change during the laser pulse irradiation and did not have significant contribution on the prompt drop of LIDT. Considering the fact that the surface temperature is repeatedly raised by laser irradiation, there is a possibility that the alumina added on the material slightly changed the surface optical properties during the laser irradiation without any changes on the bulk properties. Although this is only speculation, further TEM (transmission electron microscope) analysis on the damaged ODSC sample is currently underway and will reveal the damage mechanism.
Figure 6 shows the SEM image of the surface of sample (vii) damaged by laser irradiation. It is seen that the damage was formed along with the processing line. The temporal evolution of the reflected laser power for the sample of (vii) is shown in Fig. 7 . When the sample has rough surface, the reflected laser power decreased gradually. It is likely that for the material that has a relatively rough surface, convex surfaces were damaged first, and consequently, the reflection ratio also decreases gradually.
3.4 LIDT of Ag mirrors
Figure 8 shows the multi-pulse LIDT of Ag in Table 1 as a function the pulse number. The effects of the multi-pulse LIDT of (viii) cutting Ag and (ix) polished Ag were experimentally investigated up to 1.1 × 105 and 8.3 × 104 pulses, respectively. For the sample (viii), the LIDT was 1.4 J/cm2 with 1.1 × 105 pulses. The expected multi-pulse LIDT for 107 pulses was 1.3 J/cm2 for sample (viii), and 0.80 J/cm2 for sample (ix). The open circles in Fig. 8 represent the point where the surface was not damaged even though the laser irradiation performed up to 2.6 × 105 pulses. The laser irradiation at 1.3 J/cm2 did not cause the damage, and this will support the idea that the LIDT can be extrapolated by the power law more than 2.6 × 105 pulses.
The multi-pulse LIDT of Ag mirrors shows the power law to the number of pulses. The value of the slope of the sample (viii) was −0.0079, which was smaller than the value of (ii) cutting Cu of −0.0373. Since the slope was shallow, the life time of Ag mirror was expected to be longer than Cu mirror. Comparing between (viii) and (ix) samples, the LIDT decreased about 30% as increasing the roughness by 30%.
4. Extrapolation to ITER TS system
In ITER TS system, it would be important to evaluate the multi-pulse LIDTs at the number of pulses of 107-109. The expected multi-pulse LIDTs at 107 and 109 pulses are summarized in Table 2 . Here, only the data explained by the power law, i.e., the sample (i), (ii), (viii), and (ix), were used and other data were omitted from Table 2.
The maximum LIDT at 107 pulses was 1.3 J/cm2 for the sample (viii), i.e., cutting pure Ag, and second one was 1.1 J/cm2 for the sample (ii), i.e., cutting pure Cu, and third one was 9.6 × 10−1 J/cm2 for the sample (i), i.e., OFHC-Cu. If the surface roughness of mirror was good (Ra. □ about 10 nm), the lifetime of mirrors seemed to become longer. The expected LIDT for polished pure Ag was relatively high even though the LIDT at lower pulse number was less than those of Cu samples. Thus, without considering the other influences than the pulsed heat loads from laser pulses, Ag mirror is one of the best choices for ITER TS system. In ITER TS system, it is necessary to consider the influence such as neutron load and other particle bombardment, though the fluence should be low. It will be of importance to investigate synergistic effects of nuclear radiation and plasma irradiation on the multi-pulse LIDT in future.
The LIDT for higher pulses were predicted assuming that the tendency for higher pulse numbers can be explained in the same manner as in the range of less than 106 pulses. However, there are no experiments at the moment for the pulse numbers higher than 107; we cannot exclude the possibility that the LIDTs are not on the power law when the pulse number exceed some value higher than 107. Thus, before actual installation, it will be of importance to check that the mirrors can be endured higher number of pulses at the laser power lower than the threshold value predicted in Table 2.
Moreover, for Ag mirrors, corrosion can be a problem. If the surface property was deteriorated before installing to vacuum vessel, the LIDT would significantly decrease. One option is a silver alloy that has durability to corrosion. It is of interest to investigate the LIDT further for silver alloys experimentally; the issues remained as a future work.
The multi-pulse LIDT of the OFHC-Cu mirror for 1.8 × 106 pulses was 1.0 J/cm2, and single shot LIDT was 3.5 J/cm2. The LIDT of cutting Cu mirror for 1.6 × 105 pulses was 1.2 J/cm2. The LIDT of the cutting Ag mirror for 1.1 × 105 pulses was 1.4 J/cm2. The LIDT of the polished Ag mirror for 8.3 × 104 pulses was 0.90 J/cm2.
The slope of the multi-pulse LIDT graph for Ag mirrors was less than that for Cu mirrors. The expected single shot LIDT on OFHC - Cu was the highest of the mirrors; however, the expected multi-pulse LIDT for 107 pulses on cutting Ag was the highest. Ag mirror would be one of the best choices for ITER TS system. It will be of importance to investigate synergistic effects of laser irradiation and radiation or plasma irradiation in future.
Damage threshold was highly dependent on surface condition. Even on the same metal material, as the roughness became poor, the LIDT was lower. And even the sample of the same material and same degree of roughness, the behavior of the LIDT was very different by the difference in surface processing method. The behavior would be related to the roughness and the hardness of the surface. The multi-pulse LIDT of the low number pulse may depend on the hardness of the surface. However the value of high number of pulses may depend on the surface processing methods and the roughness. For the samples polished with Al2O3 powder, the LIDT promptly dropped when the number of pulse exceeded 102-103 even though the roughness was comparable to cutting sample. This was probably due to small amount of Al2O3 remnants existed on the surface. In the same manner, on aluminum oxide dispersion strengthened copper (ODSC) samples, the LIDT sharply dropped when the pulse number exceeded 103. Although the reason to cause the sudden drop in LIDT for ODSC was not fully understood yet, the surface condition might have been changed during the laser pulse irradiation without changing the mechanical property of bulk material.
We thank Prof. T. Jitsuno and Dr. S. Motokoshi from Osaka University for useful discussion and comments. This work was supported in part by NIFS collaborative research program (NIFS12KLEH023). This work was supported in part by a Grant-in-Aid for Young Scientists (A) 23686133 from Japan Society for the Promotion of Science.
References and links
1. A. J. H. Donné, A. E. Costley, R. Barnsley, H. Bindslev, R. Boivin, G. Conway, R. Fisher, R. Giannella, H. Hartfuss, M. G. Hellermann, E. Hodgson, L. C. Ingesson, K. Itami, D. Johnson, Y. Kawano, T. Kondoh, A. Krasilnikov, Y. Kusama, A. Litnovsky, P. Lotte, P. Nielsen, T. Nishitani, F. Orsitto, B. J. Peterson, G. Razdobarin, J. Sanchez, M. Sasao, T. Sugie, G. Vayakis, V. Voitsenya, K. Vukolov, C. Walker, K. Young, and I. T. P. A. T. G. Diagnostics, “Chapter 7: diagnostics,” Nucl. Fusion 47(6), S337–S384 (2007). [CrossRef]
2. R. M. Wood, Laser Damage in Optical Materials (IOP Publishing, 1986).
3. V. Voitsenya, A. F. Bardamid, V. N. Bondarenko, W. Jacob, V. G. Konovalov, S. Masuzaki, O. Motojima, D. V. Orlinskij, V. L. Poperenko, I. V. Ryzhkov, A. Sagara, A. F. Shtan, S. I. Solodovchenko, and M. V. Vinnichenko, “Some problems arising due to plasma-surface interaction for operation of the in-vessel mirrors in a fusion reactor,” J. Nucl. Mater. 290–293, 336–340 (2001). [CrossRef]
4. D. V. Orlinski, V. S. Voitsenya, and K. Y. U. Vukolov, “First mirrors for diagnostic systems of an experimental fusion reactor I. Simulation mirror tests under neutron and ion bombardment,” Plasma Dev. Oper. 15(1), 33–75 (2007). [CrossRef]
5. A. Litnovsky, P. Wienhold, V. Philipps, G. Sergienko, O. Schmitz, A. Kirschner, A. Kreter, S. Droste, U. Samm, P. Mertens, A. H. Donné, D. Rudakov, S. Allen, R. Boivin, A. McLean, P. Stangeby, W. West, C. Wong, M. Lipa, B. Schunke, G. De Temmerman, R. Pitts, A. Costley, V. Voitsenya, K. Vukolov, P. Oelhafen, M. Rubel, and A. Romanyuk, “Diagnostic mirrors for ITER: A material choice and the impact of erosion and deposition on their performance,” J. Nucl. Mater. 363–365, 1395–1402 (2007). [CrossRef]
6. R. W. Moir, “Grazing incidence liquid metal mirrors (GILMM) for radiation hardened final optics for laser inertial fusion energy power plants,” Fusion Eng. Des. 51–52, 1121–1128 (2000). [CrossRef]
7. V. S. Voitsenya, V. G. Konovalov, M. F. Becker, O. Motojima, K. Narihara, and B. Schunke, “Materials selection for the in situ mirrors of laser diagnostics in fusion devices,” Rev. Sci. Instrum. 70(4), 2016–2025 (1999). [CrossRef]
8. S. Kajita, T. Hatae, and V. S. Voitsenya, “Assessment of laser transmission mirror materials for ITER edge Thomson scattering diagnostics,” J. Plasma Fusion Res. 3, 032 (2008). [CrossRef]
9. A. Gorshkov, I. Bel’bas, M. Maslov, V. Sannikov, and K. Vukolov, “Laser damage investigations of Cu mirrors,” Fusion Eng. Des. 74(1-4), 859–863 (2005). [CrossRef]
10. C. S. Lee, N. Koumvakalis, and M. Bass, “A theoretical model for multiple-pulse laser-induced damage to metal mirrors,” J. Appl. Phys. 54(10), 5727–5731 (1983). [CrossRef]
11. N. Kawagoishi, H. Nisitani, E. Kondo, A. Shimamoto, and R. Tashiro, “Influence of Stress change on the fatigue behavior and fatigue life of aluminum Oxide-dispersion-strengthening copper alloy at room temperature and 350 degree,” J. Soc. Mat. Sci. Japan 53(2), 182–187 (2004). [CrossRef]
12. A. Gorshkov, I. Bel’bas, V. Sannikov, and K. Vukolov, “Frequency laser damage of Mo mirrors,” Fusion Eng. Des. 66–68, 865–869 (2003). [CrossRef]
14. Y. Jee, M. F. Becker, and R. M. Walser, “Laser-induced damage on single-crystal metal surfaces,” J. Opt. Soc. Am. B 5(3), 648–659 (1988). [CrossRef]
15. S. Kohara, Kinzoku Zairyou Gairon (Introduction to Metal Material), (Asakura, 1991). [in Japanese]
16. S. J. Zinkle, “DOE/ER-0313/16, Semiannual Progress Report,” (Oak Ridge National Lab, 1994).