Open Access
Add to BibSonomyAbstract
Multi-pulse laser induced damage threshold (LIDT) for metallic mirrors are important issue for laser diagnostics in future fusion devices. In this paper, the mechanism of multi-pulse LIDT and the influence of the slip formation and oxidization in atmosphere were investigated experimentally with a Nd:YAG pulse laser whose pulse width and wavelength are ∼5 ns and 1064 nm, respectively. From detailed surface analysis of laser irradiated part by transmission electron microscopy (TEM), it was found that the miniaturization of crystal size and slip formation were observed on damaged area. Oxidization feature was also revealed from the TEM analysis. It was shown that the multi-pulse LIDT could be increased under vacuum condition compared with that in air atmosphere.
© 2013 Optical Society of America
1. Introduction
In future experimental nuclear fusion devices including ITER, many metallic mirrors will be used for optical diagnostics. For collection optical system, one of the issues is the deterioration of optical property by the particles from plasmas [1, 2]. Especially, the influence of the deposition layer should be serious in addition to the increase in the surface roughness by sputtering. Recently, several methods are being develop to recover the optical properties by removing the deposited layers using laser pulses, plasmas, and so on [3, 4]. In addition to the collection optics, metallic mirrors will be used for laser transmission for laser diagnostics. Especially, for laser Thomson scattering diagnostics, since high power pulsed laser will be used, the mirrors will be totally broken once the damage is formed on the surface. It is inevitable to consider the laser induced damage threshold (LIDT) to design the diagnostic systems.
To design the laser transmission mirrors, a multi-pulse LIDT should be considered [5]. For edge Thomson scattering diagnostics [6], a Nd:YAG laser with the laser wavelength of 1.06 μm, and the pulse energy, and pulse width of 5 J and 20–30 ns, will be used [7]. Assuming that the duration of plasmas discharge is 500 s and that the mirror will be used for 104 plasma pulses, the number of the laser pulses subjected to the mirror will be on the order of 108–109. In [8], multi-pulse LIDT for various materials are investigated by comparing the existed LIDT data and combining the with numerical calculations. It is thought that materials with high reflectance at the wavelength, namely, silver (Ag), copper (Cu), and gold (Au), are the candidate materials for the laser transmission materials for the edge Thomson scattering diagnostics. Considering workability, price, and corrosion behavior, Cu is a likely to be a good choice for the material.
Among various metals, Cu has been widely used for metallic mirrors for high power lasers, and the LIDT of Cu has been relatively investigated [5]. However, even for Cu, the experimental investigation of multi-pulse LIDT is quite limited. In [9], the multi-pulse LIDT for Cu at 10.6 μm was investigated up to 103 pulses, and the theoretical model of multi-pulse LIDT was presented. In [10], the multi-pulse LIDT for Cu at 1.06 μm was investigated up to 105 pulses. However, for any metals, there were no investigation of multi-pulse LIDT for higher number of pulses until a recent report from the author’s group about the multi-pulse LIDT of Cu up to 106[11]. In actual fusion relevant condition, mirrors will be exposed to laser pulses in vacuum. Moreover, they will be subjected to high energy ions, neutrons, and gamma rays, though the fluences are not so high. However, there are no report about the influences of those irradiations on LIDT and the multi-pulse LIDT under vacuum condition.
In this study, the mechanism of multi-pulse LIDT is discussed from the detailed surface analysis with transmission electron microscope (TEM). The analysis shows that the miniaturization of crystal, slip formation, and oxidization occurs during the laser irradiation. The effects of the neutron irradiation on the multi-pulse LIDT was demonstrated by using a ion beam irradiated Cu sample. Furthermore, the difference between the multi-pulse LIDTs of Cu in vacuum and air conditions are shown. The expected multi-pulse LIDT under ITER conditions are discussed based on the experimental results.
2. Experimental setup
Figure 1(a) and 1(b) show schematics of experimental setup for multi-pulse LIDT measurement in the air and vacuum conditions, respectively. A Nd:YAG (yttrium aluminum garnet) laser (Continuum, SLII-10) was used for the pulsed laser source. The wavelength, the pulse width, and the repetition frequency were 1.064 μm, 5–7 ns, and 10 Hz, respectively. The multi-pulse LIDT measurement was conducted in air and vacuum conditions.
The laser pulse energy was controlled using an attenuator. The laser power was continuously monitored during the LIDT measurement as splitting the fraction of the laser power using a beam splitter to monitor the laser power. A lens was used to control the beam size at the mirror surface. The reflected laser beam power was also monitored using a power meter. The incident angle of the laser to the mirror was approximately 6–7 degree. The spatial power profile of the laser beam was measured with a laser beam profiler. The power profile was well fitted with Gaussian profile and the radius was approximately ∼0.5 mm.
A turbo molecular pump was used for vacuuming. The background pressure was approximately 5 × 10−5 Pa, and a mirror sample was installed to the vacuum chamber. Since the accessibility to the mirror was less than that in the air atmosphere, the measurement system was simplified. Although the laser beam power was not monitored in vacuum, it was confirmed that the laser power was stable enough to measure the LIDT. The incident angle of the laser to the mirror was 22.5 degree in Fig. 1(b). The angle was greater in Fig. 1(b) compared with that in Fig. 1(a); however, since the polarization of the laser to the mirror was p-polarization, the angular dependence of the absorptance and the spot size compensate with each other [12], and the influence of the angular dependence disappears.
Table 1 summarizes the mirror samples used. Two different types of Cu mirror sample (i) and (ii), were used for the experiments. Sample (i) was prepared by polishing the surface with diamond turning tool. Sample (ii) is the commercially available oxygen free high conductivity Cu mirror (Kugler Co.). The sample (ii) had better surface nature in roughness and a hard surface.
Table 1. The roughness and hardness of samples (i) and (ii).
The LIDT was measured from the temporal evolution of reflected laser power. Figure 2 presents typical temporal evolutions of the reflected laser power in air and vacuum conditions, respectively. The pulse energy of the laser beam was fixed once the laser irradiation started, and the laser irradiation continued before damage was formed. Multi-pulse LIDT is the function of the number of the pulse. In the measurement, the power was fixed and the number of the pulse was measured. By changing the laser power, the LIDT was measured as a function of the number of the pulse.
Fig. 2 Typical Temporal evolutions of the laser power reflected from mirror surface. Two cases in vacuum and air conditions were presented.
Before the damage was formed, the reflectance of the surface was higher than 95%. However, the reflected laser power was promptly dropped once damages were formed on the surface especially in the air. In vacuum, the reflected laser power dropped slower than in the air.
For e.g., the experiment in Fig. 2 was conducted at the laser pulse energy of 2.2 J/cm2. The damage was formed at the pulse number was 6.4×104. Thus, from the experiment, we can say that the LIDT at the pulse number of 6.4×104 was 2.2 J/cm2. When the experiments were conducted in air atmosphere, the pulse energy promptly decreased to almost zero. On the other hand, in vacuum, the laser power decreased slowly after the damage was started to be formed on the surface. The difference will be discussed later.
To demonstrate the neutron irradiation, copper ion irradiation was carried out with a tandem accelerator at Kyushu University. The energy of the copper ion was 2.4 MeV. The ion flux measured with a Faraday cup was 1.0×1015 m−2s−1. The detail of the damages formed by the ion beam and the depth profiles are shown later.
3. Results and discussion
3.1. Multi-pulse LIDT of Cu in air atmosphere
Figure 3 shows the measured multi-pulse LIDT for two different Cu mirrors in the air condition. The damage power thresholds are plotted as a function of the number of the pulse in logarithmic scale for both axes. The LIDT decreases as increasing the number of the pulse. The LIDT of sample (i) was lower than that of sample (ii), especially when the number of the pulse is low, i.e., lower than 103 pulses. The difference may be caused by the difference in the surface roughness. When the number of the pulse was higher than 104, the LIDTs for samples (i) and (ii) were almost consistent with each other.
The scattering of the data is mainly caused by the fact that the damage formation is a stochastic process. For the LIDT measurement, it is difficult to obtain vertical error experimentally, because the experiments are conducted with a fixed laser power and the number of the pulse where the damage is formed is different every time. It is possible to obtain the horizontal error experimentally, if several experiments are conducted at the same laser pulse energy. In the present experiments, however, we conducted the irradiation with changing the pulse energy every time and showed the scattering of the data. For e.g., in Fig. 3, the LIDT at N = 100 for sample (ii) was ∼2.5 J/cm2 on average, but the LIDT seems to have a standard deviation of ∼0.5 J/cm2. The standard deviation may depend on the sample grade and should decrease with the number of the laser pulse.
3.2. Demonstration of neutron irradiation
The copper ions were irradiated to the sample (i) at the energy of 2.4 MeV to simulate the neutron irradiation damages. Expected damages by the ion beam irradiation were calculated by SRIM-2012 code. Figure 4(a) shows a projected picture of the calculated ion trajectory and cascade collision feature of 2.4 MeV copper atoms on a copper sample. Five hundreds of copper ions were injected to the surface and the traces were shown in Fig. 4(a). Purple dots represent the position where cascade collision event occurred. It is seen ions are injected to the surface and some ions reached the depth of ∼1 μm. Figure 4(b) presents the calculated depth profiles of dpa and apa, where dpa and apa stand for displacement per atom and atoms per atom, respectively. It is seen that the copper atoms are injected to ∼ 800 nm, and depth of the formed displacement is ≤1 μm.
Fig. 4 Calculation results of the Cu ion irradiation damage with SRIM-2012 code. (a) is the projected picture of the calculated ion trajectory and cascade collision feature and (b) is the calculated depth profiles of dpa and apa.
In addition to the calculation, the damages formed by the ion beam irradiation were observed by TEM. Figure 5(a) shows the TEM micrograph of the Cu sample exposed to the ion beam. The sample was prepared with focused ion beam (FIB) milling technique, and the thickness of the sample was ∼90 nm. Before cutting the sample by the FIB, tungsten was deposited on the top of the surface. Thus, the black part seen above the surface in TEM corresponds to the tungsten deposition layer. There are many black lines and black dots, representing long dislocations and small clusters of defects. It is unlikely that these defects and dislocations change the thermal conductivity and optical reflectivity significantly. In general, defects and dislocation loops suppress the motion of dislocations; it is thought that slip on mirrors could be suppressed by these defects. Indeed, in previous irradiation experiments for Cu mirror with Cu ion beam at a dose of 15 dpa, the degradation in reflectivity was not observed at 600 nm [13].
Figure 5(b) is an enlarged TEM image of the sample. It is seen that long black lines are composed of many dislocation loops. Black dots with high contrast should correspond to self-interstitial type dislocation loops. Black dots with low contrast could be self-interstitial type dislocation loops without the satisfaction of the Bragg condition or/and stacking fault tetrahedrons. In Fig. 5(c), a cross sectional view of the sample is shown. There is a difference in contrast around 1100 nm. It is likely that self-interstitial type dislocation loops formed by the FIB process can be suppressed at the places where many vacancies exist; the depth of 1100 nm would corresponds to the area where damages by the ion beams were formed. The depth was well agree with the simulation results shown in Fig. 4(b). Two samples were prepared with different irradiation doses, i.e. 5.5 and 16.5 dpa at peaks.
In Fig. 6, multi-pulse LIDT for the samples without irradiation, with irradiation at 5.5 and 16.5 dpa are shown. The LIDTs for the three samples are almost the same; in addition, interestingly, for the low pulse number less than 103, the LIDT with irradiation was higher than that without irradiation for 16.5 dpa. Although Fig. 6 lacks the LIDT for 5.5 dpa mirrors below 104 pulses, it seems that the LIDT in the low pulse number region for 5.5 dpa is slightly higher than those without irradiation, since the LIDT for 5.5 dpa at 104 seems to still have higher value compared with those for 10–100 pulses without irradiation. In ITER, the dose for the laser transmission mirror should be significantly lower than that for the first wall, which is expected to be on the order of 1 dpa [14]. Even though the dose in the present experiment is significantly higher than the expected dose for the laser transmission mirrors, the influence of the irradiation did not appear on the LIDT. It can be said that this is a good news for laser transmission mirrors. From previous irradiation study, it was presented that the reflectance of Cu mirror did not change by the formation of vacancies [13, 1]. In addition to the reflectivity, the present study revealed that the irradiation would not be harmful to the laser material interaction.
Fig. 6 Multi-pulse LIDT for OFHC Cu mirrors in the air without irradiation and with irradiation at 5.5 and 16.5 dpa.
3.3. Mechanism of multi-pulse LIDT
Figure 7(a) shows a SEM micrograph of sample (ii) after the irradiation with laser pulses. The left hand side corresponds to the laser central region, while the right hand side corresponds to slightly edge part of the laser beam. The TEM samples were prepared by FIB milling from the central part and the edge part of laser beam from the sample. In the central part of the laser beam, ripples are formed in dense, while in the edge part, the ripples are formed but not so dense.
Fig. 7 (a) A SEM micrograph of sample (ii) after the irradiation with laser pulses. (b) and (c) cross sectional views of the TEM micrographs of the sample. (b) corresponds to the central part the laser beam and (c) edge part of the laser beam.
Figure 7(b) and (c) shows cross sectional views of the TEM micrographs of the sample. Figure 7(b) and (c) correspond to the central part and edge part of the laser beam, respectively. Flat surface before the laser pulse irradiation became rough after the irradiation. And the height of the roughness was approximately 0.5–1 μm. Near the surface, up to 4–5 μm, the crystal size became much smaller than the original. These miniaturization of crystal occurs when strong stress is put on. The miniaturization should occur by the laser irradiation not by the polishing process, because the phenomenon was not observed on the edge part, as shown in Fig. 7(c). It is thought that the increase in the temperature in response to the laser pulse mounted the stress on the surface, and, consequently, it changes the crystal size and surface roughness.
In Fig. 7(c), though the surface was not so rough, a small protrusion is observed on the surface. Interestingly, beneath the protrusion, slip was formed. And along the slip line, size of the crystal became smaller. Slip line connected to deeper region, and it seems that slip deformation occurred at the depth of 4–5 μm.
Figure 8(a)–(c) shows expanded TEM images of the sample at the laser beam center region. As seen Fig. 8(a), top of the surface layer in 100–150 nm was blackend in the TEM image. It is likely that strong stress was deposited by repetitive thermal expansion. Figure 8(b) and (c) are the bright field image (BFI) and dark field image (DFI) from the same position. In the diffraction pattern shown in the inset of Fig. 8(c), diffraction ring can also be seen. In DFI, many white spots can also be observed in the surface thin layer 100 nm. The white spot is probably originated from the oxidized copper. It is likely that oxidized layer is formed during the damaging process, and the oxidization enhances the damages by the next pulses. As seen in Fig. 2, the difference was seen in the damaging process in air and vacuum conditions. In air, the reflected power was promptly dropped after the damage was formed, while the power slowly decayed in vacuum. After the sample was damaged in the air, the color of the sample was changed to black around the laser irradiation region. It is likely that the oxidization contribute to proceed damages in the air. Once the oxidization was initiated, the surface reflectance was changed and consequently damage was formed in a catastrophic manner.
Fig. 8 Expanded TEM images of the sample at the laser beam center region. (b) and (c) are the bright field image (BFI) and dark field image (DFI) from the same position.
3.4. Multi-pulse LIDT of Cu in vacuum and extrapolation to higher pulses
Concerning the multi-pulse LIDT experiments for metallic mirrors, to our knowledge, all the experiments reported have been conducted in the air. However, in ITER, the mirrors will be irradiated by the laser pulses in vacuum condition.
Figure 9 shows the multi-pulse LIDT for OFHC-Cu in air and vacuum conditions. The multi-pulse LIDT in vacuum was higher than that in air with increasing the number of the pulses. The slope in the double logarithmic scale plot was shallow, indicating that the LIDT should be much higher with increasing the number of the pulse. In Fig. 9, open squares show the experimental data where the damage was not formed on the surface for the number of the pulse higher than 105. As predicted from Sec. 3.3, in vacuum, oxidization process does not occur, and consequently, the LIDT becomes higher than that in air atmosphere.
Here we extrapolate the LIDT to the expected ITER condition, i.e. 108–109 pulses by assuming that the relation between the LIDT and the number of the pulse can be expressed with the power law. It has been revealed that the pulse number dependence of LIDT can be explained with a power law similar as the fatigue of metals [15]. The data in Fig. 9 were fitted with power function. The obtained function was (3.76±0.32)N(−0.083±0.015) for the data in air atmosphere and (3.73±0.97)N(−0.049±0.025) for the data in vacuum. Interestingly, the extrapolated value at the single shot is consistent with each other and with the single shot damage threshold of 3.5 J [11] within the error.
Using the data in the air condition, the expected LIDT is 0.82 (+0.35, −0.25) J for 108 pulses and 0.67 (+0.33, −0.22) J/cm2 for 109 pulses. In vacuum condition, the expected LIDT is 1.51 (+1.51, −0.80) J/cm2 for 108 pulses and 1.35 (+1.50, −0.75) J/cm2 for 109 pulses. Although the ambiguity is still large, the LIDT in vacuum could be twice higher than that in the air when the pulse number is higher than 108 pulses. Especially, in vacuum, the data is still not sufficient to extrapolate it to greater number of the pulse expected in ITER.
The LIDT could slightly vary with the laser beam size. The dependence has been investigated in detail previously [16]. The ratio of the power thresholds at the laser radius r and the infinite laser radius can be obtained using the laser size and thermal properties of material, and it is thought that the same principle can also be applied to multi-pulse LIDT. In the present situation, the ratio was approximately 1.05, meaning that the present value is ∼5% greater than that in a large radius case. Considering the influence, the expected LIDT for 108 pulse is 0.78 (+0.33, −0.24) J/cm2 under air condition and 1.44 (+1.44, −0.76) J/cm2 under vacuum condition, and that for 109 pulse is 0.64 (+0.31, −0.21) J/cm2 under air condition and 1.29 (+1.42, −0.72)J/cm2 under vacuum condition.
4. Conclusions
Multi-pulse laser induced damage threshold (LIDT) was investigated experimentally using Nd:YAG laser with the time duration of ∼ 5 ns in the air and vacuum conditions. From the data in the air condition, the expected LIDT was 0.78 (+0.33, −0.24) J/cm2 for 108 pulses and 0.64 (+0.31, −0.21) J/cm2 for 109 pulses, while in vacuum condition, it was 1.44 (+1.44, −0.76) J/cm2 for 108 pulses and 1.29 (+1.42, −0.72) J/cm2 for 109 pulses. The values in vacuum was twice higher than that in the air. The mechanisms of LIDT was investigated by the detailed surface analysis using TEM (transmission electron microscope). It was found that the miniaturization of crystals occurred near the surface, and slips were formed. The slip formation lead to the formation of protrusions on the surface. Also, it was found that oxidization occurred in the air condition. In the air, when the slip formation and oxidization were initiated, a significant reduction in the optical reflectance would occur, leading to damages in a catastrophic manner.
In ITER, since the pulse duration is approximately five times longer than the present laser pulse, the expected LIDT will be two to three times higher than the values in the present study. Considering that the factor corresponds to a safety factor and assuming that the laser pulse energy will be ∼5 J [7] for ITER edge Thomson scattering, the necessary laser irradiated area on the mirror is 3.9 cm2 without considering the angle dependence on the LIDT. It is worthwhile to note that the area is more than an order of magnitude smaller than the previously estimated one [17], and the present value is more reliable. Before fabricating the mirror, it is recommended to confirm experimentally that the damage is not formed for higher pulse numbers such as 109 in vacuum at the designed laser pulse energy.
Acknowledgments
This work was supported in part by a Grant-in-Aid for Young Scientists (A) 22246120 from Japan Society for the Promotion of Science (JSPS). This work was supported in part by NIFS collaborative research program( NIFS12KLEH023). This work was conducted as a part of the collaborative research with RIAM Kyushu University.
References and links
1. D. V. Orlinski, V. S. Voitsenya, and K. Y. Vukolov, “First mirrors for diagnostic systems of an experimental fusion reactor i. simulation mirror tests under neutron and ion bombardment,” Plasma Devices and Operations 15, 33–75 (2007) [CrossRef] .
2. D. V. Orlinski, V. S. Voitsenya, and K. Y. Vukolov, “First mirrors for diagnostic systems of an experimental fusion reactor ii. the mirror tests on the large fusion devices under operation,” Plasma Dev. Operations 15, 127–146 (2007) [CrossRef] .
3. A. Litnovsky, M. Laengner, M. Matveeva, C. Schulz, L. Marot, V. Voitsenya, V. Philipps, W. Biel, and U. Samm, “Development of in situ cleaning techniques for diagnostic mirrors in ITER,” Fusion Eng. and Des. 86, 1780–1783 (2011) [CrossRef] .
4. A. Widdowson, J. Coad, G. de Temmerman, D. Farcage, D. Hole, D. Ivanova, A. Leontyev, M. Rubel, A. Semerok, A. Schmidt, and P.-Y. Thro, “Removal of beryllium-containing films deposited in JET from mirror surfaces by laser cleaning,” J. Nucl. Mater. 415, S1199–S1202 (2011) [CrossRef] .
5. 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, 2016–2025 (1999) [CrossRef] .
6. S. Kajita, T. Hatae, and O. Naito, “Optimization of optical filters for ITER edge thomson scattering diagnostics,” Fusion Eng. and Des. 84, 2214–2220 (2009) [CrossRef] .
7. T. Hatae, J. Howard, N. Ebizuka, H. Yoshida, M. Nakatsuka, H. Fujita, K. Narihara, I. Yamada, H. Funaba, Y. Hirano, H. Koguchi, S. Kajita, and O. Naito, “Progress in development of the advanced thomson scattering diagnostics,” Journal of Physics: Conference Series 227, 012002 (2010) [CrossRef] .
8. S. Kajita, T. Hatae, and V. S. Voitsenya, “Assessment of laser transmission mirror materials for ITER edge Thomson scattering diagnostics,” Plasma Fusion Research 3, 032 (2008) [CrossRef] .
9. C. S. Lee, N. Koumvakalis, and M. Bass, “A theoretical model for multiple-pulse laser-induced damage to metal mirrors,” J. Appl. Phys. 54, 5727–5731 (1983) [CrossRef] .
10. A. Gorshkov, I. Bel’bas, M. Maslov, V. Sannikov, and K. Vukolov, “Laser damage investigations of Cu mirrors,” Fusion Eng. and Des. 74, 859–863 (2005) [CrossRef] .
11. M. Sato, S. Kajita, R. Yasuhara, N. Ohno, and Y. Tawara, “Assessment of multi-pulse laser-induced damage threshold of metallic mirrors for thomson scattering system,” Optics Express 21, 9333–9342 (2013) [CrossRef] [PubMed] .
12. J. F. Figueira and S. J. Thomas, “Damage thresholds at metal surfaces for short pulse IR lasers,” IEEE J. Quantum Electron. 18, 1381 (1982) [CrossRef] .
13. V. S. Voitsenya, V. G. Konovalov, A. F. Shtan’, S. I. Solodovchenko, M. F. Becker, A. F. Bardamid, K. I. Yakimov, V. T. Gritsyna, and D. V. Orlinskij, “Some problems of the material choice for the first mirrors of plasma diagnostics in a fusion reactor,” Rev. Sci. Instrum. 70, 790–793 (1999) [CrossRef] .
14. J. Linke, P. Lorenzetto, P. Majerus, M. Merola, D. Pitzer, and M. Rödig, “EU development of high heat flux components,” Fusion Sci. Technol. 47, 678–685 (2005).
15. M. F. Becker, C. Ma, and R. M. Walser, “Predicting multipulse laser-induced failure for molybdenum metal mirrors,” Appl. Opt. 30, 5239–5246 (1991) [CrossRef] [PubMed] .
16. J. Porteus, D. Decker, W. Faith, D. Grandjean, S. Seitel, and M. J. Soileau, M. J. “Pulsed laser-induced melting of precision diamond-machined Cu, Ag, and Au at infrared wavelengths,” IEEE J. Quantum Electron. 17, 2078–2085 (1991) [CrossRef] .
17. S. Kajita, S. Takamura, N. Ohno, and T. Nishimoto, “Alleviation of helium holes/bubbles on tungsten surface by use of transient heat load,” Plasma Fusion Res. 2, 009 (2007) [CrossRef] .
