Open Access
Add to BibSonomy
Abstract
Metal-dielectric low dispersion mirrors (MLDM) have a promising application prospect in petawatt (PW) laser systems. We studied the damage characteristics of MLDM and found that the damage source of MLDM (Ag + Al2O3+SiO2) is located at the metal-dielectric interface. We present the effect of the interface on the femtosecond laser damage of MLDM. Finite element analysis shows that thermal stress is distributed at the interface, causing stress damage which is consistent with the damage morphology. After enhancing the interface adhesion and reducing the residual stress, the damage source transfers from the interface to a surface SiO2 layer, and the damage threshold can be increased from 0.60 J/cm2 to 0.73 J/cm2. This work contributes to the search for new techniques to improve the damage threshold of MLDM used in PW laser systems.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
In petawatt (PW) laser systems [1–5], the damage performance of optical film components directly affects the operation stability of the laser system. Low-dispersion mirrors (LDMs) is the most frequently used optical reflectors in the entire system. LDM must withstand high power density with the construction of the PW laser system [4]. Moreover, the pulse width is inversely proportional to the bandwidth; tens to hundreds of nm bandwidths are necessary for pulses of durations less than or approximately 100 fs [6]. For instance, in the Station of Extreme Light laser project led by SIOM, Shanghai, China, the laser uses coherent beam combination that generate four 30-PW pulses to deliver 1.5 kJ in 15 fs for the target deliver of 100 PW [4]. Hence, the corresponding reflection bandwidth requires at least 200 nm. Metal mirrors are commonly used because metals exhibit good reflectivity in the visible and infrared wavelengths; however, the damage threshold of the metal mirrors cannot meet the requirement of PW laser systems. Compared with metal mirrors, all dielectric mirrors exhibit a satisfactory reflectivity close to 100%; however, they cannot provide a broader reflective bandwidth and the mechanical stress can significantly increase when the number of layers and size increases [7–9]. To avoid the aforementioned problems, metal-dielectric mirrors with a broad reflective bandwidth and high laser induced damage threshold (LIDT) gradually became the research emphasis for high-power ultrafast pulse lasers [10–14].
Scientists across the world conducted significant research on metal-dielectric low-dispersion mirrors (MLDMs). In the 40-fs broadband low dispersion mirror thin film damage competition [11], lots of low-dispersion mirrors were tested. J.B.Oliver used plasma-ion-assisted electron-beam evaporation to fabricate MLDM on 10 inch aperture substrates [13]. Václav Škoda studied different substrates and technologies for fabricating MLDM. The LIDT values of MLDM are regularly higher in BK7 substrate than in fused silica samples and MLDM containing a Ag layer, which was prepared using the Ion beam deposition method, however, the LIDT values of MLDM are lower in BK7 samples than in MLDM containing a classically evaporated Ag layer [15]. Victoria investigated the bandgap, peak electric field, and manufacturing technology dependence of the femtosecond damage threshold of more than 30 real-life femtosecond multilayer mirrors [16]. The research work on MLDM mainly focuses on testing and characterization. However, there are few studies focusing on the damage mechanism of MLDM. In the current PW laser systems, large-diameter MLDM are also used [17]. In the future, with the development of 100-PW-level laser systems, MLDM with broad reflection bands will be more widely used. Enhancing the damage threshold of MLDM is always a challenging problem in high-power laser system regimes. Interface properties are one of the main factors influencing the LIDT values of dielectric mirrors. Stolz observed voids in the multilayer interface, which decreased the interface quality [18]. Gao investigated the laser-induced damage behavior of narrow-band interference filters irradiated using a Nd:YAG laser at 1064 nm under the single-pulse mode and found that the damage was attributed to the Ta2O5 interfaces [19]. McInnes studied the damage characteristics of a ZnS/MgF2 Fabry–Perot filter and found that the damage initiated at the ZnS boundaries. A study showed high defect density and weak mechanical strength at the boundaries [20]. Because of the different properties of the layer materials, the interfaces differ considerably. The effect of the dielectric multilayer interface on the threshold has been widely reported [19–21]. The method of film deposition, the layer materials and even the layer deposition sequence will affect the damage resistance of the interface. However, the damage investigation of MLDM is undetailed, and the interfacial effect of MLDM on the threshold remains unknown. Victoria used an advanced design to show that the electric field at the interface decreases, and the damage threshold can be increased by 40%, the threshold of MLDM can reach 1.06J/cm2 for s-polarization [16]. These results verify that the interfacial problems of MLDM exist; however, the damage mechanism of the metal-dielectric mirror interface has not been studied in detail.
In this report, we found that the damage source of MLDM is located at the interface of metal and dielectric layers. Consequently, we present the first results reporting the effect of the interface on metal-dielectric mirrors in femtosecond regimes. The electron beam evaporation technique is used to deposit MLDM (Ag + Al2O3+SiO2). The damage threshold test is performed for all samples using femtosecond laser pulses. Damage morphologies reveal that the typical damage morphology of MLDM is stress damage. Finite element analysis shows that thermal stress is distributed at the interface which is consistent with damage morphology. After annealing, the residual stress in the layers is decreased and the interface adhesion is enhanced. As expected, the damage threshold is increased by 22%, and the damage source transfers from the interface to surface SiO2. Overall, it is concluded that low adhesion is the main reason for the low damage threshold of metal-dielectric mirror, and interface enhancement is an effective approach for increasing the damage threshold. This work contributes to the search for new techniques to improve the damage threshold of MLDM used in PW laser systems.
2. Experiment
Two types of MLDM were fabricated using the electron beam evaporation technique, and they comprised a single Ag layer (150 nm) + Al2O3 transition layer (30 nm) + SiO2 protective layer (547.88 nm). The whole assembly was operated within a vacuum environment, down to 2 × 10−3 mbar, in this work. Metal coatings were deposited at ambient temperature and as quickly as possible for maximum reflectance, which was determined to be approximately 10 nm/s. All dielectric layers were deposited at a deposition rate of 1.5 nm/s. The difference between the two samples is that one sample was annealed in an atmospheric environment. Further, the Ag mirror and SiO2 dielectric layer were fabricated. Four different group names are shown in Table 1.
Table 1. Layer systems studied.
The damage threshold test was performed for all samples in the 1-on-1 mode [22], i.e., each location on the sample was irradiated using only one laser pulse. A mechanical shutter was used to achieve 1-on-1 mode. The interval time is 1 ms and it can meet the laser pulse frequency of 1 KHz. 10 sites were irradiated by the same fluence. The experimental setup is shown in Fig. 1.
Fig. 1. Experimental setup of the 1-on-1 damage threshold test for our mirrors. The tests were performed at 45° angle of incidence (AOI) and P polarization.
We used laser pulses generated by the Ti:sapphire ultrashort and ultrahigh power laser system, which delivers 30-fs pulses with a repetition rate of 1 kHz and a central wavelength of 800 nm. The spectral range is 800 (±35 nm). Measuring pulse width 30 fs refers to the pulse width before the pulse reaches the sample. The measurement spot size is 0.221(±0.006) mm2. CCD imaging method is used to get the measurement spot size. The fluctuation range of the spot area is between 0.215-0.227 mm2 and the error range is 3%. The pulse energy is strictly consistent before and after the test, we do believe that the experimental errors on the actual pulse energy is in fact lower, on the order of 1%. According to these laser parameters, the relative error of the damage threshold measurement can be assessed below 5%. The incident angle was 45° for P polarization. A charged-coupled device camera was used to detect the extent of damage. In this experiment, the damage threshold of the samples was determined using the maximum value of all laser fluences with 0% damage probability.
3. Results
3.1 Reflectivity and dispersion
The reflectivity of MLDM with and without annealing is shown in Fig. 2. The reflectivity decreases after annealing. This phenomenon has been reported in a previous study [23]. Ag generally exhibits a close-packed structure, and there are some voids with the same size as a Ag atom [24]. Ag diffuses in the dielectric layer when it receives more energy from the outside than the migration barrier. Generally, the probability of Ag diffusion can be expressed as $\textrm{W}(E )\propto \textrm{exp}({ - E/{k_B}T} )$, where $\textrm{W}(E )$ is the probability of the Ag atom receiving energy greater than E, ${k_B}$ is the Boltzmann constant, and T is the ambient temperature. Therefore, after annealing, a transition layer of Al2O3 and Ag is formed, which decreases the reflectivity. These data suggest that adhesion at the interface is enhanced by the presence of the transition layer. The group delay dispersion of low dispersion mirror we designed is only ±5 fs2, as shown in Fig. 2.
3.2 Damage threshold
The damage thresholds of different layer structures are shown in Fig. 3. The threshold value is the average value of four same samples. The damage thresholds of four samples are 0.38 (±0.02), 0.60 (±0.01), 0.73 (±0.01), and 1.1 (±0.01) J/cm2. It is known that the absorption of the metal is sufficiently high, therefore, metal mirrors cannot achieve a high damage threshold. The damage threshold significantly increases after depositing a dielectric protective layer (548 nm SiO2). The damage threshold of MLDM can be increased from 0.6 J/cm2 to 0.73 J/cm2 after annealing. Single SiO2 film exhibits the highest damage threshold, which is 1.1 J/cm2.
Fig. 3. Laser damage threshold of four group samples. The standard deviation (the error bar) among same samples is considered.
3.3 Damage morphology
The damage morphologies of the two MLDMs were observed using the Zeiss–Auriga scanning electron microscope. As shown in Figs. 4 and 5, the damage morphologies of two samples are different. For II, at the central area of the damage pit, there is metal ablation, as shown in Fig. 4(a). The typical damage morphology is stress damage, as shown in Fig. 4(b). A focused ion beam (FIB) was used to observe the section morphology. As shown in Fig. 4(c), from the cross-section morphology of the crater rim, we can observe that damage source occurs at the interface of the Ag and Al2O3. However, for sample III, as shown in Fig. 5(b), the damage pit is more regular and no apparent stress damage is observed. As shown in Fig. 5(a), there is no metal ablation at the central area of the damage pit. It indicates that the metal layer has not been damaged. The cross-section morphology shown in Fig. 5(c) reveals that the SiO2 layer is damaged. Moreover, the Ag layer is intact, and the Al2O3 transition layer still covers the metal layer. Thus, currently, we have evidence that the increase in the damage threshold is attributed to the transfer of the damage source from the interface to the surface dielectric layer.
Fig. 4. Typical damage morphologies of MLDM without annealing (II). (a) Damage morphology at the center of the damage pit. There is metal ablation. (b) Damage morphology at a fluence of 0.62 J/cm2 near the damage threshold. The damage caused by stress. (c) Cross-section morphology at the edge of the damage pit.
Fig. 5. Typical damage morphologies of MLDM with annealing (III). (a) Damage morphology at the center of damage pit. There is no metal ablation. (b) Damage morphology at a fluence of 0.75 J/cm2 near the damage threshold. (c) Cross-section morphology at the edge of the damage pit.
4. Discussion
Atkinson [25,26] proposed a method for calculating the residual stress in coating by measuring the curvature of the substrate before and after coating. The formula can be expressed as:
where ${t_s}$ and ${t_f}$ are the thickness of the substrate and film, respectively, K is the curvature of the substrate, and $E_s^{\prime}$ is the elasticity modulus of the substrate. If ${t_f}$ is considerably smaller than ${t_s}$, Eq. (1) can be simplified asThe parameters used in the calculation are shown in Table 2.
Table 2. The parameters used in the calculation of residual stress.
Figure 6 shows the surface flatness of MLDM before (a) and after (b) the annealing procedure. After annealing, the MDLM surface is flatter and shows tensile stress. The power value of MLDM before and after annealing is 0.025 and 0.003 wave, respectively, and the wavelength is 632.8 nm. The residual stress values calculated using Eq. (4) before and after annealing are 26.9 and 3.2 MPa, respectively. Combined together, these results suggest that annealing can reduce the residual stress of MLDM.
In the femtosecond laser regime, the damage characteristics of the material are closely related to the electric field distribution. Femtosecond pulsed lasers have broad spectrum characteristics, and different frequency components will affect the damage of MLDM. Figure 7(a) is the laser spectrum used in the experiment. The spectral range is 750-850 nm.The normal electric field intensity (NEFI) distribution under different frequency components of MLDM is calculated using the commercial software TFCalc [Fig. 7(b)]. The electric field intensity at the interface is around 0.45. Therefore, these data cannot explain why the initial damage occurs in the interface.
Fig. 7. (a) Incident laser spectrum. (b) Electric field distribution under different frequency components.
Based on the previously calculated electric field distribution, we can determine the heat distribution according to the following heat transfer equation: $\mathrm{\rho }{C_p}\frac{{\partial T}}{{\partial t}} = \nabla ({k\cdot\nabla T} )+ Q$, where $\mathrm{\rho }$ is the density, ${C_p}$ is the heat capacity, k is the thermal conductivity, and Q is the heating source. In the temperature field calculation, left and right boundaries are set as the period boundary conditions and the upper and lower boundaries are set as the thermal radiation boundaries. The convective heat flux is ${q_0} = h\cdot({{T_{ext}} - T} )$, where h is the heat transfer coefficient and set as 6W/(m2·K) and ${T_{ext}}$ is the ambient temperature and set as initial temperature ${T_0}$ (293 K). Figure 8(b) shows the profile of Ag temperature. The lattice temperature basically does not change before 30 fs, because this is the electron heating stage. After the pulse irradiation, the lattice temperature begins to increase. Eventually, the lattice temperature reaches the relative equilibrium state. The highest temperature is 942 K, which does not exceed the melting point of Ag (1235 K). The calculated temperature distribution of the MLDM is shown in Fig. 8(a). The parameters used in the calculation are shown in Table 3. The input laser energy density is 0.75 J/cm2. Although the SiO2 layer exhibits a high electric field intensity, its temperature increases slowly because of its small absorption coefficient. The temperature is mainly concentrated in the metal layer.
Fig. 8. (a) Calculated temperature distribution. The high temperature area of the Ag film is mainly concentrated on the surface. (b) Lattice temperature for MLDM under pulse irradiation. The temperature in Ag close to the interface to Al2O3. The laser fluence is 0.75 J/cm2.
Table 3. Parameters and values used in the calculation.a
The vertical distribution of the temperature results shows an uneven distribution of stress. Hooke’s law is used to calculate the thermal stress distribution in the MLDM. The thermal stress can be simplified and expressed as ${\varepsilon _t} = \alpha (T )({T - {T_{ref}}} )$, where $\alpha $ is the coefficient of thermal expansion, ${T_{ref}}$ is the temperature of the surrounding environment (293.15 K), and T is the temperature distribution of MLDM after thermal diffusion. In the thermal stress calculation, air is excluded and the surrounding boundaries are set as a free boundary. The initial strain and stress are zero. The parameters used in the calculation are listed in Table 3.
The calculation of thermal stress distribution is shown in Fig. 9. The regime of high thermal stress is in the Ag layer and the interface. In general, thermal stress is released at the weak site (the interface of Ag and Al2O3). Therefore, by combining the findings of damage morphology and theoretical model, we can conclude that under the 30-fs layer pulse, the Ag layer absorbs a large amount of energy, resulting in an uneven thermal stress distribution, which is released at the interface. Under the combination of residual and thermal stresses, the stress overcomes the adhesion in the interface, which causes the dielectric layer to be peeled. The primary effect of annealing is the formation of a dense transition layer of Al2O3 and Ag and reduction in the residual stress inside MLDM. The presence of a transition layer enhances the adhesion of the interface and reduces the probability of stress damage.
Fig. 9. Calculated thermal stress distribution. The thermal stress is mainly concentrated on the interface.
5. Conclusions
Here, we investigate the damage property of MLDM. Results indicate that the damage source of MLDM is at the interface of metal and dielectric layers. Moreover, we present results reporting the interfacial effect on metal-dielectric mirrors. By comparing the damage threshold with and without interfacial enhancement, we observe that interface enhancement has a positive effect on the damage threshold. The damage threshold can be increased from 0.60 J/cm2 to 0.73 J/cm2. Damage morphology shows that the damage of MLDM without annealing is initiated from the Ag and Al2O3 interface. After interface enhancement, the damage is initiated from the SiO2 dielectric layer. The theoretical model shows that the electric field peak is not located at the metal interface, but the temperature peak is observed at the metal surface. Further analysis shows that the stress is distributed at the Ag and Al2O3 interface, which is consistent with the damage morphology. Thus, it is concluded that low adhesion is the main reason for the low damage threshold of metal-dielectric mirrors. Through interface enhancement, the adhesion at the interface is improved and residual stress in layers is reduced; thus, the probability of stress damage at the interface is reduced. This work contributes to the search for new techniques to improve the damage threshold of MLDM or metal-dielectric grating used in PW layer systems.
Funding
National Key Research and Development Program of China (2018YFE0118000); National Natural Science Foundation of China (11904376); Research on the Key Technology of TMT Large Aperture Wide Angle Spectral Dichroic Mirror (U1831211); NSAF Joint Fund (U1630140); Youth Innovation Promotion Association of the Chinese Academy of Sciences (2017289); Strategic Priority Research Program of CAS (XDB1603).
Disclosures
The authors declare no conflicts of interest.
References
1. W. Li, Z. Gan, and L. Yu, “339 J high-energy Ti:sapphire chirped-pulse amplifier for 10 PW laser facility,” Opt. Lett. 43(22), 5681–5684 (2018). [CrossRef]
2. I. Y. Skobelev, S. N. Ryazantsev, D. D. Arich, P. S. Bratchenko, A. Y. Faenov, T. A. Pikuz, P. Durey, L. Doehl, D. Farley, C. D. Baird, K. L. Lancaster, C. D. Murphy, N. Booth, C. Spindloe, P. McKenna, S. B. Hansen, J. Colgan, R. Kodama, N. Woolsey, and S. A. Pikuz, “X-ray absorption spectroscopy study of energy transport in foil targets heated by petawatt laser pulses,” Photonics Res. 6(4), 234–237 (2018). [CrossRef]
3. X. Liang, X. Xie, and J. Kang, “Design and experimental demonstration of a high conversion efficiency OPCPA pre-amplifier for petawatt laser facility,” High Power Laser Sci. Eng. 6(4), e58 (2018). [CrossRef]
4. C. N. Danson, C. Haefner, and J. Bromage, “Petawatt and exawatt class lasers worldwide,” High Power Laser Sci. Eng. 7(3), e54 (2019). [CrossRef]
5. Q. G. Wang, N. Cui, and Y. P. Yao, “High peak power 1.0 µm and tri-wavelength 1.3 µm Nd:ScYSiO5 crystal lasers,” Chin. Opt. Lett. 17(11), 111403 (2019). [CrossRef]
6. J. C. Bellum, E. S. Field, and T. B. Winstone, “Low Group Delay Dispersion Optical Coating for Broad Bandwidth High Reflection at 45° Incidence, P Polarization of Femtosecond Pulses with 900 nm Center Wavelength,” Coatings 6(1), 11 (2016). [CrossRef]
7. J. Chen, H. Huang, and Y. Zhang, “Reducing electric-field-enhancement in metal-dielectric grating by designing grating with asymmetric ridge,” Sci. Rep. 8(1), 5228 (2018). [CrossRef]
8. J. N. Dong, J. Fan, S. D. Mao, Y. P. Lan, Y. G. Zou, H. Z. Wang, J. B. Zhang, and X. H. Ma, “Effect of annealing on the damage threshold and optical properties of HfO2/Ta2O5/SiO2 high-reflection film,” Chin. Opt. Lett. 17(11), 113101 (2019). [CrossRef]
9. C. Shan, Y. A. Zhao, Y. Q. Gao, X. H. Zhao, G. H. Hu, W. X. Ma, and J. D. Shao, “Laser-induced defects in optical multilayer coatings by the spatial resolved method,” Chin. Opt. Lett. 17(3), 031403 (2019). [CrossRef]
10. R. Y. Chen, Y. Z. Wang, and K. L. Guo, “Design, fabrication and laser damage comparisons of low-dispersive mirrors,” Proc. SPIE 1106, 110630F (2019). [CrossRef]
11. R. A. Negres and C. J. Stolz, “40-fs broadband low dispersion mirror thin film damage competition,” Proc. SPIE 10014, 100140E (2016). [CrossRef]
12. C. J. Stolz, L. M. Sheehan, M. K. Von Gunten, R. P. Bevis, and D. J. Smith, “150-ps broadband low dispersion mirror thin film damage competition,” Proc. SPIE 9632, 96320C (2015). [CrossRef]
13. J. B. Oliver, J. Bromage, C. Smith, D. Sadowski, C. Dorrer, and A. L. Rigatti, “Plasma-ion-assisted coatings for 15 femtosecond laser systems,” Appl. Opt. 53(4), A221–A228 (2014). [CrossRef]
14. V. Csajbók, Z. Bedőházi, B. J. Nagy, and P. Dombi, “Ultrafast multipulse damage threshold of femtosecond high reflectors,” Appl. Opt. 57(2), 340–343 (2018). [CrossRef]
15. V. Skoda and J. Vanda, “A study of metal-dielectric mirrors technology with regard to the laser-induced damage threshold,” Proc. SPIE 10014, 1001424 (2016). [CrossRef]
16. V. Csajbók, L. Szikszai, B. J. Nagy, and P. Dombi, “Femtosecond damage resistance of femtosecond multilayer and hybrid mirrors,” Opt. Lett. 41(15), 3527–3530 (2016). [CrossRef]
17. A. Hervy, L. Gallais, G. Cheriaux, and D. Mouricaud, “Femtosecond laser-induced damage threshold of electron beam deposited dielectrics for 1-m class optics,” Opt. Eng. 56(1), 011001 (2017). [CrossRef]
18. S. C. Weakley, C. J. Stolz, and Z. L. Wu, “Role of starting material composition in interfacial damage morphology of hafnia-silica multilayer coatings,” Proc. SPIE 3738, 318–324 (1999). [CrossRef]
19. W. D. Gao, M. Q. Zhan, and Z. S. Fan, “Laser-induced damage of Ta2O5/SiO2 narrow-band interference filters under different 1064 nm Nd:YAG laser modes,” Appl. Surf. Sci. 250(1-4), 195–202 (2005). [CrossRef]
20. A. McInnes, C. MacDonald, D. Gibson, and A. Wilson, “Investigation and Modelling of Laser Damage Properties of Fabry-Perot Filters,” Proc. SPIE 1438, 14381A (1989). [CrossRef]
21. H. Hu, Z. Fan, and F. Luo, “Laser-induced damage of a 1064-nm ZnS/MgF2 narrow-band interference filter,” Appl. Opt. 40(12), 1950–1956 (2001). [CrossRef]
22. ISO 21254-1:2000 (E), Laser and laser-related equipment—determination of laser-induced damage threshold of optical surfaces—Part 1: 1-on-1 test.
23. J. B. Wu and Y. X. Jin, “Effect of Annealing temperature on metal/Dielectric multilayers for fabricating broadband pulse compression gratings(in Chinese),” J. Inorg. Mater. 29(10), 1087–1092 (2014). [CrossRef]
24. X. L. Sun and J. D. Shao, “Effect of Alumina and Chromium Interlayers on Microstructures and Optical Properties of Thin Ag Films on Glass Substrates,” Chin. Opt. Lett. 23(9), 2576 (2006).
25. C. Klein, “How accurate are Stoney’s equation and recent modifications,” J. Appl. Phys. 88(9), 5487–5489 (2000). [CrossRef]
26. S. Shao, Z. X. Fan, and J. D. Shao, “Evolutions of residual stress and microstructure in ZrO2 thin films deposited at different temperatures and rates,” Thin Solid Films 445(1), 59–62 (2003). [CrossRef]
27. K. Guo, Y. Wang, and R. Chen, “Effects of structural defects on laser-induced damage of 355-nm high-reflective coatings sputtered on etched substrates,” Opt. Mater. 89, 173–177 (2019). [CrossRef]
28. K. Guo, Y. Wang, and R. Chen, “Effects of ion beam etching of fused silica substrates on the laser-induced damage properties of antireflection coatings at 355 nm,” Opt. Mater. 90, 172–179 (2019). [CrossRef]
