The efficiency of B4C, Mo and Zr barrier layers to improve thermal stability of Mg/Co multilayer up to 400 °C is investigated. Multilayers were deposited by direct current magnetron sputtering and characterized using X-ray and extreme ultraviolet reflection. The results suggest that B4C barrier layer is not effective due to drastic diffusion at Mg-B4C interface. Although introducing Mo barriers improves the thermal stability from 200 to 300 °C, it increases the interface roughness and thus degrades the optical performances. On the contrary, Zr barriers can significantly increase the thermal stability of Mg/Co up to 400 °C without optical performance degradation. Thus, Mg/Zr/Co/Zr is suitable for EUV applications requiring both optimal optical performances and heat resistance.
© 2011 OSA
High reflective multilayer mirrors are widely used as optical elements for applications such as extreme ultraviolet (EUV) microspectroscopy , high harmonic femtosecond chemistry , solar astrophysics imaging [3,4], and synchrotron radiation . Mg-based multilayers, such as Mg/SiC, Mg/Y2O3 and Mg/Co are promising in the 25-40nm wavelength range, the Mg L3 absorption edge being located at 25nm [6–10]. Mg/SiC obtained a peak reflectance of 44% at 31.2nm at 10° of incidence . Mg/Co has a peak reflectance of 40.3% at 30.5nm at 10° . Tri-material Mg/Sc/SiC reported has a peak reflectivity of 48.7% at 36.8 nm .
Multilayers mirrors usually endured a high flux of incident light or high heat loads in applications such as synchrotron radiation and solar imaging. However, Mg is known to have a low melting point (650 °C) , making it difficult to improve the thermal stability of Mg-based multilayers. Previous studies demonstrated that Mg/SiC and Mg/Y2O3 are thermally stable below the temperature of 200 °C, but deteriorate drastically at higher temperatures [6,10]. According to previous research, Mg/Co can be stable at 200 °C . Further investigation on the heat resistance of this new multilayer has not been reported. Compared with Mg/SiC, Mg/Y2O3 and Mg/Sc/SiC, Mg/Co has narrower bandwidth which produced better spectral resolution , making it more attractive for applications requiring a narrow spectral bandwidth, such as monochromatic solar imaging. Thus, it is important to investigate and improve the thermal stability of Mg/Co for such practical applications.
To improve the thermal stability, diffusion barrier layers can be inserted between the Mg and Co layers. B4C is a stable ceramic and also a common diffusion barrier layer. The efficiency of B4C barrier layer has been demonstrated for some multilayers. B4C diffusion barriers with thicknesses between 0.3 and 1.0 nm increased the thermal stability from 150 up to 400 °C for Mo/Si  and from 100 up to 200 °C for Sc/Si . Mo and Zr barriers are expected to be efficient to prevent diffusion in Mg/Co multilayer, since both Mo and Zr have high melting point and do not chemical react with Mg or Co below 650 °C [12,15]. Moreover, the introduction of barrier layer in Mg/Co can also improve reflectance when the order of the materials within a period is correctly chosen [16,17]. Thus, in this paper, we investigate the thermal stability of Mg/Co multilayer and then, that of Mg-Co based multilayers where B4C, Mo and Zr barriers have been introduced to enhance thermal stability.
Four sets of multilayer samples with a gamma value of 0.83 (dMg/(dMg + dCo)) have been prepared, including Mg/Co, Mg/B4C/Co/B4C, Mg/Mo/Co/Mo and Mg/Zr/Co/Zr. The number of period is 30, and the period thickness is near 17.0 nm. All the samples were deposited onto polished silicon wafers (100) with a size of 30×30 cm2 using direct current (DC) magnetron sputtering method. A 3.5 nm-thick B4C capping layer was deposited onto the surface of each sample to prevent oxidation. Following deposition, each sample was cut into nine equal small pieces of 10×10 cm2 for heat resistance study. The small samples were mounted on a plate heated by a wire-wound furnace in a vacuum chamber with a base pressure of 3×10−4 Pa. The samples were heated from room temperature to 300, 350 or 400°C, respectively, and held for 1 hour. The temperature was monitored by a thermocouple gauge attached the back of plate.
X-ray and EUV reflectometry measurements were carried out to test the changes in the multilayers before and after annealing. Grazing incident X-ray reflection (GIXR) measurements were made in the θ-2θ reflection geometry mode, using an X-ray diffractometer working at Cu Kα line (0.154nm). The fit of the GIXR curves performed with Bede Refs software was used to determine individual layer thickness, interface roughness, and layer density . EUV reflectance measurements were made at 45°, using the reflectometer operating on the Spectral Radiation Standard and Metrology Beamline and Station (U26) at the National Synchrotron Radiation Laboratory (NSRL), China. The details on deposition and measurement procedure have been described elsewhere .
3. Results and discussion
3.1 Thermal stability of Mg/Co multilayer
The measured X-ray and EUV reflectance curves of Mg/Co multilayers before and after annealing are shown in Fig. 1 . MgCo_1, MgCo_2, MgCo_3 and MgCo_4 stand for the as-deposited sample and those annealed at 300, 350 and 400°C, respectively. The period thickness of MgCo_1 MgCo_2 and MgCo_3 are 16.9, 16.7 and 16.7nm respectively.
As shown in Fig. 1(a), upon annealing at 300 °C, the high-order Bragg peaks become broadened and slightly shift towards higher angle, which correspond to a slight period thickness contraction and inter-diffusion increase. The structure of MgCo_3 encountered significant degradation upon 350 °C annealing, as the high Bragg orders are not well-defined any more. Interface roughness values derived from GIXR curves were 0.4, 0.5 and 1.2nm for MgCo_1, MgCo_2 and MgCo_3, respectively. Following 400 °C annealing, the multilayer structure of MgCo_4 is completely destroyed since no Bragg peak is presented in corresponding curve. Consistent with GIXR measurements, the measured EUV reflectance in Fig. 1(b) shows slight reflectance decrease at 300 °C and then, notable decline at 350 °C, while no reflectance is obtained upon 400 °C annealing. Previous work has demonstrated that Mg/Co multilayer was stable up to 200 °C annealing . However, at higher temperatures, its structure changed and optical performances degraded. Hence, interface engineering such as barrier introduction is required to improve the thermal property of Mg/Co.
3.2 Mg/Co multilayer with B4C barrier layers
Figure 2(a) presents the GIXR curves of Mg/B4C/Co/B4C and Mg/Co both made of 10 periods. The thickness of B4C layer is 0.9 nm. The period thickness of Mg/Co multilayer and Mg/B4C/Co/B4C is 8.0 and 9.8 nm, respectively. The results suggest that the quality of Mg/B4C/Co/B4C is worse than that of Mg/Co. To identify which interface, Co-B4C or Mg-B4C, was responsible for the poor performance of the Mg/B4C/Co/B4C, Mg/B4C and Co/B4C multilayers were prepared and characterized by GIXR measurements. Figure 2(b) shows the GIXR curve of a 120 bi-layered Co/B4C multilayer with a period thickness of 3.6 nm. The well-defined and intense Bragg peaks indicate abrupt interfaces in Co/B4C. According to the fit of the experimental curve, the thicknesses of Co and B4C are estimated to be 1.8 and 1.7nm, respectively, while the interface widths (σ) is 0.6 nm for σB4C-on-Co and σCo-on-B4C. As already reported in our previous work , Mg/B4C multilayer had a peak reflectivity of only 0.2% at 30.4nm: it is a direct consequence of its poor structural quality resulting from significant inter-diffusion at the interfaces. Chemical reaction at Mg-B4C interface, if any, is not clear in this work and will be investigated in a forthcoming study. Therefore, the stable B4C ceramic, though effective in other multilayers as a diffusion barrier layer [13,14], is not suitable for Mg/Co, mainly due to the poor structural quality of Mg-B4C interface.
3.3 Mg/Co multilayer with Mo barriers
Figure 3(a) shows the GIXR measurement of Mg/Mo/Co/Mo multilayer. Mg/Mo/Co/Mo_1, _2, _3 and _4 stand for the as-deposited sample and annealed at 300, 350 and 400 °C, respectively. Their periods are 17.2, 17.2, 17.8 and 17.8 nm, respectively. No significant structural changes are observed for Mg/Mo/Co/Mo_2 upon annealing at 300 °C. Following annealing at 350 °C, the Bragg peaks of Mg/Mo/Co/Mo_3 are broadened and their positions are significantly shifted towards smaller angle: this suggests increase of both inter-diffusion and period thickness. Upon 400 °C annealing, less Bragg peaks are now observed, indicating that the structure quality becomes worse. Thus, the introduction of Mo as barrier layer improves the thermal resistance of Mg/Co multilayer from 200 up to 300°C. According to the fit of the experimental reflectivity curve, the interface roughness is 0.7 nm for both MgMoCoMo_1 and _2. It was not possible to fit Mg/Mo/Co/Mo_3 and _4 reflectivity curve: given their poor interface quality, a multilayer model cannot be introduced for fitting purpose. Corresponding changes can be observed in the EUV reflectance measurements shown in Fig. 3(b). The EUV peak reflectance decreased while the peak position shifts towards longer wavelength upon 350 and 400 °C annealing. The peak reflectance of Mg/Mo/Co/Mo (34.6%) is lower than that of as-deposited Mg/Co (42.2%) for two reasons: the high absorption of Mo and the increased interface roughness. Moreover, the full width at half maximum (FWHM) of Mg/Mo/Co/Mo (5.5nm) is much larger than that of Mg/Co (0.4nm).
3.4 Mg/Co multilayer with Zr barrier layers
Figure 4 presents the experimental X-ray and EUV reflectivity curves of as-deposited and annealed Mg/Zr/Co/Zr multilayers. It can be seen that Mg/Zr/Co/Zr multilayer remains almost stable upon temperature increase. When annealing up to 400 °C, the EUV peak reflectance only decreases slightly, while no significant structural change can be deduced from the analysis of the GIXR curves. This slight change of reflectance can be attributed to the slight differences of samples. The interface roughness value extracted from the fit of the experimental curves of Fig. 4(a) is 0.5 nm for the four samples. In addition to its good thermal stability, Mg/Zr/Co/Zr is characterized by a high EUV reflectivity (44.5%) close to that of Mg/Co (42.6%). Thus, Zr is a suitable barrier layer for Mg/Co multilayer, since its introduction at both interfaces improves the thermal stability from 200 to 400 °C, without reducing EUV reflectance.
For comparison, Fig. 5 shows the evolution, as a function of the annealing temperature, of the normalized EUV reflectance decrease (NRD) measured for the Mg/Co, Mg/Mo/Co/Mo and Mg/Zr/Co/Zr multilayers. The value of NRD, defined as:Figs. 1(b), 3(b) and 4(b), respectively. As mentioned above, the Mg/Mo/Co/Mo is stable up to 300 °C but is not efficient anymore for higher temperatures. On the contrary, Mg/Zr/Co/Zr can resist heat treatment up to 400 °C. Thus, introducing Zr as diffusion barriers can successfully improve the thermal stability of Mg/Co multilayer.
According to our investigations, the optical performances of Mg/Co multilayer remain stable if the multilayer is not annealed higher than 200°C. In order to enhance this temperature limit, barrier layers made of B4C, Mo and Zr are introduced into Mg/Co. From the analysis of the evolution, as a function of the annealing temperature, of the X-ray and EUV reflectivity curves, B4C is not suitable for Mg/Co multilayer, since inter-diffusion may take place at Mg-B4C interface. The introduction of Mo improves thermal stability up to 300 °C but is accompanied of a loss of EUV reflectivity. The introduction of Zr significantly improves the thermal stability of Mg/Co (up to 400°C) without degrading EUV reflectance. In conclusion, the introduction of Zr barrier layer constitutes an efficient method to upgrade the thermal stability of Mg/Co multilayer for EUV applications. This thermal resistance is mandatory for applications such as astronomical observation and synchrotron radiation where high thermal charges are encountered.
This work is supported by the National Natural Science Foundation of China (Grant No. 10825521, 11061130549 and 10905042), and by 973 program (Grant No. 2011CB922203) and by Agence Nationale de la Recherche (Grant No. 2010-INTB-902-01).
References and links
1. Y. Kondo, T. Ejima, H. Takatsuka, and M. Watanabe, “Microscopic ultraviolet photoelectron spectroscopy using He-I and He-II resonance lines,” Surf. Rev. Lett. 9(1), 521–527 (2002). [CrossRef]
2. N. Dudovich, O. Smirnova, J. Levesque, Y. Mairesse, M. Yu. Ivanov, D. M. Villeneuve, and P. B. Corkum, “Measuring and controlling the birth of attosecond XUV pulses,” Nat. Phys. 2(11), 781–786 (2006). [CrossRef]
3. D. L. Windt, S. Donguy, J. Seely, B. Kjornrattanawanich, E. M. Gullikson, C. C. Walton, L. Golub, and E. DeLuca, “EUV multilayers for solar physics,” Proc. SPIE 5168, 1–11 (2004). [CrossRef]
4. M. Suman, M. G. Pelizzo, D. L. Windt, and P. Nicolosi, “Extreme-ultraviolet multilayer coatings with high spectral purity for solar imaging,” Appl. Opt. 48(29), 5432–5437 (2009). [CrossRef] [PubMed]
5. Y. Y. Platonov, D. Broadway, B. DeGroot, B. Verman, B. Kim, G. Gutman, J. Wood, J. Rodriguez, and N. Grupido, “Deposition of X-Ray Multilayers on Long Size Substrates for Synchrotron Applications,” Proc. SPIE 3152, 231–239 (1997). [CrossRef]
6. T. Ejima, A. Yamazaki, T. Banse, K. Saito, Y. Kondo, S. Ichimaru, and H. Takenaka, “Aging and thermal stability of Mg/SiC and Mg/Y2O3 reflection multilayers in the 25-35 nm region,” Appl. Opt. 44(26), 5446–5453 (2005). [CrossRef] [PubMed]
7. J. T. Zhu, Z. S. Wang, Z. Zhang, F. L. Wang, H. C. Wang, W. J. Wu, S. M. Zhang, D. Xu, L. Y. Chen, H. J. Zhou, T. L. Huo, M. Q. Cui, and Y. D. Zhao, “High reflectivity multilayer for He-II radiation at 30.4 nm,” Appl. Opt. 47(13), C310–C314 (2008). [CrossRef] [PubMed]
8. J. Zhu, S. Zhou, H. Li, Q. Huang, Z. Wang, K. L. Guen, M.-H. Hu, J.-M. André, and P. Jonnard, “Comparison of Mg-based multilayers for solar He II radiation at 30.4 nm wavelength,” Appl. Opt. 49(20), 3922–3925 (2010). [CrossRef] [PubMed]
9. K. Le Guen, M.-H. Hu, J.-M. André, P. Jonnard, S. K. Zhou, H. Ch. Li, J. T. Zhu, Z. S. Wang, and C. Meny, “Development and interfacial characterization of Co/Mg periodic multilayers for the EUV range,” J. Phys. Chem. C 114(14), 6484–6490 (2010). [CrossRef]
10. H. Maury, P. Jonnard, K. Le Guen, J.-M. André, Z. Wang, J. Zhu, J. Dong, Z. Zhang, F. Bridou, F. Delmotte, C. Hecquet, N. Mahne, A. Giglia, and S. Nannaronne, “Thermal cycles, interface chemistry and optical performance of Mg/SiC multilayer,” Eur. Phys. J. B 64(2), 193–199 (2008). [CrossRef]
11. A. Aquila, F. Salmassi, Y. Liu, and E. M. Gullikson, “Tri-material multilayer coatings with high reflectivity and wide bandwidth for 25 to 50 nm extreme ultraviolet light,” Opt. Express 17(24), 22102–22107 (2009). [CrossRef] [PubMed]
12. A. A. Nayeb-Hashemi and J. B. Clark, “The Mg-Zr (Magnesium-Zirconium) System,” Bull. Alloy Phase Diagrams 6(3), 246–250 (1985). [CrossRef]
13. T. Bottger, D. C. Meyer, P. Paufler, S. Braun, M. Moss, H. Mai, and E. Beyer, “Thermal stability of Mo/Si multilayers with boron carbide interlayers,” Thin Solid Films 444(1-2), 165–173 (2003). [CrossRef]
14. P. Jonnard, H. Maury, K. Le Guen, J.-M. André, N. Mahne, A. Giglia, S. Nannarone, and F. Bridou, “Effect of B4C diffusion barriers on the thermal stability of Sc/Si periodic multilayers,” Surf. Sci. 604(11-12), 1015–1021 (2010). [CrossRef]
15. A. A. Nayeb-Hashemi and J. B. Clark, Phase Diagrams of Binary Magnesium Alloys (ASM International, 1988), pp 204–205.
16. K. Le Guen, M.-H. Hu, J.-M. André, P. Jonnard, S. K. Zhou, H. C. Li, J. T. Zhu, Z. S. Wang, N. Mahne, A. Giglia, and S. Nannarone, “Introduction of Zr layers in Mg/Co multilayers for EUV applications,” Appl. Phys., A Mater. Sci. Process. 102, 69–77 (2011).
17. K. Le Guen, M.-H. Hu, J.-M. André, S. K. Zhou, H. Ch. Li, J. T. Zhu, Z. S. Wang, C. Meny, A. Galtayries, and P. Jonnard, “Observation of an asymmetrical effect when introducing Zr in Mg/Co multilayers,” Appl. Phys. Lett. 98(25), 251909 (2011). [CrossRef]
18. M. Wormington, C. Panaccione, K. Matney, and D. Bowen, “Characterization of structures from X-rayscattering data using genetic algorithms,” Philos. Trans. R. Soc. Lond. A 357(1761), 2827–2848 (1999). [CrossRef]