Laterally graded periodic Mo/B₄C multilayer for extreme ultraviolet wavelength of 6.8-11.0 nm

EUV and soft X-ray are powerful sources used in microscopy, astronomical observation, biological science and material science. However, the weak refraction of EUV and soft X-ray in all materials, coupled with high absorption, makes conventional refraction-based optics not suitable for EUV and soft X-ray region. Multilayer mirrors, consisted of alternate layers with high electron density materials (Mo, W, Sc, etc.) and low electron density materials (Si, C, B₄C, etc.), can reflect the EUV and soft X-ray radiations. High electron density materials are selected as the absorber while the low electron density ones are used as the spacer. On the basis of Bragg's formula, high reflectance can be achieved when period thickness of multilayer satisfies the Bragg's formula either for a particular wavelength or a particular grazing angle due to the constructive interference of all the interfaces^[1-3].

A problem for uniform multilayer mirrors when applied in EUV and soft X-ray region is the small working wavelength range due to their single period thickness structure. There are two ways to solve this problem: depth graded structured multilayer and laterally graded structured multilayer. For a depth graded structured multilayer, the period thicknesses vary with depth so that each bilayer is effectively tuned to a different X-ray wavelength. It can achieve broadband reflectance along the whole multilayer mirror^[4-5]. For a particular wavelength at a fixed incidence angle, only one bilayer period thickness satisfies the Bragg's formula in a depth graded multilayer. Thereby, the reflectance is relatively low compared with uniform multilayer mirrors. Furthermore, due to the aperiodic structure of depth graded multilayer, spectral bandwidth will be broadened and significantly influence the spectral resolution. As an alternative, laterally graded multilayer mirror has been studied in previous articles^[6-7]. For a laterally graded multilayer structure, period thicknesses vary continuously along the horizontal direction. Either the desired working wavelength or incidence angle can be simply achieved by translating the laterally graded multilayer to the corresponding position with the particular periodic thickness. Laterally graded multilayer not only simplifies the experiment alignment but also achieves high reflectance over a broad working wavelength with high spectral resolution. Chain Liu et al designed two different masks and fabricated two linearly graded W/C multilayers with average gradient of 0.027 nm/mm for tunable X-ray double-monochromator applications^[6]. Transmittance of the double-multilayer monochromator was more than 50% in the 9.9 to 10.1 keV range. Using graded multilayers in a double-monochromator configuration, one can conveniently adjust the bandpass and peak positions of the transmitted beams. Laterally graded multilayer provides an attractive way for polarization studies on magnetic materials and useful applications in EUV metrology, astronomy and microscopy in EUV and soft X-ray range^[8].

The design of laterally graded multilayer for synchrotron radiation reflector and polarization analysis requires determining the best materials combination. B₄C-based multilayers have shown relatively high theoretical reflectance in the previous researches owing to the low absorption coefficient of B at wavelengths longer than the B-K edge near 6.6 nm^[9]. Higher theoretical reflectivities at 1.5° off normal incident angle can be achieved with La/B₄C and Ru/B₄C multilayers in 7 nm wavelength, which are 63.4% and 43.5% respectively. However, the reaction between the lively La and B₄C at the interface and the serious diffusion at the interface of Ru and B₄C caused a dramatic loss to the reflectance^[10-12]. Mo/B₄C stack has stable physical and chemical properties. It can form multilayer structure with smooth interfaces and good thermal stability^[12-14]. Mo/B₄C multilayer was more widely used in the practical multilayer application for soft X-ray near 7 nm. In this paper, a Mo/B₄C layered system is designed for synchrotron radiation as a broadband reflecting mirror for extreme ultraviolet wavelength interval of 6.8-11.0 nm. Laterally graded [Mo/B₄C]₆₀ multilayer sample was prepared by magnetron sputtering on silicon wafer. The quality and optical property of the sample are evaluated by grazing incidence X-ray reflectivity and synchrotron radiation reflectance measurements.

1. Experimental details

Mo/B₄C laterally graded multilayer films of 60 bilayers were grown on polished Si (100) wafers (30 mm wide and 70 mm long) by direct current (DC) magnetron sputtering in argon of 99.999% purity. The base pressure was well below 1.5×10^-4 Pa. During the deposition process, a 0.15 Pa argon pressure was used in the deposition chamber. Two 4-inch(10.16 cm) ring sputtering sources were used. The target purities were 99.95% for Mo and 99.5% for B₄C. The plasma discharges were established with a DC power of 120 W for B₄C target and a DC power of 40 W for Mo target.

The individual Mo and B₄C layer thicknesses in laterally graded multilayers were adjusted by two specially formed deposition flux shaping masks fixed on the top of shield cans and directly above the sputtering targets. The film thickness distribution of magnetron sputtering and the profile-coating technique have been extensively studied in the previous studies^{[7, 15]}. The profile of the mask aperture was designed according to the required thickness profile. For masks used in the deposition process, the width of the aperture (perpendicular to the substrate moving direction) was 110 mm, and the length (along the substrate moving direction) at different transverse positions was determined by calculating the total material accumulated at these positions when the substrate passes over the aperture. As shown in Fig. 1, the substrate was moving linearly at a constant speed during the deposition. The distance from the target to the mask was 60 mm and from the mask to substrate was 10 mm. The substrate faced downward and it was driven by a computer-controlled stepper-motor to pass each magnetron source.

Figure  1.  Schematic diagram of the configuration used for the graded sputter deposition

DownLoad: Full-Size Img PowerPoint

After deposition, the multilayers were investigated with grazing incidence X-ray reflectivity (GIXRR) on a five-circle diffractometer with the Cu-Kα line at 8.04 keV (0.154 nm). The size of X-ray spot was 100 μm×2 mm. GIXRR measurements were made in θ-2θ geometry over the range of 0° < θ < 8°. Fits to the GIXRR data were used to determine layer thicknesses and interfacial parameters. EUV reflectance measurements were performed with synchrotron radiation (SR) by reflectivity-meter at the National Synchrotron Radiation Laboratory (NSRL). The reflectivity measuring device and measuring method have been described in Ref.[16]. An 1800 L/mm monochromator grating was used in the measurement. The wavelength resolution was better than 0.05 nm and the size of focused X-ray spot was 1×5 mm². The sample stage was placed in a vacuum chamber with a pressure greater than 1.5×10^-4 Pa. EUV reflectance measurements of the multilayer sample were made in the wavelength range of 6.0-11.0 nm with an incident angle of 45°. The reflectance measurements were performed every 5 mm interval along the gradient direction of the multilayer sample.

2. Results and discussion

The results of grazing incidence X-ray reflectivity measurements are shown in Fig. 2. For the convenience of comparison, the reflection curves were, in turn, translated vertically by two orders of magnitude.

Figure  2.  GIXRR curves measured at 8.04 keV of different point location x along the mirror

DownLoad: Full-Size Img PowerPoint

Fig. 2 shows that the Bragg reflective peaks shift monotonically towards the large angle direction as the measured location x changes from 5 mm to 68 mm. It can be noticed that the higher order Bragg peaks are spread out. This is due to the fact that the X-ray beam has a lateral size of 2 mm along the direction of the D-spacing gradient. Hence, the X-ray effectively probes a 2 mm area with spread D-spacing, leading to the widening of the Bragg peaks. The angle positions of the peaks are determined by the multilayer period thickness and the incident X-ray wavelength according to the Bragg's law. Therefore, it can be inferred from the GIXRR curves that the shift of Bragg peaks represents a variation of the multilayer period thickness at different location x along the sample. Based on the Bragg's formula, the multilayer period thicknesses D were calculated according to the angle value of the Bragg peaks. The calculated results are listed in Table 1.

Table  1.  Measured results of the GIXRR shown in Fig. 2 and the synchrotron radiation (SR) shown in Fig. 4

measured location x/mm	SR reflective peak λ/nm	FWHM, Δλ/nm	fitted period thickness D by SR/nm	calculated period thickness D by GIXRR/nm
5	10.60	0.31	7.98	7.82
10	10.12	0.29	7.57	7.44
15	9.67	0.27	7.19	7.03
20	9.25	0.27	6.83	6.74
25	8.80	0.26	6.47	6.47
30	8.41	0.23	6.10	6.16
35	8.05	0.20	5.86	5.83
40	7.69	0.19	5.59	5.56
45	7.39	0.18	5.34	5.33
50	7.09	0.17	5.11	5.10
55	6.79	0.16	4.88	4.88

 | Show Table

DownLoad: CSV

Fig. 3 shows the distribution of multilayer period thicknesses. The horizontal axis represents the location x along the sample in the long direction, and the vertical coordinate represents the multilayer period thicknesses. It can be concluded that multilayer period thicknesses vary continuously along the horizontal direction of the sample from 7.82 nm to 4.39 nm with an average gradient of 0.054 nm/mm as the measured location x increases from 5 mm to 68 mm.

Figure  3.  Period thickness D of measured location x along the sample

DownLoad: Full-Size Img PowerPoint

The EUV reflectivity of multilayer was measured by synchrotron radiation (SR) at the incident angle of 45°. Fig. 4 shows the measured and fitted results. The measured reflectance data are plotted as open circles, and solid lines represent the fitted curves. The measured results indicate that different location along the gradient direction on the multilayer mirror is effectively tuned to a different X-ray wavelength in reflecting geometry. The reflection peak is located at 10.6 nm near the edge of the mirror where location x=5 mm, and it has the trend of shifting to short wave-length direction when the location x along the mirror increases. Ultimately, the reflection peak shifts to 6.79 nm as x increases to 55 mm. Reflectivities of all measured points on the mirror are about 10%. FWHM (full width at half maximum) of the reflectance peaks varies from 0.13 nm to 0.31 nm with the increase of multilayer period.

Figure  4.  Synchrotron radiation measured reflectivity data (open circles) at NSRF for different location x along the mirror with interval 5 mm and their fitted curves (lines) by IMD software

DownLoad: Full-Size Img PowerPoint

The measured synchrotron radiation reflectance curves were fitted using the IMD software to determine the layer thickness^[17]. The fitting results are also listed in Table 1. By comparison, it can be noticed that the multilayer period thicknesses calculated from SR and GIXRR measurements are different (offset between 0 and 0.16 nm). This can be explained as that the measured locations were not strictly the same for two different measurements.

3. Conclusion

In summary, linearly lateral graded [Mo/B₄C]₆₀ multilayer sample was deposited by DC magnetron sputtering method on silicon substrate. GIXRR analysis demonstrates that the sample has a period thickness (D) varying linearly from 4.39 nm to 7.82 nm with an average gradient of 0.054 nm/mm in the horizontal direction. EUV reflectance measurement indicates that the multilayer achieves high reflectivity over a wavelength range of 6.8-11.0 nm. Moreover, FWHM of the reflectance peaks varies from 0.13 nm to 0.31 nm with the increase of multilayer period thicknesses, indicating a high spectral resolution for the multilayer sample.