Open Access
Add to BibSonomyAbstract
Reactive sputtering with a mixture of argon and nitrogen (N2 partial pressure of 4%, 8%, and 15%) as the working gas is used to develop the high reflectance Pd/B4C multilayers for soft X-ray region application. Compared to the pure Ar fabricated sample, the interface roughness of the nitridated multilayer is slightly increased while the compressive stress is essentially relaxed from -623 MPa (pure Ar) to -85 MPa (15% N2). A maximum reflectance of 32% is measured at the wavelength of 9.5 nm for the multilayer fabricated with 15% N2. After storing the multilayers in an air environment for 6–17 months, a distinct aging effect is observed on the nitridated samples. The transmission electron microscopy results indicate that a large part of the top layers of the nitridated samples is deteriorated with severe interdiffusion, essential decrease in d-spacing, and compacted multilayer structure. The deterioration is less pronounced for the multilayers fabricated with a higher ratio of N2. Energy dispersive X-ray spectroscopy reveals that the concentration of nitrogen and boron in the degraded area is much reduced compared to the intact layers. A primitive model of upward diffusion of nitrogen and boron is proposed to explain the aging effects of the nitridated structure.
© 2017 Optical Society of America
1. Introduction
Reactive sputtering with the addition of nitrogen to the sputtering gas is a widely used technique to fabricate nitride coatings and multilayer structures for mechanical and optical applications. Besides the widely studied transition metal nitrides to improve the hardness and corrosion/wear resistance of different mechanical components [1,2], the multilayer mirrors for extreme ultraviolet (EUV) and soft X-ray (SXR) optics is another notable application. These multilayers usually consist of a periodic layer structure with a bilayer thickness (d-spacing) of only few nanometers. Based on the constructive interference of waves reflected from different interfaces, a high reflectance can be achieved at the specific EUV or SXR wavelength [3]. This makes the multilayer widely applied in the astronomical telescopes, synchrotron radiation and free electron laser facilities, and hot dense plasma diagnostics. Interface imperfections of the nanoscale multilayer, including roughness and interdiffusion, are the most crucial factors to achieve the high reflectance. To improve the interface structure, reactive sputtering with nitrogen can be used. It has been successfully applied to reduce the interface roughness of multilayer systems, like Al/SiC [4], Co/C [5] and W/B4C [6] for developing related EUV and SXR mirrors. It can also significantly suppress the interdiffusion between neighboring layers as demonstrated in the Cr/Sc [7], La/B [8], and Pd/Y [9] multilayers, due to the formation of stable metal nitrides. If the optical constants of the layers were not severely changed, a higher reflectance can thus be generated compared to the multilayer fabricated by the traditional Ar sputtering. Moreover, the nitridation of layers can reduce the intrinsic stress of the multilayer and avoid the deformation of substrates or film delamination [5,6,10]. The thermal stability of the nitridated multilayer can also be improved, which is important for the mirrors working under high heat load [11,12]. Given to these advantages, reactive sputtering with nitrogen has become an effective method to improve the performance and extend the limits of conventional multilayer optics.
Pd/B4C is an important multilayer structure given to its applications for normal incidence SXR mirrors at the waveband of 6.7 −12 nm [13], and for grazing incidence hard X-ray monochromators [14]. Although a maximum SXR reflectance of 43% was achieved at 9.1 nm, it is much lower than the theoretical value of 60% [9] due to mainly the relatively large interface width existing in the structure. The Pd/B4C multilayer also suffers from a large compressive stress of the coatings [15]. To overcome these limitations, we studied the Pd/B4C multilayers fabricated by reactive sputtering with a mixed working gas of argon and nitrogen. Windt and Gullikson recently reported their results on studying the nitridated Pd/B4C multilayers, while the measured reflectance was very low, below 10% at the wavelengths of around 8.4-8.7 nm [15].
In the present paper, we demonstrate that adding 15% nitrogen (partial pressure) to the working gas (Ar) can generate a relatively high reflectance of 32% with a small compressive stress of only −85 MPa as compared with −623 MPa for the multilayer fabricated with pure Ar. More importantly, a distinct temporal change of the nitridated multilayer structure is observed. A large part of the surface layers is deteriorated with severe interdiffusion and essential compaction of the layers, while the deterioration is less pronounced with a higher nitrogen concentration.
2. Fabrication and characterization of the Pd/B4C multilayers (as-deposited)
The Pd/B4C multilayers were fabricated by the direct current magnetron sputtering technique. The base pressure before deposition was 5.0 × 10−5 Pa. Both high purity argon (99.999%) and a mixture of argon and nitrogen (99.999%) were used as the working gas. The ratio of nitrogen (4%, 8%, 15%) is defined as the partial pressure of N2 to the total pressure in the gas cylinder. The working pressure during deposition was 0.1 Pa. During the deposition process, the same working gas was applied for the growth of both Pd and B4C layers. The multilayers were deposited on both superpolished silicon wafers (~0.2 nm RMS roughness) and circular quartz substrates of 1.0 mm thickness and 30 mm diameter. The former one is for the internal layer structure and reflectivity characterization, and the latter one is for the stress measurements.
The Pd/B4C multilayers were designed for operation at around 9.5 nm wavelength at the near normal incidence (5° incidence angle). The period (d-spacing) of the multilayers was d = 5.0 nm with the thickness of Pd layers of about 2.5 nm, i.e. the thickness ratio dPd/d was close to 0.5 for all samples studied. The number of bilayers was N = 80 to provide the maximum reflectance. A capping layer of Pd with around 2.5 nm thickness was placed on the top of all samples. Four different multilayer samples were fabricated by using pure argon, and a mixture of argon and nitrogen with three mixing ratios of 4%, 8%, and 15% nitrogen as the working gas.
The multilayer structures were first characterized using grazing incidence X-ray reflectometry (GIXR) and X-ray scattering measurement at the Cu-Kα emission line (λ = 0.154 nm). The reflectivity curves are shown in Fig. 1 showing that the periods of all multilayers are similar to each other, with slight variation for the different samples. For the multilayer fabricated in pure argon (Fig. 1(a)), up to the 7th order Bragg peak is observed that indicates relatively sharp interfaces and small interfacial roughness. After adding 4%-15% nitrogen to the working gas, the layer structures are deteriorated and only 5 order Bragg peaks are observed. The high order Bragg peaks disappearing could be caused by either increase in the interfacial roughness or decrease in the amplitude of corresponding Fourier harmonics in the dielectric constant depth-distribution. The last could be initiated by several reasons including variation in the interlayers thickness and/or the thickness ratio as well as possible changes in the layers density after nitridation. To decide between these factors we performed analysis of X-ray scattering, because just interfacial roughness is responsible for its appearance.
Fig. 1 X-ray reflectivity (λ = 0.154 nm) versus the grazing angle of the as-deposited Pd/B4C multilayers (thick grey lines), the same multilayers after 6 months storage (thin red lines), and 17 months storage (dashed blue lines). Four samples were fabricated by magnetron sputtering using the working gas of pure Ar (a) and a mixture of Ar and N2, where the content of nitrogen (its partial pressure) was 4% (b), 8% (c) and 15% (d).
The scattering measurements were performed using the detector scan mode at the fixed incidence angle corresponding to the 1st Bragg peak angular position, which is slightly different for different samples. The detector scans of all samples are shown in Fig. 2. The non-specular scattered intensity can be used to compare the interface roughness of the different multilayers. The curves were shifted slightly along the angle-axis for better comparison. Here, the scattering angle is counted to the sample surface. The small peaks at the 2.6-2.7° scattering angle correspond to the resonant non-specular scattering (RNS) arising due to interference of waves scattered by conformal interfacial roughness component while the wings between the specular and RNS peaks are caused mainly by the scattering from nonconformal interfacial roughness [16,17]. As shown in Fig. 2, for all nitridated samples, the average normalized intensity of scattering wings with the scattering angle of 0.2°-0.75° and 1.3°-2.5° (indicated by arrows) is two to five times larger than that of the pure Ar sample. We thus can conclude that the non-conformal interface roughness of all multilayers fabricated with adding of nitrogen is essentially enhanced as compared to the mirror fabricated with pure Ar and this is the main reason of high order Bragg peaks disappearing in Fig. 1. As will be shown in the TEM measurements in the following section, the non-conformal roughness growth is caused mainly by a random polycrystallization of the layers during nitridation, the nano-grains being distributed stochastically in different layers. At the same time, the conformal roughness is less affected by the nitridation process as the RNS peaks are still persist in Fig. 2.
Fig. 2 X-ray scattering measurements (λ = 0.154 nm) of the four as-deposited Pd/B4C multilayers, fabricated with pure Ar and different ratio of mixture of Ar and N2.
The internal stress of the multilayer structures was determined by the sample-curvature measurement. The Fizeau interferometer was used to measure the surface contour and the radius of curvature of the substrates before and after deposition. After that, the stress was calculated according to the Stoney equation [18]. The stress measurement results are shown in Fig. 3. The multilayer fabricated with pure Ar exhibits a large compressive stress of −623 MPa. After adding 4% or 8% N2 in the working gas, the stress is significantly reduced to around −160 MPa. With a higher ratio of N2 of 15%, the stress further decreases down to −85 MPa. The trend of the stress reduction of the nitridated Pd/B4C multilayer is similar to the recently reported results of Windt et al. [15]. However, the multilayer stress remains at a compressive state in our experiments, instead of a reversed tensile state as shown in [15].
Fig. 3 Measured stress as a function of the content ratio of N2 gas for the fabricated Pd/B4C multilayers.
The soft X-ray reflectivity measurements were performed 20 days after the fabrication at the Bending magnet for Emission Absorption and Reflectivity (BEAR) beamline at the ELETTRA synchrotron [19]. All samples were measured at an incidence angle of 5 degree (off normal). The results are presented in Fig. 4. The peak reflectivity of the mirror fabricated with pure Ar achieves 45% (Fig. 4(a)). This is similar to the result reported by Windt et al., 43% at λ = 9.1nm [13]. Adding 4% nitrogen to the working gas results in a dramatic decrease in the reflectivity down to about 18% (Fig. 4(b)), while a further increase in the nitrogen ratio causes a certain enhancement of the reflectivity to 28% with 8% nitrogen (Fig. 4(c)) and 32% with 15% nitrogen ratio in the working gas (Fig. 4(d)). The dramatic decrease of reflectivity after nitridation and the relative enhancement with a higher ratio of N2 were also observed in [15]. Nevertheless, the reported reflectivity of the nitridated multilayer in [15] was below 10%. Given to the essentially relaxed stress, the 32% reflectivity can be of practical interest for applications and a higher reflectivity may be expected using a higher ratio of nitrogen. The general drop of the reflectivity after nitridation can be caused by two reasons: 1) the increased interface roughness as indicated in Fig. 2; 2) the enhanced absorption from the incorporated nitrogen impurity and the possibly formed boron nitrides in the B4C layers (the absorption coefficient of BN is ~70% larger than B4C at the wavelength of 9.5 nm). It is noted that the reflectivity oscillations at the short wavelength side of the Bragg peak are slightly enhanced for the multilayers fabricated with a smaller ratio of N2 (inset of Fig. 4(b)). This may imply a different structure existing inside the multilayer which can also be related to the reduced reflectivity, and it will be further discussed in the next section.
Fig. 4 Spectral dependence of the SXR reflectivity of Pd/B4C multilayers measured promptly after deposition (solid spheres) and after six months keeping in air (open spheres), fabricated by pure Ar (a), Ar + 4%N2 (b), Ar + 8%N2 (c), Ar + 15%N2 (d).
3. Characterization of the Pd/B4C multilayers after 6 months aging
The samples were kept in air environment with ~20% humidity at room temperature for 6 months. Then both the hard X-ray (HXR) and SXR reflectivities were re-measured to analyze the temporal stability of Pd/B4C multilayer structures. The results are presented in Fig. 1 (thin red curves) and Fig. 4.
Figure 1(a) demonstrates that the HXR reflectivity of the mirror fabricated with pure Ar coincides almost completely with the reflectivity measured immediately after deposition. The peak value of the SXR reflectivity is even slightly increased after 6 months aging with a little shifted peak position (Fig. 4(a)). Given to the small incident beam size of the SXR measurement (< 1x1mm2), this can be caused by the different measurement position on the sample and a slight non-uniformity of the multilayer period over the sample surface. Nevertheless, a good temporal stability is inherent in the structure fabricated with pure Ar.
For the mirror fabricated with 15% nitrogen working gas, both HXR and SXR reflectivity curves are changed insignificantly after 6 months aging, while the reflectivity of the Bragg peaks in Fig. 1 are slightly decreased and the position is shifted to the large angle in contrary to the as-deposited sample. The peak value of the SXR reflectivity in Fig. 4(d) decreases by about 2%.
However, the reflectivity curves are dramatically deformed for the aged multilayers fabricated with adding 4% and 8% nitrogen to the working gas. Splitting of the Bragg peaks is clearly observed in the HXR curves including the first order Bragg peak, which is only slightly affected by imperfections of internal structure of multilayer, as a rule. The newly splitted peaks appear at the large angle side of the original Bragg peak indicating that part of the layers were changed with smaller d-spacings. Moreover, the height of the first order Bragg peaks decreases significantly. The peak value of the SXR reflectivity also decreases by 5%-8% compared to the as-deposited cases. In addition, a broad sidelobe is observed at the short wavelength side of the SXR reflectivity curve of multilayer fabricated with 4% nitrogen (Fig. 4(b)). The sidelobe of the 8% doped sample is also enhanced compared to the as-deposited state (Fig. 4(c)). These facts point clearly to a drastic deterioration of the internal structure of multilayers fabricated with 4% and 8% nitrogen.
Transmission electron microscopy (TEM) was used to characterize thoroughly the changes occurred in the internal structure of the multilayers after 6 months aging. The selected area electron diffraction (SAED) was used to study the crystallization state. The TEM measurements were performed with an FEI Tecnai G2 F20 equipment in Materials Analysis Technology Inc. The samples were prepared with the focused ion beam technique. The TEM images are shown in Fig. 5, where the Pd layers appear as dark, and the B4C layers as bright. Figure 5 consists of three columns corresponding to different composition of the working gas: pure Ar, Ar + 4%N2, and Ar + 15%N2. The first row (A) shows TEM images of the whole cross-section of multilayers, row (B) shows the TEM images with higher resolution, row (C) shows SAED patterns of the samples, and row (D) displays the surface area of the samples. Notice that high resolution images and SAED patterns of both the top and the bottom parts of the multilayer stack are presented for the sample fabricated with 4% nitrogen.
Fig. 5 TEM and electron diffraction images of the multilayer fabricated by pure Ar (A0, B0, C0, D0), 4% N2 (A4, B4, C4, D4), and 15% N2 (A15, B15, C15, D15). For the 4% N2 sample, Figs. (B4 top, C4 top), (B4 mid), and (B4 bottom, C4 bot.) correspond to the top, middle and bottom part of the stack.
Image (A0) demonstrates that all 80 bilayers remain intact after 6 months aging for the multilayer fabricated with pure Ar. Only several layers were found to be slightly delaminated (pointed by the arrows) that can be caused by the large compressive stress inherent to this multilayer. The high-resolution image B0 shows that the interfaces between layers are sharp and smooth demonstrating a high quality of the structure. The measured layer thickness of Pd and B4C proved to be 2.6 nm and 2.5 nm, respectively.
However, image (A4), where the internal structure of the multilayer fabricated with 4% nitrogen in the working gas is shown, demonstrates unexpected result: uppermost part of the structure containing about 50-60 bi-layers was totally destroyed during 6 months aging, while the lowermost part looks intact. The structure of the top, middle transition area, and bottom layers can be better seen in the high-resolution images (B4). The light element layers are barely perceptible and the bi-layer thickness in the top part of the stack decreases down to around 3.7 nm (image (B4) top) due to essential decrease in the light element layers thickness. The periodicity of layers in the top stack is deteriorated. This fact allows to explain the splitting peaks observed in the GIXR curves. Due to decreasing thickness of layers in the top stack, the total thickness of multilayer structure was compressed by about 12% (comparing images (A4) and (A0), presented in the same scale). The layer image in the middle transition area between the degraded and preserved structure (image (B4) mid) implies a severe diffusion of light elements which caused the compaction of the layer structure. A tendency of upward diffusion is observed. The bottom layers display a periodic structure with the prescribed d-spacing of 5.1 nm, while the interfaces width seems to be a little larger as compared with that in image (B0). In addition, polycrystalline grains were observed in the layers.
The TEM images of the multilayer fabricated with adding 15% N2 to the working gas are shown in Fig. 5, right column (A15-B15). Only a small part at the top of the stack, containing about 15 bilayers, is changed after 6 months. The deterioration is not as pronounced as for the 4% N2 sample, which is in agreement with the GIXR results. The high-resolution image of the bottom layers (B15) exhibits a well-ordered layer structure with 5.2 nm d-spacing. An obvious polycrystallization of the layers is observed that increases interfacial roughness as compared with the sample deposited with pure Ar (image (B0)). The relatively large interface roughness of the nitridated multilayer agrees with the scattering measurement results (Fig. 2).
The electron diffraction images of the samples studied are shown in Fig. 5, row C. Owing to the large number of different possible borides and nitrides that can be formed, it is difficult to identify uniquely the exact structure of each diffraction pattern. Nevertheless, the diffraction features are numbered from (1) to (6), based on the calculated d-spacings, to compare the crystallization changes. The multilayer deposited with pure Ar displays a broadened polycrystalline ring (1) (image (C0)), which is identified as Pd (111) [15]. After introducing 15% N2 to the working gas, four diffraction rings, (1) to (4), are observed indicating an enhanced crystallization degree (image (C15)). Besides the palladium phase (1), the same as found in the pure Ar sample, new phases of boron nitride and, maybe, palladium boride are formed (rings (2) - (4)). For the sample deposited with adding 4% N2, nano-beam diffraction was used to measure separately the patterns at the top (image (C4) top) and bottom (image (C4) bottom) of the multilayer stack. The preserved bottom layers exhibit similar crystalline structures as the 15% N2 sample, except the missed phase (2). Two new phases of (5) and (6) are formed in the deteriorated top layers as compared to the bottom part. The 6 months aging is thus accompanied by the formation of new polycrystalline structures.
The TEM images of the near surface areas of the different samples were shown in Fig. 5, row D. The surface of the sample fabricated with pure Ar was found to be flat and sharp (image (D0)). However, a rough diffusion area containing polycrystalline grains was formed at the surface of the samples fabricated with 4% and 15% of nitrogen (image (D4) and (D15)). The thickness of the diffusion area is only 3-4 nm. This also implies a tendency of upward diffusion while part of the atoms may escape the multilayer stack. Otherwise, the surface diffusion layer would be much thicker than presently observed.
4. Pd/B4C multilayers after 17 months aging
The multilayer samples were further monitored for 17 months after deposition. The GIXR results show that the structure of the pure Ar sample is still almost the same as the as-deposited state (Fig. 1(a), dashed line). The 4% N2 sample was fully deteriorated with an even smaller d-spacing compared to the structure of 6-months aging (Fig. 1(b) dashed line). With the higher ratio of N2, the further deterioration is also less pronounced and the 15% N2 sample is only a little changed compared to 6 months (Fig. 1(d) dashed line). To better study the structural deterioration, the elemental distribution in the intact and degraded area of the 15% N2 sample was measured using the energy dispersive X-ray spectroscopy (EDX) and electron energy loss spectroscopy (EELS) analysis in the TEM set up. Several line scans were performed continuously from the intact layers to the surface degraded area of the multilayer to study the relative changes. Figure 6(a) is the TEM image of the near surface area of the 15% N2 sample and the number of the degraded bilayers is similar to 6 months. Figure 6(b) shows an EDX line scan performed along the red line area indicated in Fig. 6(a). Given to the lack of standard calibration sample, the absolute concentration of different elements is not accurate. Nevertheless, the relative changes of the content of each element are reliable. In the good layer structure, clear periodic oscillations of Pd, N, and B profiles are observed, where the depth distributions of N and B are the same and the one of Pd is opposite. Although the oscillation of C is hard to recognize in this EDX result, it is clearly observed in another EELS line scan which confirms the same depth distributions of B and C. It is clear that nitrogen is mainly incorporated in the B4C layers rather than in palladium. This implies that nitrogen is easier to react with boron and carbon atoms [20]. A similar result was observed in the nitridated Co/C [21] and W/B4C [6] multilayers, where nitrogen was mainly contained in the carbon and boron carbide layers, instead in Co or W layers. A small amount of oxygen is also detected in the layers which displays a similar oscillation as boron. It means that the oxygen impurity is mostly contained in the B4C layers, probably due to the very negative enthalpy of formation of B2O3 [20]. However, in the degraded layers, the periodic oscillation of different elements disappears due to the severe inter-diffusion. The content of nitrogen and boron decrease significantly compared to the good layers, while carbon, palladium, and oxygen remain similar. This trend is confirmed by both EDX and EELS results. It is indicated that nitrogen and boron are the main diffusing elements in the deterioration process, most probably diffusing upwards. This process depletes the nitrogen and boron atoms in the near surface area and results in the much thinner light element layers and the total compaction of the multilayer.
Fig. 6 TEM image (a) and EDX line scan (b) of the surface area of the multilayer fabricated with 15% N2 after 17 months storage.
5. Discussion
The temporal stability of Pd/B4C multilayers fabricated in pure Ar environment is recently reported by Morawe et al [22, 23]. Surface deterioration of the multilayer with d-spacing below 4 nm is observed and the degraded area also exhibits a smaller d-spacing as compared to the normal layers. This is attributed to the significant oxidation of the surface B4C layers forming boron oxide particles on the surface and depleting boron in the particular system of Pd-B4C [22]. In our experiments, all multilayers have the d-spacing of 5 nm and the sample fabricated by pure Ar is very stable. This is consistent with the reported result. However, the nitridated samples exhibit a temporal deterioration. Based on the SXR, GIXR, TEM and EDX measurements, we propose a primitive model to explain this distinct phenomenon.
During the reactive sputtering process, nitrogen atoms were mainly incorporated in the B4C layers forming chemical B-N and C-N bonds. Due to the relatively low concentration of nitrogen and the bombardment of the deposited particles, the nitrides are largely non-stoichiometric and contain defects. The thermodynamic non-equilibrium and relative high concentration of defects can provide the driving forces for the diffusion of nitrogen and boron atoms and make the multilayer less stable than the pure Ar fabricated sample [24]. It was also observed in the nitridated Cr/Sc multilayer that the multilayer evolved into two parts with slightly different periods after annealing at 330 °C [11]. This was explained by the redistribution of N and recrystallization during annealing. On the other hand, oxygen can diffuse into the surface layers from air and react with boron atoms given to the more negative enthalpy of formation of e.g. B2O3 than BN [20]. These two reasons make the deterioration first occurred in the surface layers and part of the nitrogen and boron atoms in the surface layers escape from the multilayer that created the surface diffusion area. This resulted in the low concentration of boron and nitrogen in the degraded area and the layer compaction. The escape of nitrogen is understandable still it is volatile. For the escaped boron atoms, one of the possible explanations is that they diffuse and aggregate on the multilayer surface and form boron oxide particles with relatively large size as observed earlier in [22]. It is interesting that no enrichment of oxygen was observed in the degraded layers that is consistent with the results reported in [22]. Probably, the oxidation only occurs at the very surface area of the multilayer which may accelerate the upward diffusion of boron and nitrogen in the top part of the stack. Due to the limited surface oxidation, the bottom part of the multilayer structure remains intact. This deterioration may start quite soon after the fabrication, since the SXR measurement after 20 days already shows a small sidelobe beside the Bragg peak of the sample fabricated with 4% N2 (Fig. 4(b)). As the ratio of nitrogen used in the sputtering process increases, more nitrogen was incorporated in the layers. The composition of the formed nitrides was closer to the stoichiometric value and the diffusion coefficient of nitrogen decreased with the higher concentration [24,25]. This makes the multilayer structure more stable. Thus, the surface degradation of layers was less pronounced with the larger ratio of N2 used.
For the Pd/B4C multilayer fabricated with 15% nitrogen, a reflectance of 30% was still obtained at 9.4 nm, despite the ~15 bilayers deteriorated at the surface. If a larger N2 ratio of above 15% is used in the sputtering process, a better stability and higher reflectance may be expected. Moreover, the multilayers in this experiment all terminate with Pd at the surface which was also used as the capping layer. However, the Pd/B4C structure fabricated in pure Ar environment with Pd at the surface was recently proved to be most unstable, compared to the other capping layer materials like B4C, Al2O3, and Si [22]. Thus, the temporal stability may be further improved with other capping layers.
6. Summary
Soft X-ray Pd/B4C multilayers were fabricated by reactive sputtering technique with different working gases: pure Ar and a mixture of argon and nitrogen at different partial pressure of N2. (4%, 8%, and 15%). The nitridated multilayers demonstrate an essential decrease in the compressive stress as compared to the sample fabricated with pure Ar. At the same time, the interfacial roughness increases slightly due to an enhanced crystallization of the layers. Moreover, an unexpected physical phenomenon was observed in the nitridated samples: severe deterioration and compaction of the upper part of the multilayer structures occurs after 6-months storage, while the lower part of layers is still retained. It is implied that the temporal stability of the nitridated Pd/B4C multilayers is poorer than the ones fabricated with pure Ar. The deterioration was less pronounced for the larger ratio of nitrogen (15%) used in the sputtering process. An upward diffusion of nitrogen and boron from the bottom to the top of the stack was proposed as the reason for the poor temporal stability, according to the structural characterization results. Nevertheless, this model is primitive, more experiments and characterization are needed to explain the origins and detailed process for the deterioration. Given to the present results, a larger amount of N2 (>15%) or different capping layers may provide a better temporal stability for the case of nitridated Pd/B4C multilayers. The observed temporal changes of the nitridated multilayer structure can provide useful guidance for the development of EUV and X-ray multilayer mirrors and also other related thin film devices fabricated by reactive sputtering with nitrogen.
Funding
National Natural Science Foundation of China (NSFC) (11443007, 11505129); National Key Scientific Instrument and Equipment Development Project (2012YQ13012505, 2012YQ24026402); Shanghai Pujiang Program (15PJ1408000).
Acknowledgments
The SXR reflectivity measurements were carried out at the ELETTRA synchrotron in the framework of Proposal No. 20150417.
References and links
1. S. Wang, D. Antonio, X. Yu, J. Zhang, A. L. Cornelius, D. He, and Y. Zhao, “The hardest superconducting metal nitride,” Sci. Rep. 5(1), 13733 (2015). [CrossRef] [PubMed]
2. H. C. Barshilia, B. Deepthi, N. Selvakumar, A. Jain, and K. S. Rajam, “Nanolayered multilayer coatings of CrN/CrAlN prepared by reactive DC magnetron sputtering,” Appl. Surf. Sci. 253(11), 5076–5083 (2007). [CrossRef]
3. E. Louis, A. E. Yakshin, T. Tsarfati, and F. Bijkerk, “Nanometer interface and materials control for multilayer EUV-optical applications,” Prog. Surf. Sci. 86(11-12), 255–294 (2011). [CrossRef]
4. D. L. Windt and J. A. Bellotti, “Performance, structure, and stability of SiC/Al multilayer films for extreme ultraviolet applications,” Appl. Opt. 48(26), 4932–4941 (2009). [CrossRef] [PubMed]
5. J. A. Bellotti and D. L. Windt, “Depth-graded Co/C multilayers prepared by reactive sputtering,” Proc. SPIE 7437, 743715 (2009). [CrossRef]
6. D. L. Windt, “Reduction of stress and roughness by reactive sputtering in W/B4C X-ray multilayer films,” Proc. SPIE 6688, 66880R (2007). [CrossRef]
7. N. Ghafoor, F. Eriksson, E. Gullikson, L. Hultman, and J. Birch, “Incorporation of nitrogen in Cr/Sc multilayers giving improved soft x-ray reflectivity,” Appl. Phys. Lett. 92(9), 091913 (2008). [CrossRef]
8. D. S. Kuznetsov, A. E. Yakshin, J. M. Sturm, R. W. van de Kruijs, E. Louis, and F. Bijkerk, “High-reflectance La/B-based multilayer mirror for 6.x nm wavelength,” Opt. Lett. 40(16), 3778–3781 (2015). [CrossRef] [PubMed]
9. D. Xu, Q. Huang, Y. Wang, P. Li, M. Wen, P. Jonnard, A. Giglia, I. V. Kozhevnikov, K. Wang, Z. Zhang, and Z. Wang, “Enhancement of soft X-ray reflectivity and interface stability in nitridated Pd/Y multilayer mirrors,” Opt. Express 23(26), 33018–33026 (2015). [CrossRef] [PubMed]
10. R. Conley, N. Bouet, J. Zhou, H. Yan, Y. Chu, K. Lauer, J. Miller, L. Chu, and N. Jahedi, “Advanced multilayer Laue lens fabrication at NSLS-II,” Proc. SPIE 8502, 850202 (2012). [CrossRef]
11. F. Eriksson, N. Ghafoor, L. Hultman, and J. Birch, “Reflectivity and structural evolution of Cr/Sc and nitrogen containing Cr/Sc multilayers during thermal annealing,” J. Appl. Phys. 104(6), 063516 (2008). [CrossRef]
12. P. Emma, R. Akre, J. Arthur, R. Bionta, C. Bostedt, J. Bozek, A. Brachmann, P. Bucksbaum, R. Coffee, F.-J. Decker, Y. Ding, D. Dowell, S. Edstrom, A. Fisher, J. Frisch, S. Gilevich, J. Hastings, G. Hays, P. Hering, Z. Huang, R. Iverson, H. Loos, M. Messerschmidt, A. Miahnahri, S. Moeller, H.-D. Nuhn, G. Pile, D. Ratner, J. Rzepiela, D. Schultz, T. Smith, P. Stefan, H. Tompkins, J. Turner, J. Welch, W. White, J. Wu, G. Yocky, and J. Galayda, “First lasing and operation of an ångstrom-wavelength free-electron laser,” Nat. Photonics 4(9), 641–647 (2010). [CrossRef]
13. A. J. Corso, P. Zuppella, D. L. Windt, M. Zangrando, and M. G. Pelizzo, “Extreme ultraviolet multilayer for the FERMI@Elettra free electron laser beam transport system,” Opt. Express 20(7), 8006–8014 (2012). [CrossRef] [PubMed]
14. A. Rack, T. Weitkamp, M. Riotte, D. Grigoriev, T. Rack, L. Helfen, T. Baumbach, R. Dietsch, T. Holz, M. Krämer, F. Siewert, M. Meduna, P. Cloetens, and E. Ziegler, “Comparative study of multilayers used in monochromators for synchrotron-based coherent hard X-ray imaging,” J. Synchrotron Radiat. 17(4), 496–510 (2010). [CrossRef] [PubMed]
15. D. L. Windt and E. M. Gullikson, “Pd/B4C/Y multilayer coatings for extreme ultraviolet applications near 10 nm wavelength,” Appl. Opt. 54(18), 5850–5860 (2015). [CrossRef] [PubMed]
16. D. G. Stearns and E. M. Gullikson, “Nonspecular scattering from extreme ultraviolet multilayer coatings,” Physica B 283(1-3), 84–91 (2000). [CrossRef]
17. I. V. Kozhevnikov, “Analysis of X-ray scattering from a rough multilayer mirror in the first order perturbation theory,” Nucl. Instrum. Methods A 498(1-3), 482–495 (2003). [CrossRef]
18. L. B. Freund and S. Suresh, Thin Film Materials-Stress, Defect Formation and Surface Evolution (Cambridge University, 2003).
19. S. Nannarone, F. Borgatti, A. DeLuisa, B. P. Doyle, G. C. Gazzadi, A. Giglia, P. Finetti, N. Mahne, L. Pasquali, M. Pedio, G. Selvaggi, G. Naletto, M. G. Pelizzo, and G. Tondello, “The BEAR beamline at Elettra,” AIP Conf. Proc. 705, 450–453 (2004). [CrossRef]
20. C. Montcalm, P. A. Kearney, J. M. Slaughter, B. T. Sullivan, M. Chaker, H. Pépin, and C. M. Falco, “Survey of Ti-, B-, and Y-based soft x-ray-extreme ultraviolet multilayer mirrors for the 2- to 12-nm wavelength region,” Appl. Opt. 35(25), 5134–5147 (1996). [CrossRef] [PubMed]
21. M. Wen, Q. Huang, S. Ma, W. Li, R. She, J. Peng, A. Giglia, I. V. Kozhevnikov, H. Feng, Z. Zhang, and Z. Wang, “Improvement of interface structure and polarization performance of Co/C multilayers by incorporation of nitrogen,” Opt. Express 24(24), 27166–27176 (2016). [CrossRef] [PubMed]
22. C. Morawe, R. Supruangnet, and J. C. Peffen, “Structural modifications in Pd/B4C multilayers for X-ray optical applications,” Thin Solid Films 588, 1–10 (2015). [CrossRef]
23. R. Supruangnet, C. Morawe, J. C. Peffen, H. Nakajima, S. Rattanasuporn, P. Photongkam, N. Jearanaikoon, and W. Busayaporn, “Chemical modification of B4C cap layers on Pd/B4C multilayers,” Appl. Surf. Sci. 367, 347–353 (2016). [CrossRef]
24. L. Hultman, “Thermal stability of nitride thin films,” Vacuum 57(1), 1–30 (2000). [CrossRef]
25. T. Moskalioviene and A. Galdikas, “Stress induced and concentration dependent diffusion of nitrogen in plasma nitrided austenitic stainless steel,” Vacuum 86(10), 1552–1557 (2012). [CrossRef]
