short communications\(\def\hfill{\hskip 5em}\def\hfil{\hskip 3em}\def\eqno#1{\hfil {#1}}\)

Journal logoJOURNAL OF
SYNCHROTRON
RADIATION
ISSN: 1600-5775

New transmission-type X-ray filters consisting of amorphous multilayer films

aAdvanced Technology Research Laboratories at Futtsu, Nippon Steel Corporation, 20-1 Shintomi, Futtsu, Chiba 293, Japan, and bAdvanced Technology Research Laboratories at Hiyoshi, Nippon Steel Corporation, 3-35-1 Ida, Nakahara-ku, Kawasaki 211, Japan
*Correspondence e-mail: takagi@kaiseki.re.nsc.co.jp

(Received 4 August 1997; accepted 22 January 1998)

New transmission-type X-ray filters have been developed. The filters consist of X-ray-amorphous metal (less than 30 Å)/metalloid (∼10 Å) multilayer films sputter-deposited on X-ray-transparent polymer substrates. Such metal/metalloid multilayer films show only very broad diffraction peaks, since the metal and metalloid layers forming the multilayer films are usually X-ray amorphous if the layers are sufficiently thin. The filters use the wavelength-dependent absorption phenomena near absorption edges of elements to reduce the intensity of transmitted X-rays, without generating any crystalline sharp peaks which cause serious problems in experiments such as fluorescence XAFS measurements. The multilayer-film filters were prepared by a multi-target magnetron sputtering deposition technique, paying special attention to the homogeneity of the layer thickness by spinning substrates of the films. The filters are useful in reducing the intensity of undesirable fluorescence emissions and improving the signal-to-background ratios of data acquired in various measurements using a solid-state detector.

1. Introduction

It is important to improve signal-to-background ratios in order to obtain high-quality X-ray data. Transmission-type X-ray filters which use wavelength-dependent absorption phenomena near absorption edges of elements have been devised and are commonly used (Stern & Heald, 1979[Stern, E. A. & Heald, S. M. (1979). Rev. Sci. Instrum. 50, 1579-1582.]).

Such transmission-type X-ray filters have some common problems. If the filters contain crystalline materials, they do not only absorb X-rays of a certain wavelength region and thus reduce background levels, but also generate diffraction X-ray beams which sometimes add an undesirable background to the spectra. This phenomenon causes particularly serious problems when large solid-angle detectors such as large-area/multiple-element solid-state detectors (SSDs) are used for data acquisition. Under a high-flux synchrotron radiation beam, even small portions of crystalline phases in the filters cause unignorable diffraction intensities. X-ray wavelength scanning, which is an indispensable procedure in XAFS experiments, increases the chance of encountering such problems.

To solve these problems, new transmission-type X-ray filters consisting of X-ray-amorphous multilayer films, from which no sharp crystalline diffraction peaks are observed, have been developed and applied to fluorescence XAFS measurements.

2. General concept of X-ray-amorphous multilayer filters

The X-ray-amorphous multilayer-film filters can be prepared on polymeric film substrates which are almost transparent in the hard X-ray region (5–20 keV). Polyimide films are the most suitable, because of their thermal stability and mechanical toughness (Seymour & Carraher, 1981[Seymour, R. B. & Carraher, C. E. (1981). Polymer Chemistry, p. 451. New York: Marcel Dekker.]). Such multilayer films for X-ray optics are usually prepared by a multi-target sputtering deposition technique.

Generally, sputter-deposited multilayer films show only broad diffraction peaks, if the constituent layers are sufficiently thin, since a new layer starts to deposit onto an older one before the original layer grows thick enough to crystallize (Takagi et al., 1985[Takagi, Y., Flessa, S. A., Hart, K. L., Pawlik, D. A., Kadin, A. M., Wood, J. L., Keem, J. E. & Tyler, J. E. (1985). Proc. SPIE, 563, 66-69. ]). Although the maximum layer thicknesses in which elements can remain in amorphous states depend on the elements, their chemical states and the deposition conditions, the maximum thickness of metal layers is ∼30 Å, while that of metalloids (C, B or Si) is ∼100 Å. Therefore, by inserting a light-element metalloid such as C or Si, which is transparent to hard X-­rays, between thin metal layers, we can obtain a multilayer-film filter consisting of X-ray-amorphous layers. A typical transmission diffraction image from an X-ray-amorphous multilayer film, a Co(18 Å)/C(20 Å) multilayer film, taken by an image plate, is shown in Fig. 1[link]. Only one broad peak, attributable to the first-order peak of the amorphous structure of the metal and metalloid layers, was observed. Another advantage in using transmission-type filters is that any sharp diffraction peaks from the layer structures in the low-angle region, which appear in the reflection geometry, do not appear in the transmission geometry.

[Figure 1]
Figure 1
Typical transmission image-plate diffraction image of the X-ray-amorphous multilayer-film deposited on a polyimide film substrate, Co(18 Å)/C(20 Å) – (1100 + 1100 layers on both sides of the substrate) analysed by BAS 2000 (Fuji Film Co. Ltd) at λ = 1.7 Å.

The absorption mechanism of incident X-rays by X-ray-amorphous multilayer filters is no different to that of the traditional crystalline transmission filters. Absorption efficiency per thickness of the metal layers in the filter will not be much different from that of the traditional crystalline filters, since the densities of the amorphous phases are only slightly lower than those of the crystalline phases of the metals commonly used in the filters.

3. Fabrication technique of X-ray amorphous multilayer-film filters

The desired multilayer films can most easily be prepared by using a multi-target magnetron sputtering deposition technique. However, special attention must be paid to obtaining homogeneous layer thicknesses in order to achieve uniform filter absorption characteristics. The film substrates were spun in the substrate plane at an appropriate spinning rate to compensate for the slow shutter speed, as shown schematically in Fig. 2[link] (Takagi et al., 1995[Takagi, Y., Iijima, I., Akiyama, N. & Yamada, H. (1995). Jpn Patent P01972555. (In Japanese).]). The improvement in the layer homogeneity using this method is shown in Fig. 3[link]. A similar technique has been developed by Kortright et al. (1993[Kortright, J. B., Gullikson, E. M. & Denham, P. E. (1993). Appl. Opt. 32, 6961-6968.]).

[Figure 2]
Figure 2
Substrate-spinning mechanism for homogenizing the layer thickness of the multilayer filter (Takagi et al., 1995[Takagi, Y., Iijima, I., Akiyama, N. & Yamada, H. (1995). Jpn Patent P01972555. (In Japanese).]).
[Figure 3]
Figure 3
Dependence of the layer thickness in an Hf/Fe multilayer film on the position from the spinning centre of the substrate, with and without spinning (10 turns s−1). The bilayer thicknesses of Hf and Fe were estimated from the low-angle diffraction peaks from X-ray diffraction of the layer structure (Takagi et al., 1995[Takagi, Y., Iijima, I., Akiyama, N. & Yamada, H. (1995). Jpn Patent P01972555. (In Japanese).]).

The multilayer films were deposited a few hundred layers at a time alternately on both sides of the polyimide films, in order to avoid warping and peeling of the films due to internal stress during the deposition.

4. Application of the filters in fluorescence XAFS experiments

X-ray-amorphous multilayer-film filters were prepared for use in improving the signal-to-background ratio and decreasing the dead-time of multi-element SSDs in fluorescence XAFS experiments.

The prepared filter was mounted on an acrylic holder in order to fit in front of the aperture of the multi-element SSD used in various types of fluorescence XAFS experiments at PF-BL3A (Mizutani et al., 1994[Mizutani, T., Imafuku, M., Kikuchi, T. & Takagi, Y. (1994). Photon Factory Activity Report 12, p. 28. Photon Factory, Tsukuba 305, Japan.]).

The wavelengths and energies of the absorption edges and the fluorescence transitions referred to in the following sections are listed in Table 1[link].

Table 1
Wavelengths and energies of the absorption edges and fluorescence transitions related to the present study (Weast, 1984[Weast, R. C. (1984). Editor. Handbook of Chemistry and Physics, 64th ed., pp. E133-E180. Florida: CRC.])

Wavelength (Å) Element Designation Energy (keV)
1.0723 Pt LIII Absorption edge 11.562
1.31304 Pt Lα1 LIIIMV 9.4423
1.32432 Pt Lα2 LIIIMIV 9.3618
1.38059 Cu K Absorption edge 8.9803
1.38109 Cu Kβ2 KMIV,V 8.9770
1.392218 Cu Kβ1,3 KMII,III 8.90529
1.48807 Ni K Absorption edge 8.33165
1.540562 Cu Kα1 KLIII 8.04778
1.544390 Cu Kα2 KLII 8.02783
1.60815 Co K Absorption edge 7.70954
1.74346 Fe K Absorption edge 7.11120
1.7442 Fe Kβ5 KMIV,V 7.1081
1.75661 Fe Kβ1,3 KMII,III 7.05798
1.936042 Fe Kα1 KLIII 6.40384
1.939980 Fe Kα2 KLII 6.39084
2.07020 Cr K Absorption edge 5.9888
43.68 C K Absorption edge 0.28383

4.1. Example 1: Cr/C multilayer-film filters to reduce the intensity of Fe K-edge-related fluorescence in Cu Kα fluorescence XAFS measurements

Cu Kα fluorescence XAFS measurements are very useful in clarifying the crystal structures of Cu precipitates (∼1 at.%) in Cu-added steels (Takagi, Okitsu et al., 1993[Takagi, Y., Okitsu, Y., Mizutani, T. & Imafuku, M. (1993). Photon Factory Activity Report 11, p. 35. Photon Factory, Tsukuba 305, Japan.]; Imafuku et al., 1996[Imafuku, M., Mizutani, T., Takagi, Y. & Okitsu, Y. (1996). Proceedings of the International Conference on Microstructures and Functions of Materials (ICMFM 96), edited by N. Igata, Y. Hiki, I. Yoshida & S. Sato, pp. 25-28. Department of Materials Science and Technology, Science University of Tokyo, 2641 Yamazak, Noda 278, Japan.]). In such an experiment, it is important to reduce the intensity of Fe K-edge-related fluorescence transitions from the matrix Fe phase, while keeping the weak Cu Kα fluorescence intensity from the precipitates intact, in order to obtain data with a high signal-to-background ratio. A Cr(15 Å)/C(10 Å) (200 layers) multilayer-film filter was prepared on a polyimide film substrate for this purpose. Energy spectra obtained using the multi-element SSD with and without the filters are shown in Fig. 4[link]. The Cr/C multilayer film absorbs Fe Kα and Kβ fluorescence photons while the intensities of Cu Kα and Kβ fluorescence photons are little changed, since the energy of the Cr K-edge is much closer to the energies of the Fe K-edge-related fluorescence photons than the corresponding Cu K-edge fluorescence photons.

[Figure 4]
Figure 4
Fluorescence energy spectra from the Cu-added steel taken by the multi-element SSD system, with and without the Cr/C multilayer filter, at PF-BL3A.

4.2. Example 2: Co/C multilayer-film filters to reduce the intensity of Cu K-edge-related fluorescence in Pt Lα fluorescence XAFS measurements

Pt LIII-edge fluorescence XAFS measurements are very useful in the clarification of the states of Pt included in the quench-and-melt-growth Y–Ba–Cu oxide superconductors (Takagi, Tanaka et al., 1993[Takagi, Y., Tanaka, M., Mizutani, T., Takebayashi, S. & Imafuku, M. (1993). Photon Factory Activity Report 11, p. 34. Photon Factory, Tsukuba 305, Japan.]; Tanaka et al., 1995[Tanaka, M., Morita, M., Takebayashi, S., Hayashi, S., Takagi, Y. & Hashimoto, M. (1995). Proceedings of the 7th International Symposium on Superconductivity (ISS 94), edited by K. Yamafuji & T. Morishima, pp. 693-696. Berlin: Springer-Verlag. ]). In this experiment, it is very important to reduce Cu Kα fluorescence intensity from the YBa2Cu3O7 and Y2BaCuO5 matrix phases to avoid an increase in the SSD dead-time, while keeping intact the Pt LIII-edge-related fluorescence intensity from the minor phases. For this purpose, a Co(18 Å)/C(20 Å) (200 layers) multilayer-film filter was prepared.

Cu Kα fluorescence photons are absorbed by the Co/C multilayer filter to a much greater extent than Pt Lα photons, since the photon energies of Cu Kα and Kβ fluorescence are only a little larger than that of the Co K-edge, while the energy of Pt Lα is much greater. The dependence of the filter performance on thickness was examined by changing the number of layers of the filter inserted in front of the SSD. The results are summarized in Fig. 5[link]. The intensity ratio of Cu Kα to Pt Lα was reduced more than that of Cu Kβ to Pt Lα, as the wavelength of Cu Kα is closer than that of Cu Kβ to the Co K-edge. Good linearity of the intensity ratios (especially that of Cu Kα to Pt Lα) versus the number of layers verifies our assumption that the C layers are almost transparent for all fluorescence photons in this X-ray wavelength region. This is due to the low energy of the C K-edge compared with the energies of both the Cu K- and Pt LIII-edges. The dead-time was improved by more than 20% compared with an unfiltered case.

[Figure 5]
Figure 5
Dependence of fluorescence intensity ratios I(Cu Kα)/I(Pt Lα) and I(Cu Kβ)/I(Pt Lα) on the number of layers of the Co/C multilayer filter.

5. Conclusions

Although systematic studies of the performance of the transmission-type X-ray-amorphous multilayer-film filters must be carried out, it is clear that the filters are very effective tools in improving the signal-to-background ratios in various X-ray experiments by reducing undesirable background intensity. The easy preparation method and broad range of applicable elements also give powerful advantages over other types of filters. The combination of multiple metal elements as constituent layers will also enable the preparation of band-pass-type amorphous filters.

The filters will also be useful in the detection of very weak diffraction arcs taken by image plates, since an amorphous filter can improve the signal-to-background ratio without adding new diffraction arcs.

Acknowledgements

The technical contributions of Toshiharu Kikuchi of Nippon Steel Corporation in devising and drawing the filter assembly were very helpful in these experiments. We also thank Toshiyuki Mizutani of Nippon Steel Corporation for his contribution to the fluorescence XAFS experiments.

References

First citationImafuku, M., Mizutani, T., Takagi, Y. & Okitsu, Y. (1996). Proceedings of the International Conference on Microstructures and Functions of Materials (ICMFM 96), edited by N. Igata, Y. Hiki, I. Yoshida & S. Sato, pp. 25–28. Department of Materials Science and Technology, Science University of Tokyo, 2641 Yamazak, Noda 278, Japan.
First citationKortright, J. B., Gullikson, E. M. & Denham, P. E. (1993). Appl. Opt. 32, 6961–6968.  CrossRef CAS PubMed
First citationMizutani, T., Imafuku, M., Kikuchi, T. & Takagi, Y. (1994). Photon Factory Activity Report 12, p. 28. Photon Factory, Tsukuba 305, Japan.
First citationSeymour, R. B. & Carraher, C. E. (1981). Polymer Chemistry, p. 451. New York: Marcel Dekker.
First citationStern, E. A. & Heald, S. M. (1979). Rev. Sci. Instrum. 50, 1579–1582.  CrossRef PubMed CAS Web of Science
First citationTakagi, Y., Flessa, S. A., Hart, K. L., Pawlik, D. A., Kadin, A. M., Wood, J. L., Keem, J. E. & Tyler, J. E. (1985). Proc. SPIE, 563, 66–69.   CrossRef CAS
First citationTakagi, Y., Iijima, I., Akiyama, N. & Yamada, H. (1995). Jpn Patent P01972555. (In Japanese).
First citationTakagi, Y., Okitsu, Y., Mizutani, T. & Imafuku, M. (1993). Photon Factory Activity Report 11, p. 35. Photon Factory, Tsukuba 305, Japan.
First citationTakagi, Y., Tanaka, M., Mizutani, T., Takebayashi, S. & Imafuku, M. (1993). Photon Factory Activity Report 11, p. 34. Photon Factory, Tsukuba 305, Japan.
First citationTanaka, M., Morita, M., Takebayashi, S., Hayashi, S., Takagi, Y. & Hashimoto, M. (1995). Proceedings of the 7th International Symposium on Superconductivity (ISS 94), edited by K. Yamafuji & T. Morishima, pp. 693–696. Berlin: Springer-Verlag.
First citationWeast, R. C. (1984). Editor. Handbook of Chemistry and Physics, 64th ed., pp. E133–E180. Florida: CRC.

© International Union of Crystallography. Prior permission is not required to reproduce short quotations, tables and figures from this article, provided the original authors and source are cited. For more information, click here.

Journal logoJOURNAL OF
SYNCHROTRON
RADIATION
ISSN: 1600-5775
Follow J. Synchrotron Rad.
Sign up for e-alerts
Follow J. Synchrotron Rad. on Twitter
Follow us on facebook
Sign up for RSS feeds