Nuclear Instruments and Methods in Physics Research A 470 (2001) 210­214 An extended anomalous fine structure ofX-ray quasi-Bragg diffuse scattering from multilayers V.A. Chernova, N.V. Kovalenkob, S.V. Mytnichenkoc,* a Siberian SR Centre, Budker Institute of Nuclear Physics, 11 Lavrentyev Ave., 630090 Novosibirsk, Russia b Budker Institute of Nuclear Physics, 11 Lavrentyev Ave., 630090 Novosibirsk, Russia c Institute of Solid State Chemistry, 18 Kutateladze Str., 630128, Novosibirsk, Russia Abstract An X-ray quasi-Bragg diffuse scattering anomalous fine structure technique was probed near the absorption Ni K- edge to study the interfacial structure of the Ni/C multilayer deposited by the laser ablation. Like other combinations of the EXAFS and diffraction techniques, this method has a spatial selectivity and was shown qualitatively to provide atomic structural information from the mixed interfacial layers. The possibilities and advantages of this technique are discussed. r 2001 Elsevier Science B.V. All rights reserved. PACS: 68.55.@a; 61.10.kw Keywords: Multilayers; X-ray diffuse scattering; EXAFS-spectroscopy 1. Introduction there are two X-ray structural techniques that can provide the direct atomic structural information The recent vigorous advancement ofphysics and from interfacial layers. chemistry ofthe thin films, multilayers, magnetic EXAFS spectroscopy of the atoms-markers [1]: and semiconductor superlattices was caused by The principle ofthis method is that at the stage of their unique properties principally differing from multilayer growth, the atoms-markers having the those ofbulk materials. There is no doubt that same chemical properties as the constituting atoms these unique properties are immediately associated are embedded at a required depth, then the usual with the surface and interface structures. Of fluorescent EXAFS technique can be used. Never- special interest is the chemical or atomic arrange- theless, this technique has its own evident limita- ment ofinterfaces. Thus, the development ofX- tion. ray structural methods allowing one to obtain this X-ray standing waves [2]: Owing to the fact that information is a very real problem. At present, the yields ofsecondary processes are proportional to the standing wave intensity, one can perform *Corresponding author. Siberian SR Centre, Budker Insti- localized probes ofthe atomic structure. The main tute ofNuclear Physics, 11 Lavrentyev Ave., 630090 Novosi- problem ofthis technique is the drastic modifica- birsk, Russia. Tel.: +7-3832-394013; fax: +7-3832-342163. tion ofthe diffraction (the extinction depth and so E-mail address: s.v.mytnichenko@inp.nsk.su (S.V. Mytnichenko). on) as the photon energy moves through the 0168-9002/01/$ - see front matter r 2001 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 8 - 9 0 0 2 ( 0 1 ) 0 1 0 4 0 - 3 V.A. Chernov et al. / Nuclear Instruments and Methods in Physics Research A 470 (2001) 210­214 211 absorption edge. To some degree, this problem can due to the fact that this system was previously be overcome with the use ofspecial substrates that studied by us in details [1,7­9]. The presence of can generate the standing waves [3] so that the strong intermixing between the metal and carbon diffraction from the studied thin films can be layers was revealed by EXAFS spectroscopy and excluded from consideration. Nevertheless, from X-ray diffraction in low- and high-angle ranges. At the technical standpoint, the last method is not a thickness smaller than about 2 nm, the metal universal. Besides, due to the fact that the standing layers are saturated by carbon and are present wave size is comparable to the multilayer period, in the carbide glass-like phase instead ofthe the ratio between the interfacial and bulk signals is metal one. The interfacial regions have the not large, which also restricts this method. highest carbon concentration. This strong inter- The diffraction Anomalous Fine Structure mixing is specific to the laser ablation technique (DAFS) technique is widely used currently [4]. and is explained by the ballistic effect due to high- This method is based on the fact that the energy- energy atom (ion) bombardment ofthe growing dependence ofthe diffraction intensity near the layer. absorption edge provides a local chemical envir- onment information just as the EXAFS spectrum. In spite ofa more complicated data treatment 2. Experimental compared to the standard EXAFS spectroscopy, this method has a very important advantage: this The Ni/C multilayer was deposited by the laser is the spatial selection. Unlike EXAFS spectra ablation technique on the flat silica wafer with a which are contributed from all specimen atoms, surface roughness of 0.4 nm. The number of DAFS spectra provide information on the atoms bilayers was 30. The other parameters were involved in diffraction. obtained from the usual reflectivity scans: the Like the diffraction, the off-specular resonant multilayer period, L, was 4 nm, and the Ni-rich diffuse scattering has the spatial selectivity also layer comprised approximately 0.4 ofthis period due to the fact that only the ``roughness atoms'' or 1.6 nm. The effective roughness parameter, s, localized at the interfacial layers 2s ¼ 2Os2rough þ was found to be equal to 0.3 nm. According to s2mix thick, where srough is the roughness dispersion our previous studies [9] this value is determined and 2smix is the thickness ofthe mixed layers, are by the presence ofintermixed layers, whereas the involved in this scattering. Thus, the Diffuse true roughness dispersion, srough, is smaller Scattering Anomalous Fine Structure (DSAFS) (B0.1­0.2 nm). spectra contain information on the local chemical The measurements were performed using SR environment ofthe ``roughness atoms''. The main ofthe VEPP-3 storage ring ofthe Siberian purpose ofthis work was to test the DSAFS SR Centre at Budker INP, which operates at technique. 2 GeV and with a maximum stored current of The resonant quasi-Bragg diffuse scattering [5,6] 165 mA. The diffractometer ofthe ``anomalous was used in this study for two reasons. Firstly, this scattering'' station [10] with a primary channel- scattering is caused by the interfacial imperfections cut single-crystal Si(1 1 1) monochromator and coherently repeated from one layer to another the scintillation detector based on an FEU-130 (interfacial cross-correlation). Thus, when the photomultiplier with a NaI(Tl) scintillator incident angle differs greatly from the Bragg angle, were used. The experimental setup is shown in this scattering has a clear kinematic nature. It Fig. 1. simplifies the data treatment considerably. The The energy scan through the Ni K-edge was second reason is the high intensity ofthis performed in such a manner that the diffuse scattering, which allows one to obtain the experi- scattering intensity was always measured at the mental data for a reasonable time. same point in q-space. The standard fluorescent The Ni/C multilayer deposited by the laser EXAFS spectrum was obtained concurrently with ablation technique was chosen as a test sample the basic measurements. 212 V.A. Chernov et al. / Nuclear Instruments and Methods in Physics Research A 470 (2001) 210­214 Fig. 1. The experimental setup: y0, y1 and yB are the incident, scattered and Bragg angles, respectively; E is the photon energy; S1, the primary slit (100 mm) providing an energy resolution of Fig. 3. The shield effect due to strong absorption: mðEÞ is a bulk about 1 eV; S2, the secondary slit (B2 mm) were used to select absorption attenuation coefficient obtained from the fluorescent quasi-Bragg diffuse scattering. The energy scan was performed EXAFS measurements. The roughness cross-correlation was in such a manner that the diffuse scattering intensity was always assumed to be complete. measured at the same point in q-space. Though y0, y1 and yB are changed during this scan, the momentum transfer, q and off-specular angle, o ¼ y02y1 (0.21) were kept constant. Fig. 4. The dependence of wðkÞk3 obtained from the DSAFS spectrum (solid curve) and standard fluorescent EXAFS (points). Fig. 2. The DSAFS spectra obtained: the experimental (lower curve) and corrected data (upper curve). The resulting Fourier transforms of k3-weighted 3. Results and discussion DSAFS and usual fluorescent EXAFS are shown in Fig. 5. The coordination shell positions, that The DSAFS spectrum obtained (Fig. 2) was may be observed using the bulk EXAFS data [1] corrected with regard to the wave extinction due to are depicted as well. The main difference, which strong absorption (Fig. 3). Ignoring the quadratic is seen at a single glance, is the splitting ofthe in the anomalous dispersion correction terms Ni­Ni(I) shell. The origin ofthis splitting can be (their contribution does not exceed a few percent), explained by the wðkÞ behavior (Fig. 4). It is well the corrected experimental data were used to known that the presence, in any shell, oftwo types obtain the oscillation part ofthe imaginary ofneighboring atoms at approximately same anomalous dispersion scattering amplitude, wðkÞ distances in approximately equal amounts cause (Fig. 4) by means ofthe Kramers­Kronig disper- the modulation ofthe fine structure oscillation. In sion relations. Further data treatment was per- other words, the oscillation amplitude alternately formed using the standard methods [11]. decreases and increases, which is seen from Fig. 4. V.A. Chernov et al. / Nuclear Instruments and Methods in Physics Research A 470 (2001) 210­214 213 Fig. 5. The resulting Fourier transforms of k3-weighted DSAFS (solid curve) and standard fluorescent EXAFS (points). Although both curves are shown in Fig. 2 in the scattering intensity falls off drastically in this case. same scale, the amplitude ofDSAFS oscillation is One more application is DSAFS studies based on much smaller compared to the EXAFS one and the superlattice satellite reflections in a high-angle this is not a surprise. Indeed, the structural range. The interfacial stress studies are of interest imperfections in the interfacial layers are at their in this case. maximum. Thus, the results are in a good qualitative agreement with our assumption that the DSAFS 4. Conclusion spectrum provides the structural information on the interfacial layers. As has already been men- The DSAFS method was shown to provide tioned, the thickness ofthe studied interfacial information on the local atomic structure and its layers is determined by the mixed layer thickness disruptions at the interfacial layers of thickness 2s in our case and is about 0.6 nm. in the multilayers. The structural method used in this work can be developed further. For example, using the linear Acknowledgements SR polarization and the corresponding experi- mental geometry ofDSAFS measurements, one We thank the staffs ofVEPP-3, optical work- can obtain the local chemical surrounding not only shops, and SSRC at BINP for their assistance. in the lateral plane, but also normal to it. Another This study was supported by the Russian Founda- possibility is the use ofthe diffuse scattering tion for Fundamental Research, Grant Nos. 99- discrimination in the direction normal to the 02-16671 and 00-02-17624. specular diffraction plane [12]. This allows one to study the local structure at the roughness imper- fections with different spatial ranges. In principal, this method can be used for the single interface or References surface, but one should bear in mind that the diffuse scattering intensity is proportional to N2, [1] V.A. Chernov, N.I. Chkhalo, S.G. Nikitenko, J. Phys. IV 7 (1997) C2­699. where N is the bilayer number. Thus, the diffuse [2] B.W. Batterman, Phys. Rev. 133 (1964) A759. 214 V.A. Chernov et al. / Nuclear Instruments and Methods in Physics Research A 470 (2001) 210­214 [3] V.A. Chernov, N.I. Chkhalo, I.P. Dolbnya, K.V. Zolotar- [9] V.A. Chernov, E.D. Chkhalo, N.V. Kovalenko, S.V. ev, Nucl. Instr. and Meth. A 395 (1995) 175­177. Mytnichenko, Nucl. Instr. and Meth. A 448 (2000) 276. [4] J.O. Cross, Ph.D. Thesis, University ofWashington, 1996. [10] BriefDescription ofthe SR Experimental Station, [5] A.V. Andreev, A.G. Michette, A. Renwick, J. Modern Preprint, INP, 90-92, Novosibirsk, 1990. Opt. 35 (1988) 1667­1687. [11] N. Binsted, J.V. Campbell, S.J. Gurman, P.C. Stephenson, [6] D.G. Stearns, J. Appl. Phys. 71 (1992) 4286­4298. SERC Darsbery Laboratory EXCURV92 Program, 1991. [7] V.A. Chernov, N.I. Chkhalo, M.V. Fedorchenko, E.P. [12] V.A. Chernov, V.I. Kondratiev, N.V. Kovalenko, S.V. Kruglyakov, S.V. Mytnichenko, S.G. Nikitenko, J. X-Ray Mytnichenko, Nucl. Instr. and Meth. A 470 (2001) 145, Sci. Technol. 5 (1995) 65. these proceedings. [8] V.A. Chernov, N.I. Chkhalo, M.V. Fedorchenko, E.P. Kruglyakov, S.V. Mytnichenko, S.G. Nikitenko, J. X-Ray Sci. Technol. 5 (1995) 389.