Metallic Multilayers Mirrors : New Possibilities for Soft X-ray Imaging and Spectroscopy

Successive evaporations of two correctly choosen elements permit now to obtain high efficient Bragg reflectors in X-UV where until now the grazing incidence optics was the only one solution. The potentialities of the X-U.V range for imaging and microscopy have been clearly underlined since the fifties (1) but pratical difficulties have damped a real development, not only in the field of X-rays analysis but also for plasmas diagnostics, X-U.V launched astronomy and synchrotron radiation. It is known that from some angstroms to some hundred of angstroms all materials had very low dispersion and pretty high absorption (2) which ruled out possibilities to built classical lens. Unfortunatly the absorption is not large enough to attenuate the electromagnetic wave to a depth comparable to one wavelength and as a consequence normal incidence mirrors had too low reflectivity for pratical uses. But, because optical indices in this range have a value sligtly smaller than one, the total external reflection phenomena can be obtained under grazing incidence. So all optical systems have used this grazing incidence solution but with the penalties of astigmatism and low aperture problems. To limit the astigmatism, optics with double cylindrical or spherical grazing incidence mirrors had been used ( 3 , 4 ) . More expensive optical solutions have been developed, mainly by astronomers (5), with combinaisons of parabolic and hyperbolic mirrors. Recently noticeable progress have been achieved in POlishing and simultaneous control of shaping, which permits to hope to bring the grazing incidence optics closer to the theoretical possibilities (6). Two new techniques, closely connected with the progress and goals for related microelectronics, have offered new possibilities to bypass the grazing incidence solution. The first one is to build adapted Fresnel zone plates with alternate transparent and absorbing rings (7). The second way is to build multilayers appropriate to the X-U.V optical index values. It is a very attractive solution because it will permit to adapt all spectromeqers and imaging systems using dielectric multilayers or mirror combinaisons well known in the visible, and also in X-rays regions using crystals. Multilayers optics are interferential systems so they are band pass reflectors or filters and they obey the Bragg law 2d.sinB = NX. Some multilayers, like the Langmuir Blodgett organic multilayers, are sometimes used in X-U.V but much more efficient and tailored systems can now be built with any periodicity by evaporating alternatively two correctly choosen elements. Such evaporated multilayers can then fill the gap between the visible and the X-ray range, which is existing presently for interferential mirrors. Like in the dielectrics mirrors the evaporation techniques permit to control all parameters related to specific need at desired wavelength For example the condition 2d = h for normal incidence can be attained. Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1984216 C2-66 JOURNAL DE PHYSIQUE The possibility of obtaining multilayers in the X-U.V range have been demonstrated in 1972 (8) using well developed matricial calculations for the visible regions by including the complex optical index values. Other class of theories, closer to X-ray diffraction method, have also been developed with the same classical approximations (9). ~nformations obtained by these different methods are similar and can be briefly summarized here to demonstrate the flexibility of the tailoring and to explain why we can obtain much more reflectivity than natural and organic crystals. At a given wavelength, the best peak reflectivity is obtained by layering two materials, noted by subscript 1 and 2, having the largest absorption ratio or equivalently the highest c1/c2, if the imaginary part of the index is used. For each value of this ratio, the relative thickness f3 of the more absorbing material over the period must be correctly adapted. It may be noted that amorphous or crystalline materials are equally convenient. The figure given here is a result of a systematic studies (lo), using I ~4 ! z lo3 f a lo2 a' ,d g 'S 2 S Z 40 60 100 200 WAVELENGTH (A) tabulated X-U.V index of all elements (Z ) , in view to select the best couple of material to be layered. Because the large number of atomic thresholds present in the X-U.V range, a given couple is suitable only in a limited range. On this figure the maximum peak reflectivity R calculated for normal incidence is indicated for each range. The corresponding f3 and also the number N of layers evaporated to get this reflectivity in normal incidence are also indicated. The band pass of such multilayers is proportional to 1/N. For several pratical reasons, all elements of the periodic table cannot be layered and some arbitrary limitations have been introduced in their choice. Other independent calculations, using different choice, attain the same values and give a confirmation (11) about these high reflectivities predictions. Tuning of the f3 value during layering permits to get alternatively best peak, integrated reflectivity or narrowest band pass (12). More generally the difficult problem to supress the overlapping order can be partially solved in the same way. For example, an equal thickness of the two materials inside the period ( 6 = 0,5) corresponds to a centro symmetric period which cancels the second order and a purely sinusoidal index variation in the period will lead to a single first order reflection. Let us look to the main evaporating systems developed up to now to produce efficient multilayers and to some pratical problems. E. SPILLER was the first to introduce an in situ X-ray reflectometer inside an electron bombardment evaporator in view to control the thickness of each layer during the growth. Such a system produces aperiodic optimized multilayers for the specific angle and the wavelength used in the in situ reflectometer. Ion bombardment techniques have also been developed and they permi$ to obtain amorphous or epitaxied multilayers (13,14). Very thin layers, down to 8 A thickness have been built by magnetron sputtering. This high rate evaporator is able to produce multilayers with a large number of periods only by rotating the substrate at a constante speed above the two cathodes. Diodes and triode sputtering also produce multilayers. Another type of evaporator, initially developed to get semi conductor and semi metal supergratings, uses the pulsed laser focused under vacuum on two targets. It is now being successfully used to produce verygoodx-UVmulti layers@). X-U.V reflectivity tests have been extensively developed to compare experimental values with theogetical prediction. Short wavelength measurements using the Cu K a line ( A = 1.54 A) are the more usual probe for the layering and to measure the 2d. These measurements are useful to eLucidate the electronic density profile obtained in the period. This parameter determines all the qualities of the multilayers. Reflectivity tests at longer wavelength are also achieved by using fluorescence excited soft X-rays lines or with window X-rays tube. Only the synchrotron source can offer a wide range of wavelengths. The high polarization of this source also permits ta demonstrate the possibility to obtain an X-U.V linear polarizator by using a 45" incidence reflexion on a multilayers (15). Attention is now given to evaluate effects of slope defects on the mirror substrate and rugosities appearing during layering. These defects enlarge the rocking curve and decrease the reflectivity. Theoretical calculations (10) and experiments are in progress to know more about this limiting factor (16). Among the numerous applications under development in X-U.V optics, a few can been selected just to illustrate the new possibilities. Historically the first direct test was the normal incidence double spherical Schwarzschild optics at 120 A (17). Multilayers coating of single shaped mirror is the simplest solution to obtain directly selected band pass image in astronomy or plasma diagnostics but focussing by bending flat layered substrates are also in progress (18). The possibility to obtain the equivalent of the Fabry Perot in X-rays had also been demonstrated by T.W. BARBEE. Linear polarizer using a 45' incidence multiiayer have been recently used to test the circularity of the light emitted at 130 A by an helical undulator set on a storage ring (19). A compact multiband spectrometer using several multilayers coupled with a streak camera close to a laser plasma experiment have been used to study simultaneously temporal evolution of several characteristics lines (20). New ideas are also blooming, for example the use of the multilayers simultaneously in transmission and reflection to get X-U.V beam splitter (21). High reflectivities and large band pass X-U.V multilayers optics can also find wide application in electron and ion microphobe or for fluorescence analysis. These reflectors can sustain high photon flux and are proved to be useful to synchrotron radiation studies. They are also an attractive solution to the resonator cavity for the X-U.V lasers now in progress. 1. KIRPATRICH P. and BAEZ A.V., J. Opt. Soc. Amer. 2 (1948) 766. 2. HENKE B.L. et al., Atomic Data and Nuclear Data Tables 27 (1982) 1. 3. SEWARD F. et al., Rev. Sci. Instrum. 9 (1976) 464. 4. MONTEL M., X-ray Microscopy and Microradiography (ed. V.E. COSSLET) p. 177 (1957) 5. CHASE R.C and SILK J.K., Appl. Optics 14 (1975) 2096. 6. FRANKS A., This conference 7. SCHMAL G. et al., This conference. 8. SPILLER E., app. Phys. Lett. 20 (1972) 365. 9. BARBEE T.W., AIP Conf. Proceedings NO75 p. 131 (1981). 10. ROSENBLUTH A.E., Thesis Univ. of Rochester (USA) (1982). 11. GAPONOV S.V. et al., Nucl. Inst. Methods 208 (1983) 227. 12. SPILLER E., AIP Conf. Proceedings no 75, p.124 (1981). 13. SCHULLER I.K. and FALCO C.M., "VLSI Electronics : Microstructure Sciences" Vo1.4, Chap. 5, p. 183 (1982). 14. LOVE W.P. et al., Phys. Rev. B24 (1981) 6193. 15. DHEZ P., Adv. Space Res. 2 (1983) 199. C2-68 JOURNAL DE PHYSIQUE 16. MANCINI D.C. and BILDERBACK D.H., Nucl. I n s t . Methods 208 (1983) 263. 17. HAEBLICH R.P., i n Repor t DESY-S.R79/12 (1979). 18. UNDERWOOD J . H . and BARBEE T.W., Nature 2 (1981) 429. 19. GLUSKIN e t a l . , Synch ro t ron Users Meet ing Brookaven, S e p t . 1983. 20. STRADLING e t a l . AIP Conf. P roceed ing N O 75, p.292 (1981). 21. SPILLER E., New X-rays O p t i c s Meeting, July 1982 Abingdon (U.K) .