Monochromation and apodization with Ti-B4C multilayers in neutron optics

Monochromation and apodization with Ti-B4C multilayers in neutron optics

PHYSICA ELSEVIER Physica B 198 (1994) 231 234 Monochromation and apodization with Ti-B4C multilayers in neutron optics t M. M a a z a a' b, ,, A. M ...

298KB Sizes 2 Downloads 20 Views

PHYSICA ELSEVIER

Physica B 198 (1994) 231 234

Monochromation and apodization with Ti-B4C multilayers in neutron optics t M. M a a z a a' b, ,, A. M e n e l l e a, J.P. C h a u v i n e a u c, B. P a r d o c, A. R a y n a l c, F. B r i d o u c, C. Sella d, T. M e g a d e m i n i e a Laboratoire Li, on Brillouin, Commissariat it I'Energie Atomique-Centre National de la Recherche Seientifique, Bat. 563, Centre d'Etudes Nuclbaires de Saclay 91191, Gif-sur- Yvette, France b Atominstitut der Osterreichischen Universithten, Sch~ttelstrasse 115. A-1020 Wien, Austria c Institut d'Optique Thborique et Appliqube, Universitb de Paris-Sud, Centre d'Orsav, Bat. 503, 91405 Orsav, France d Laboratoire de Physique des Matbriaux, Centre National de la Recherche Scientifique, 1 place Aristide Briand, 92195 Meudon-Bellevue. France Dbpartment de Physique, Universitb des Sciences et Techniques de Masuku, B.P. 595, Franceville, Gabon

Abstract

The apodization within B4C Ti multilayered monochromators in the Kagan-Afanas'ev-Bormann geometry is checked experimentally. By using the absorption and transparency of B4C and Ti layers, respectively, this apodization manifests itself by narrow and intense first Bragg order coupled with a very low surrounding reflectivity over a large momentum range. Three ion beam sputtering multilayered monochromators, constituted by 9, 25 and 50 (20/~, B4C-120/~ Ti) bilayers, are tested. The first Bragg order reflectivity is about 12.4%, 60.6% and 73.5%, respectively, with a corresponding bandwidth A K B / K n of 10.5%, 5.6% and 4.1%, respectively, with a surrounding reflectivity smaller than 1% (left) and 1% (right). Compared to the classical Ni-Ti multilayered system, the B4C-Ti multilayered system in the

Kagan-Afanas'ev Bormann geometry is better from the monochromation and apodization point of view.

1. Introduction

The recent developments in the field of thin films technology have induced a wide use in neutron optics of these devices as monochromators, polarizers broadband filters and supermirrors [1-6]. The multilayered monochromators present advantages over the conventional perfect crystal: (i) Multilayers have large period spacings, which makes them useful for monochromating cold neutrons with wavelengths greater than 6/~. (ii) The bandwidth of the wavelength of the reflected beam is large, which

* Corresponding author. Dedicated to the memory of Monsieur Jean Rossat-Mignod.

results in a greater flux at the sample. This feature is particularly useful for low-resolution studies on partially ordered systems in which a high flux of neutrons is very desirable and a large value of A2/2 can be tolerated. (iii) The reflectivities of multilayers reported in the literature have been high, usually higher than 90% for multilayers with a A spacing between 80 and 200/~. Unfortunately, the Bragg peaks given by these periodic multilayered monochromators, with a reasonable number of bilayers, are surrounded by small interference fringes called Kiessig fringes [7, 8]. To use these devices conveniently in intermediate- and high-resolution spectrometers, in which a large value of A2/2 is not acceptable, the apodization of the corresponding Bragg peaks is required. By apodization, we mean

0921-4526/94/$07.00 © 1994 Elsevier Science B.V. All rights reserved SSDI 0921-4526(93)E0544-Q

M. Maaza et al./Physica B 198 (1994) 231 234

232

to obtain Bragg peaks with good reflectivity and small bandwidth A)./2, surrounded by very small Kiessig fringe intensity. To achieve this, the use of periodic multilayers in the so-called Kagan Afanas'ev-Bormann geometry is required [9-12]. This configuration consists of a packing of strong and weak absorbing materials in which the absorbing layers (reflector layers) are thinner than the spacer ones. The absorption effect of reflector layers is taken into account in the expression of the neutron refractive index which can be expressed as [13] nj = 1 - )~2 N j { + bj - ib~'}/2n,

tl)

hj = _+ x/[(aj~/4r0

(2)

b'j = (aja/22)/2rr,

-

(aja/2)~)2],

(3) 1,

in which i 2 = - 1, 2 is the de Broglie associated neutron wavelength, Nj is the number of scattering atoms per unit volume in the jth layer, aj~ and aja are, respectively, the scattering and absorption cross-sections of the material j [14], and bj and b~' are the real and imaginary parts of the nuclear scattering length. Experimentally, the nuclear (N~.bj) and the absorption ( N i ' b y ) scattering length densities are measured. In a previous paper, the apodization phenomenon within the neutron optics devices in Kagan-Afanas'ev-Bormann geometry was discussed in a theoretical and experimental approach [15, 16]. The interpretation of this apodization can be explained in the following manner: let us consider the periodic stack in the Kagan-Afanas'ev-Bormann geometry of Fig. 1. As the neutron wave penetrates the multilayer, it undergoes successive Bragg reflections from a large number of parallel interfaces. In each reflection the amplitude of the incident neutron wave is reduced somewhat. As the nodes of the standing Bragg neutron wave are close to the strong absorbing layers, the absorption is small as in perfect crystals. For Kiessig fringes, the antinodes of the standing waves, given by the superposition of the incident and reflected waves at air multilayer and multilayer-substrate, do not correspond, generally, to the absorbing layers. So, the corresponding beam will be gradually attenuated as it passes through these layers. As

Fig. 1. Schematic diagram of a multilayer mirror in Kagan Afanas'ev B o r m a n n geometry with a cyclic ratio 7 smaller than 0.5.

narrow intense Bragg peaks coupled with a reduced surrounding intensity are obtained, to some extent, such multilayers behave as perfect crystals [17]. The aim of this work is to present experimental results about the apodization within B4C Ti multilayers, in which the B4C layer absorption is a key to the experiments.

2. Experimental results Three Ti-B4C multilayered monochromators of various bilayer number were prepared by ion beam sputtering technique at a deposition temperature of 300 K from Ti and B4C targets [16]. The distance between the targets and the substrate is about 25 cm, assuring a thickness uniformity better than 1% on a 2 x 2 cm / area. The ion beam was extracted from a 3 cm diameter ion source. Prior to deposition, the system was pumped with a cryogenic pump down to a pressure of about 10 -8 mbar

233

M. Maaza et al./Physica B 198 (1994) 231-234

while deposition is performed at a pressure of 2× 10-4mbar. The ion beam was neutralized by injecting electrons with a hot tungsten filament. Ion beam sputtering was performed with argon ions of an energy of 1.2 keV and a current of 40 mA. The multilayers are deposited on optical surface quality borosilicate glasses which are cleaned in a special cleaning solution and then rinsed several times in bidistilled water. Before sputter deposition the substrate is sputter etched for several minutes. The rate of deposition is 15/~/min for Ti and 3.5 ~,/min for BaC. Also, the targets are cleaned by presputtering. The layer thickness is measured during deposition by a calibrated quartz microbalance, and a built-in soft X-ray reflectometer is used to control the periodicity of the stack [18, 19]. The main advantage of this double control, compared to the unique conventional quartz microbalance, for thickness monitoring lies in the self-compensation of thickness errors made on the successively deposited layers. One can note that this thickness control is very important in our case; only a very small thickness fluctuation is tolerated to achieve the so-called apodization of Bragg peaks. The number of Ti-B4C bilayers deposited is fixed to 9, 25 and 50, maintaining both the period and cyclic ratio constant. The expected values are 140/~, and 14.3%, respectively, corresponding to Ti and B4C layer thicknesses of 120 A and 20 ~, respectively. Thereafter, the 9, 25 and 50 bilayer samples will be designated by M1, M2 and M3, respectively. Neutron reflectometry measurements were carried out at the ORPHEE-Lron Brillouin Laboratory 14 MW reactor located at Centre d'Etudes Nuclraires of Saclay. The time of flight "EROS" reflectometer was used. The incident neutron wavelength varies from 3 to 25 A. The grazing angle 00 and relative angular resolution A0/00 are fixed at 2.1 × 10 -1 rad and 5 × 10 z, respectively. The theoretical neutron reflectivity profiles are calculated using the standard matrix method [20] by assuming an ideal periodic structure with sharp interfaces, zero interfacial roughness and no random thickness errors in the periodicity. Experimental neutron reflectivity in logarithmic scale versus momentum K = 2n sin0o/2 of the three mirrors is reported in Fig. 2. They are compared with theoretical Ni-Ti characteristics (period, number of periods, cyclic ratio) in which the Ni

[oo00o] Exp [-

] The

[BdC-Ti]/Boro [Ni-Ti]lBoro

-1

-2

(a)

31o ~ - - -Kiessig =2 fringes ' ,---,---~

-2

-4t

,

,

I . . . . 002

I . . . . 0.04

K = 2 n sinOo/~,

0.06

[A'11

Fig. 2. ( o o o o o o ) experimental neutron reflectivity profiles in logarithmic scale versus m o m e n t u m K = 2 n s i n 0 0 / 2 at 0o = 2.1 x 10 -2 rad of (a) 9 B4C-Ti, {b) 25 B4C Ti and (c) 50 B4C-Ti compared to the computed profiles ( ) given by their corresponding N i - T i stacks.

and Ti layers are free of absorption. Two main regions are observed: the total reflection plateau and the so-called vitreous region. In this last one, only two Bragg peaks occur. The broadening of the different experimental Bragg peaks is negligible; such as for X-rays, this indicates a uniform set of

234

M. Maaza et al./ Physica B 198 (1994) 231-234

bilayer thicknesses through B 4 C - T i stacks due to the performance of the thickness control monitoring. According to Fig. 2, both the reflectivity and contrast of Kiessig fringes are smaller than those of Ni-Ti, as expected theoretically; their experimental reflectivity is smaller than 1% for all samples. They disappear partially when the bilayer number is increasing (M2 and M3). The experimental reflectivity of the first Bragg peak is of the order of 12.4%, 60.6% and 73.5% for M1, M2 and M3, respectively, while that of the second peak is much smaller (smaller than 5%), which is not the case for N i - T i profiles. Likewise, from the resolution point of view, the Bragg peaks of B4C-Ti m o n o c h r o m a t o r s are narrower than those of Ni Ti; the width at half maximum of the first Bragg order is of the order of AKa/KB 10.5%, 5.6% and 4.1% against 12.15%, 7.17% and 5.1% for N i - T i ones. Thus, these results agree with the revised theoretical calculations and confirm the high apodization of Bragg peaks coupled with a good intensity within high absorbing-low absorbing multilayered periodic stacks in K a g a n - A f a n a s ' e v - B o r m a n n geometry. The considered value of 7 was fixed to 14.3%. To enhance the Bragg peak reflectivity of the previous multilayers, the optimization of the parameter 7 for a fixed period A is required. This optimal value, which depends on the nature of the absorbing material, is found by differentiating the so-called ZeldovichVinogradov transcendental equation generalized by Megademini [-21, 22]. The simulations indicate that the maximum of the first Bragg reflectivity, in this case, is of the order of 97% and corresponds to 7op, of the order of 45% for the B4C-Ti system [15, 16].

spectrometers. This apodization will be better with G d - Y i and C d T e - T i systems (over the resonance neutron wavelength range).

Acknowledgements We are indebted to Drs. B.R. Mfiller from Berliner Electronenspeicherringgesellschaft Synchrotronstrahlung and Sincrotrone Trieste and J. Corno from Institut d'Optique Th6orique et Appliqu+e for their invaluable help. Likewise, we wish to thank Dr. O. Guiselin from L6on Brillouin L a b o r a t o r y - C o m missariat fi l'Energie Atomique for his Fresnel reflectivity program, used to achieve the calculations.

-

3. Conclusion The apodization within B4C-Ti multilayered m o n o c h r o m a t o r s in the K a g a n - A f a n a s ' e v - B o r mann geometry is checked experimentally. This apodization manifests itself by a narrow and intense first-order Bragg peak coupled with a very low surrounding reflectivity over a large m o m e n t u m range. F r o m a practical point of view, to some extent, such multilayers behave as perfect crystals and can be used in future in intermediate- and high-resolution

References [1] B.P. Schoenborn, C.L.D. Caspar and O.F. Kammerer, J. Appl. Crystallogr. 7 (1974) 508 510. [2] C.F. Majkrzack, AppL Optics 23 (1984) 3524-3528. [3] F. Mezei and P. A. Dagleish, Commun. Phys. 2 (1977)41-43. [4] O. Schaerpf, Physica B 156-157 (1989) 631-638. [5] J.B. Hayter and H.A. Mook, J. Appl. Crystallogr. 22 (1989) 35 41. [6] T. Ebisawa, T. Akiyoshi, N. Achiwa, S. Yamada and S. Okamoto, Annu. Rep. Res. Reactor Inst. Kyoto Univ 14 (1981) 10 18. [7] H. Kiessig, Ann. Phys. 10 (1931) 715-723. [8] M. M~aza, Z. Jiang, F. Samuel, B. Farnoux and B. Vidal, J. Appl. Crystallogr. 25 (1992) 789-796. [9] A.M. Afanas'ev and Yu. Kagan, Zh. Eksp. Teor. Fiz. 48 (1965) 327-335. [10] G. Bormann, Z. Phys. 42 (1941) 157-162. [11] G. Bormann, Z. Phys. 127 (1950) 297-305. [12] M. Laue, Acta. Crystallogr. 2 (1949) 106 113. [13] H. Rauch and D. Tuppinger, Z. Phys. A 322 (1985) 427 432. [15] M. M~aza, V. Mikerov, B. Pardo, J.P. Chauvineau, A. Menelle, J. Raynal, F. Bridou and T. Megademini, J. Appl. Crystallogr., submitted. [16] A. Menelle, M. Mfiaza, B. Pardo, J.P. Chauvineau, J. Raynal, F. Bridou and T. Megademini, J. Appl. Crystallogr., submitted. [17] D. Schwahn, A. Miks6vsky, H. Rauch, E. Seidl and G. Zugarek, Nucl. Instr. Meth. A 239 (1985) 229-234. [18] J.P. Chauvineau, Rev. Phys. Appl. 23 (1988) 1645-1652. [19] B. Pardo, J.M. Andre and T. Megademini, Rev. Phys. Appl. 23 (1988) 1579 1597. [20] F. Abel+s, Ann. Phys. (Paris) 5 (1950) 777-795. [21] T. Megademini, Thesis of Paris XI University (1990). [22] A.V. Vinogradov and B.Y. Zeldovich, Appl. Opt. 16 (1977) 89 103.