In situ multilayer deposition studies by x-ray energy dispersive method

J. C. Malaurent*, H. Duval*, G. Cauchon**, P. Dhez**, O. Hainaut**, M,
Idir** and J.P. Chauvineau***.

* Laboratoire de Physique des Solides, Bat 510 Univ. Paris XI, F- 91405 Orsay
** Laboratoire de Spectroscopie Atomique et Ionique, Bat 350, Univ. Paris
XI, F- 91405 Orsay
*** Institut d'Optique Theorique et Applique d'Orsay, Bat 503 Univ. Paris
XI, F- 91405 Orsay

        As demonstrated early by E. Spiller, in-situ X-ray reflectivity is
adequate to provide basic information on multilayers during deposition and
on coating methods. Only few practical methods exist because the substrate
must be fixed during coating and the vacuum compatible reflectometer shoud
be without moving part.
        Spiller's in-situ reflectometor uses a monochromatic soft x-ray
line, at fixed incidence, produced by a windowless x-ray tube inside the
evaporator. In 1994 another in-situ x-ray reflectivity method was developed
but with source and detector outside. It used a classical x-ray  tube, a
curved Si111 crystal and a CCD detector with a phosphor converter (1). The
small angular divergence of the prepared monochromatic beam is suitable to
get reflectivity at 8 keV versus incidence angle. Compared to Spiller's
method, the 8 keV photons are able to penetrate all the stack, in addition
the angular opening of the monochromatic beam give the possibility to
follow simultaneously the reflectivity from the total reflection region to
Bragg peaks.
Reflectivity by an Energy Dispersive Method (EDM) is convenient for in-situ
reflectivity tests. As demonstrated few years ago, it is well-suited to
study multilayers on specular (2) and off-specular (3) conditions over a
broad energy spectrum, typically 5 to 50 keV. By initially setting the
incidence angle, the first order Bragg peak can be brought to any given
energy. Dependence of the calculated reflectivity on possible indices
uncertainties, versus energy, were pointed out but appear to be quite low
due to the evaluated indice table accuracy (4).

        Mo/Si multilayers deposited on (111) silicon wafers by electron
beam evaporation and by ion beam sputtering have been studied
experimentally by EDM reflectivity with simultaneous monitoring by an
in-situ Spiller's type reflectometer at 44=C5. Even for the first deposited
periods, EDM permits to observe the secondary fringes appearing between the
total reflection limit and the first Bragg peak. Peak reflectivity and band
pass of the Bragg peaks have also been followed during deposition. Like
others reflectivity methods, EDM is sensitive to interface roughness during
the deposition of successive periods and roughness estimates were
performed. Experimental results were compared to corresponding calculated
models.


1. E. L=FCken, E. Ziegler, P. H=F8gh=F8j, A. Freund, E. Gerdau, A. Fontaine,
"Growth Monitoring of W/Si X-ray Multilayers by X-ray Reflectivity and
Kinetic Ellipsometry", Coatings for Extreme Region of the Electromagnetic
Spectrum, vol. 2253 Optical Interference Coatings, pp. 327-32, European
Optical Society and International Society for Optical Engineering,
Grenoble, 1994. and E. L=FCken, PhD thesis, Hamburg university and ESRF
(1996)

2. P. Dhez et al. J. X-Ray Sc. Techn. 3,176-193 (1992).

3. J.C. Malaurent et al. Opt. Comm. 113, 1-8 (1994).

4. H. Duval, Opt. Comm. to be published (1996).