The technique of neutron reflectivity has emerged less than 15 years ago. It has appeared as a key technique in the study of polymers and magnetic thin films. Problems such as polymer interdiffusion, di-block copolymers ordering have been addressed. In the last ten years, especially after the discovery of the giant magnetoresistance (GMR) effect, neutron reflectivity has successfully been applied to the study of magnetic multilayers and ultrathin films. Problems such as the magnetic ordering in rare-earth multilayers, surface magnetism and anisotropy in ultrathin magnetic layers have been solved. However, in order to push further the possibilities of neutron reflectivity, it has appeared that improvement in the flux and focalisation were necessary. In that aim, we have recently rebuilt a new polarised reflectometer with polarisation analysis at the Laboratoire Léon Brillouin, dedicated to the study of magnetic thin films.

To increase the available neutron flux, the different solutions we have used are the following :

• increase of the wavelength spectrum width dl/l by using a multilayer monochromator (compared to the very high resolution dl/l = 0.6% of a graphite monochromator).
• focalisation of a 100 mm high neutron beam onto a 15 mm high region (which is suitable for the study of small magnetic samples (typically 1cm²).
• use of a position sensitive detector (measurement of both spin states at a time and access to off-specular measurements).

The first version of the PADA reflectometer was a two axis spectrometer using a graphite monochromator which had an excellent wavelength resolution dl/l (~0.6%) a beam focalisation using three graphite crystals.

The new PADA spectrometer is mounted on the guide G2 of the reactor Orphée. The neutron beam is deviated and monochromatised by a 3 m monochromating guide (A) made of nickel-titanium multilayers. The direction of the monochromatised beam makes an angle of 2.4° relative to the direction of the main guide (see figure 2). This part of the guide has been built and mounted by the CILAS company. The angular deviation of 2.4° is however not sufficient to move enough away the output beam from the main guide at the sample position. Thus, we have mounted two 1.80 m long 2q_c-supermirror neutron guide (B) at the output of the monochromator. The total beam deviation is thus 5°, and the beam deviation at the sample position (8 m away) is of 900 mm from the main guide, making possible the mounting of the necessary protections around the main guide (50mm lead, 250mm concrete) and the sample environment (cryostat and magnetic field) (see figure 1). Since most of the studies on magnetic samples are realised on samples less than 20mm wide (and more generally 10 mm wide), we have focused the beam vertically on the sample position. The vertical focalisation is realised by a 8 m conical neutron guide (C) made of 2q_c-supermirrors (see figure 2). The vertical focalisation optics is interrupted twice. These two interruptions make it possible to mount the polarising (D) and flipping systems. The beam is polarised by transmission polarisers made of Fe/Si multilayers deposited on 50 mm high silicon substrates. These mirrors were provided by Th. Christ (from the HMI in Berlin). The incidence angle on the mirrors is 0.3°. The polarisation efficiency is 0.97. The incidence angle on the polarisers is small in order to reject long wavelengths neutrons generated by the monochromator system by reflection on the polariser and let only go through the 0.4 nm neutrons. The transmission of these polarisers is 70%. The polariser (D) is mounted 2 m away from the sample in a 50 mT field to maintain a good polarisation efficiency. The analysis system (E) is similar except that the height of the device is 80 mm to enable a full analysis of the reflected beam which is highly divergent in the vertical direction.

With these two improvements (new monochromator and focalised beam), the available flux has been multiplied by a factor 15 and is presently of 5x10⁵ neutrons/cm².s after analysis on the detector with a resolution dq = 0.03°.This large flux increase will allow two main progresses in the reflectivity experiments. It will drastically reduce the measurements times of the reflectivity curves. This will allow measurements in various conditions of magnetic field and temperature. The variation of these external parameters will make it possible to add external "strains" during the numerical fits of the reflectivity curves and thus improve their reliability. The second gain is that it is now possible to access to a diffusion vector range twice as large as before (it was previously limited by the lack of intensity). The in-depth spatial resolution can thus be multiplied by a factor two. In the present state of the spectrometer, the wavelength resolution dl/l is now of 4% and the angular resolution dq can be varied from 0.01 to 0.06°. The intensity dynamic range of measurement is now typically between 10⁵ to 10⁶ on a 1 cm² sample for an 8 hours full analysis scan and a resolution dl/l = 4% and dq = 0.05°.

In the year 1999, a position sensitive detector will be installed (currently being developed in the frame of the European XENNI program). It will allow off-specular studies on magnetic systems.  

for further details, contact :

Claude Fermon or Frédéric OTT

Lab. Léon Brillouin

LABORATOIRE LEON BRILLOUIN (CEA/CNRS) mise à jour : 30/07/99