The ID19 beamline main features derive from the requirements of having a spectrally and spatially homogeneous, highly coherent, beam at the sample position, with (maximum) dimensions ? 45 x 15 mm2, a high photon flux and a tunable photon energy in the range of 6-120 keV. These requirements lead to the choices of a long (145m) beamline, and an eleven pole, variable gap, high-magnetic field wiggler (Bmax =1.4T, Ec,max = 32 keV) as source. The long source-to-sample distance, added to the small dimensions of the source (» 0.1 mm) are necessary conditions to obtain a highly coherent beam. The components of the beamline ID19 were designed and manufactured (polished windows, filters, reduction of small monochromator vibrations, etc.) to avoid, as much as possible, spurious images and/or losses of coherence on the X-ray path [8].

The optical hutch, located just after the front end, contains the diaphragms, slits, filters and a shutter. An about 100 m long tunnel joins the first hutch to the satellite laboratory, outside the experimental hall. The monochromator and experiments hutches are located in this satellite laboratory. The monochromator hutch comprises a "stroboscopic shutter" (adjustable chopper, which also allows to tailor the heat load on the sample, when working with the white beam), slits, the main monochromator and a "variable beam" (white and monochromatic) shutter. No mirror was installed (in 1995) because even the best mirrors introduced unacceptable contrasts in the beam.

The monochromatic beam is delivered, as a function of the required experimental conditions, by one among several possible monochromators :

a) a ‘single’ crystal or multilayer (depending on the wished DE/E) located at the beginning of the experiments hutch, protected by a shielding ("blue box")
b) a ‘bendable’ monochromator diffractometer has been built in collaboration with the group of Rolf Köhler (Berlin), mainly to investigate semiconductor wafers. It permits to overcome the spatial energy dispersion associated with the extended beam surface and the small source size, as well as the curvature of the sample
c) a double crystal, fixed exit, monochromator diffracting in the vertical plane is located in the monochromator hutch. The crystals are 30 cm long (111) Si, in symmetrical Bragg position; the first crystal is water cooled. The 333 reflection of this double crystal monochromator is used when a small DE/E is required.


The beamline is designed to use either the white beam or a monochromatic beam in the experiments hutch, which is therefore a rather heavy one (the thickness of the Pb walls is typically 30 mm). The experiments hutch is equipped with a marble floor. Consequently the several available setups can easily be transferred on air pads. These setups include:

  • a "horizontal" diffractometer for ‘real time' (white beam or monochromatic) experiments, which can accommodate special stages including a liquid helium or closed cycle cryostats, an electromagnet, furnaces and straining devices (designed in collaboration with colleagues from Marseille, Nancy and Lyon). This diffractometer, holding a special high precision stage, was also used, up to now, for the microtomographic work.
  • we are presently commissioning an actual microtomographic device to better respond to this increasing fraction of the scientific (and industrial) beamline activity.
  • a ‘bendable’ monochromator diffractometer already mentioned.
  • a "high precision" vertical diffractometer designed in collaboration with the group of Bernard Capelle (Paris). It allows to optimize the angular and spatial resolution, and perform high resolution diffractometry in combination with diffraction imaging.
  • special "accurate" slits for 'section' and 'pinhole' diffraction imaging and diffractometry, and various ‘fast’ shutters operating in the 10-2 second timescale.


Detectors are a crucial part of the experimental setup. Several kinds of detectors are currently used: X-ray films and nuclear plates, proportional counters (one of them being energy sensitive), and several CCD cameras equipped with a visible light optics and scintillators. Among these cameras we should point out the outstanding performances of the FRELON camera (Fast REadout Low Noise), designed and produced at the ESRF: it has, simultaneously, a 14 bits dynamic range and a readout time (1024*1024 pixels) smaller than 0.1 second. This camera is an essential device for our work: it is used for all the tomographic experiments, and for an increasing number of diffraction imaging experiments. Several scintillators and optics allow to select an effective pixel size between 0.4 and 40 µm. The whole ensemble sits on a one meter precise translation, to easily change the sample-to-detector distance which is a crucial parameter to take advantage of the coherence of the beam. In the very near future the beamline will be equipped with a second, (2048*2048) FRELON camera.

The beamline is controlled in the ESRF standard way: VME electronics, workstations, SPEC program (including special extensions to simultaneously move ‘motors’ and efficiently record images). ID19 is the (non-multiple) beamline which produces the largest amount of data at the ESRF. This is mainly connected with the microtomographic work where, presently, up to 8 Gigabytes of data can be produced in one hour. A devoted fast link connects the beamline with the central data storage, where 1 Terabyte of disk space will be foreseen for ID19 at the end of 2000. The bottleneck resides in the reconstruction time (2-5 hours) of the 3D images. The reconstruction process was therefore reconsidered to improve it as much as possible, both on the software and the hardware side.

A connected topic is the post acquisition image processing to extract the relevant physical parameters from the 2D or 3D images. A combination of commercial software (for 3D visualization, for instance) and special programs, written in-house or resulting from collaborations with external groups (Yves Epelboin for the sophisticated diffraction imaging simulation programs, Dominique Jeulin for the processing of 3D images) is expected to cover our needs. However a lot remains to be done in this area.