The ESRF-EBS (Extremely Brilliant Source) is a 150M€ project to totally upgrade the ESRF storage ring, with a completion date in 2020. This upgrade will increase the brilliance on some of the insertion device beamlines by a factor of 100. The new lattice will replace the old double bend achromat magnet sequence between the insertion devices with a new multi bend achromat design, increasing the number of bending magnets between the straight sections to seven.
The X-ray Source
This new low-emittance ring does necessarily mean major changes for all the bending magnet beamlines at the ESRF. The most significant impact is the removal of the old bending magnets as sources from the lattice. Within the new lattice, a space has been identified in which a small source with 2 mrad acceptance could be located. The new source position is slightly shifted with respect to the current source, about 2.7 m upsream. With the smaller acceptance, the source size is reduced but with quite improved brilliance, delivering a significantly smaller beam and increased power density within the beam. Various source options were considered: using a small insertion device, exploiting the steering magnets already part of the existing lattice design, or using a small 0.86 Tesla dipole positioned between two of the new steering magnets. Finally a choice was made to use the 0.86 Tesla “short bend” as a source for XMaS, as this provided a clean uncontaminated x-ray source and considerably more flux at higher energies, since the magnetic field of the dipole is more than double the 0.4 Tesla value of the old source.
Calculated flux curves. The dashed line shows the flux from the previous 0.4 T, 3 mrad horizontal fan source, whereas the solid line shows the flux from the 2 mrad fan of the new 0.86 T short bend.
From the very first design studies of the XMaS beamline, it was decided to use a very simple optical system consisting of a directly water cooled Si<111> monochromator and a toroidal mirror in a roughly 1:1 optical geometry. This simple, but robust system (Fig. 2) has ensured that a diverse and interdisciplinary science program can be maintained. It is intuitive for users and allows for rapid changes of x-ray energy by simply rotating the monochromator Bragg axis and adjusting the distance between the first and second monochromator crystals. Downstream of the monochromator the next optical element is the toroidal mirror which focuses the beam onto the sample. In addition, for further beam conditioning, two flat harmonic rejection mirrors and a phase retarder can be employed. The largest impact of the EBS upgrade will be the movement of the source some 2.7 m upstream in the lattice. Again, to ensure that the benefits from the new source are fully maximized, the experimental hutch has been extended by 4 m downstream and the diffractometer moved to maintain the current 1:1 focus. Overall, however, the upgrade will have a limited impact on the optical design but the main optical components will be refreshed to latest standards and re-designed to match the source characteristics.
Schematic diagram of the optical configuration of the XMaS beamline. Post EBS upgrade, an additional toroidal focusing mirror will be used to extend the upper energy range of the beamline and to take advantage of the new smaller source, 2.7 m further upstream from the previous source.
- Focussed Monochromatic Beam: both the monochromator and the focussing mirror are used. The beamline is normally used in this configuration.
- Unfocussed Monochromatic Beam: the monochromator is used, but the mirror is driven below the beam which passes unfocussed into the experimental hutch. This allows a broad fan of monochromatic radiation to pass into the experimental hutch.
The old water cooled monochromator has been replaced by a new LN2 cooled system to cope with the increased thermal load from the new source,helping to reduce any heat-bumps. In addition, to exploit the flux at higher energies, the energy range will be extended from 2.035 keV to an upper range of approximately 40 keV. Directly cooled LN2 copper blocks are mounted to the sides of both the first and second silicon monochromator crystals, clamped using a pre-defined stress with a thin layer of indium metal between the copper blocks and the silicon crystal, to ensure good thermal conductivity. Both cooled crystals are thermally isolated by intermediate invar plates mounted onto ceramic balls, creating point contacts, thus minimizing thermal bridges between the crystals at cryogenic temperatures and the crystal cage at room temperature. Three high precision linear actuators control the distance between the first and second crystal, with the second crystal sitting on a three point kinematic mount. Also, the speed of the Bragg axis rotation has been increased to 1 degree per second.
Focusing Toroidal Mirrors
For focusing the beam, the same simple toroidal mirror geometry that was previously used has been retained. However, as the useable energy of the beamline is to be extended up to ~40 keV, new mirror coatings and angles of incidence were required. One particularly demanding requirement, primarily for XAS studies, was the ability to rapidly change energy between 2.035 and ~40 keV whilst maintaining the focal spot at the same position. The old primary mirror was rhodium coated, with an incident angle of 4.5 mrad. Along with the mirror coating material, the angle of incidence effectively imposes an upper-energy cutoff for the beamline. To increase the energy cutoff, therefore we need, a metallic coating with a higher density, and/or a decrease in the angle of incidence of the mirror. Platinum is a good mirror coating for higher x-ray energies, however the Pt L-edges present problems at lower energies, especially as Pt is an important catalysis material. It was therefore necessary to incorporate a second toroidal mirror, that could be translated into the beam, to maintain a clean energy spectrum across the entire 2.035 to ~40 keV energy range. A chromium coating has been selected for the second mirror with the Pt mirror covering the energy around the Cr K-edge at 5.989 keV (Fig. 3). To ensure that the mirrors could be translated seamlessly and the focal spot to remain in the same position, the incidence angle for the two mirrors must be identical. The mirrors will use an angle of 2.5 mrad as this is the best compromise between the delivered flux for both high and low energies. One problem with operating x-ray mirrors at very low angles is that any ‘real’ mirror has a finite length and the incident beam from the monochromator will overspill the mirror at low angles resulting in a loss of transmitted flux, especially at lower x-ray beam energies where the vertical fan size increases. Considerable care has been taken when specifying the mechanics for these two mirrors to enable the focused beam to remain at the same point in space as the mirrors are translated across the beam. It is non-trivial task to align two toroidal mirrors on a mechanical translation slide, as any small misalignments are magnified by the 20 meter lever from the mirror to the focal point of the beam. The motorized step resolution for the yaw motion will be <5 µrad and <5 mrad for the less sensitive roll motion. Both mirrors will have a new fixed sagittal radius and variable meridional radius, the latter being controlled by independent mirror benders. This proposed optical solution will deliver a clean energy spectrum from 2.035 to approximately 33 keV.
SHADOW calculations were performed to estimate the total flux transmitted assuming a 1.2 m long toroidal mirror, a 2 mrad horizontal fan of bending magnet radiation and an incident angle of 2.5 mrad. The mirror width was set to 120 mm, with the sagittal and tangential radii set to their theoretical values calculated by SHADOW. Incorporating realistic parameters for the optical elements has enabled us to model the new source characteristics as functions of both energy and focal length. We estimate the new focal spot size ~(60 µm horizontally x 70 µm vertically) will be significantly smaller than our current focused spot size of ~(300 µm horizontally x 600 µm vertically) but with a comparable flux (Fig. 4). As the source point has moved, aberrations in the beam shape are evident at the current diffractometer position but are minimised at the 1:1 focused position. Our experimental hutch has been extended by 4 m downstream which also allows additional beam defining and conditioning elements to be incorporated into the delivery design, increasing operational efficiency and allowing rapid switching between experimental configurations. As the majority of experiments performed on XMaS require a small footprint/small angular divergence, the more brilliant source increases usable flux density for most users by one order of magnitude or more. Further increases in the S/N ratio are derived from the reduced background as the beam is no longer being defined close to the sample.