Development of a custom-made 2.8 T permanent-magnet dipole photon source for the ROCK beamline at SOLEIL

Abstract

In August 2021, the SOLEIL storage ring was restarted after the summer shutdown with a new bending magnet made entirely of permanent magnets. Producing a magnetic field of 2.8 T, it replaced one of the 32 electromagnetic dipoles (magnetic field of 1.7 T) of the ring to allow the ROCK beamline to exploit more intense photon fluxes in the hard X-ray range, thus improving the time resolution performances of the beamline for experiments carried out above 20 keV. The reduction of the new dipole magnetic gap required to produce the higher field has led to the construction and installation of a new vacuum vessel. The realization of the new dipole with permanent magnets was a technological feat due to the very strong magnetic forces. The permanent-magnet assembly required dedicated tools to be designed and constructed. Thanks to accurate magnetic measurements, a precise modelization of the new dipole was performed to identify its effects on the electron beam dynamics. The first measurements carried out on the ROCK beamline have highlighted the expected increase in photon flux, and the operation performances remain unchanged for the other beamlines. Here, the major developments and results of this innovative project are described in terms of technology, electron beam dynamics and photon beam performance on the ROCK beamline.

Keywords: ROCK photon beamline; SOLEIL; permanent magnet dipole; synchrotron light source.

Figure 1

Comparison of the performance of the standard dipole and the Superbend. (Top) Variation of the magnetic field along the electron trajectory. (Bottom) Variation of the photon flux as a function of the photon energy (calculated at 8.7 m from the source point for an aperture of 13 mm × 2.4 mm).

Figure 2

View of the bottom part of the electromagnetic version of the Superbend. (Green) The yoke and the poles. (Red) The coils. X, Z and Y are the horizontal, vertical and longitudinal dimensions, respectively, expressed in millimeters. This design was rejected and replaced by a permanent-magnet version.

Figure 3

(Top) Partial view of the high field part (half-yoke) of the Superbend. (Bottom) Transverse view of the exit of the low field part (half-yoke) of the Superbend. (Light green) Pole. (Blue) PM blocks. (Dark green) Yoke. (Orange) Magnetic short circuit. (Red arrows) Direction of the magnetization of the PM blocks. X, Z and Y are the horizontal, vertical and longitudinal dimensions, respectively, expressed in millimeters.

Figure 4

View of the high (1) and low (2) magnetic field parts of the Superbend, with (top) and without (bottom) the yoke.

Figure 5

(Top) View from inside the storage ring of the standard electromagnetic dipole. (Bottom) Same view with the Superbend dipole.

Figure 6

The tools (in gray) used for the installation of PMs around the high field poles (top) and low field poles (bottom).

Figure 7

Distribution of the measured angular defect and magnetization for all the PM blocks.

Figure 8

The PM Superbend installed on the Hall probe bench in the SOLEIL laboratory.

Figure 9

(Top) Variation of the magnetic field along the beam trajectory in the Superbend. (Bottom) Beam trajectory (in black) deduced from the measured 2D magnetic field mapping (in colors).

Figure 10

Variation of the field integral during testing of the sensitivity to the yoke temperature. (Top) Versus time. (Bottom) Versus yoke temperature.

Figure 11

Original vacuum vessel installed in the previous ROCK electromagnetic dipole.

Figure 12

Top view layout of the new Superbend vacuum vessel. (1) Dipole part; (2) quadrupole part; (3) crotch box; (4) pumping sub-assembly; (5) quadrupole pumping port; (6) beamline outlet; (7) crotch absorber flange.

Figure 13

Layout of the transverse profile of the new vacuum vessel. The wall thickness of the high field part was reduced to 0.6 mm thanks to the small surface of the pole (only ∼50 mm × 50 mm). All the transition taper angles were fixed to 0.02 rad.

Figure 14

(Top) high (1) and low (2) field gap footprint on the dipole part of the new vacuum vessel. (Bottom) Zoom of the transition from the high (1) to low (2) field parts.

Figure 15

The Superbend vacuum vessel installed temporarily inside the standard electromagnetic dipole and equipped with heating films. The upper yoke of the electromagnetic dipole is not yet back in place in the photograph.

Figure 16

The Superbend vacuum vessel installed in the final configuration, ready to receive the Superbend.

Figure 17

Variation of the quadrupolar component (solid line) along the beam trajectory in the Superbend. The magnetic field (dotted line) is plotted as a reference.

Figure 18

Longitudinal variation of the sextupolar component (solid line) generated mainly by the high field of the Superbend. The magnetic field (dotted line) is plotted as a reference.

Figure 19

On-momentum dynamic aperture and corresponding frequency map calculated at the injection point. (Top) Without Superbend. (Bottom) With Superbend after compensating for the focusing effect. Simulations were performed with the SOLEIL version of the 6D true symplectic TRACY tracking code. The color bar indicates the tune diffusion rate after 2 × 1026 turns.

Figure 20

The Superbend installed in its place in the storage ring tunnel. In red, the frame equipment dedicated to extracting the dipole if a vacuum vessel bake-out must be performed.

Figure 21

Example of the temperature variation in the storage ring tunnel during operation. (Red) Air tunnel temperature. (Black) PM temperature. Temperatures on the vertical axis are given in °C.

Figure 22

Black lines are the flux of the monochromatic beam measured at the sample position for a full spot size, plotted as a function of X-ray energy for an electron beam current of 500 mA and 2.8 T source. Diamonds are related to measurement made with the Si(111) monochromator whereas circles relate to measurements carried out with the Si(220) monochromator. For comparison purposes, the flux measured at the sample position for the 1.7 T source and Si(220) monochromator (gray dotted lines) is reported as well as the flux calculated for perfect optics using the 2.8 T Superbend (red lines).

Figure 23

(a) Pd K-edge normalized absorption spectra recorded for a 1 wt% Pd loaded catalyst supported on CeO₂ (measurement in transmission in 5 s using the 2.8 T source and 12.5 s using the 1.7 T source). A zoom on the upper part is added. (b) Ce K-edge k ²χ(k) EXAFS spectra recorded on a CeO₂ pellet (measurement in transmission in 5 s). Dotted red lines correspond to measurements made with the 1.7 T bending magnet source and black lines to those made with the 2.8 T Superbend.

Figure 24

Comparison of k ³χ(k) EXAFS spectra extracted from full-field TXM images recorded on an alumina cylindrical extrudate impregnated with a 1 M Mo solution with the 1.7 T source and the 2.8 T source. The spectra were extracted from the pixel at the center of the extrudate but considering different binnings of the camera pixel image. The image on the left (using the 1.7 T source) has been binned to reach a pixel size of 50 µm × 50 µm whereas at the right (using the 2.8 T source) the pixel size is 16.25 µm × 16.25 µm (Barata, 2023 ▸).