Beamline Description

The layout of FERMI and the main properties of its light (high-brilliance, short-pulse length, variable polarization, and coherence) have been described elsewhere [1-3]. The 100-4 nm wavelength range is covered by two distinct undulator chains: the long-wavelength FEL-1 (100-20 nm) and the short-wavelength FEL-2 (20-4 nm). The LDM beamline is operating in the full photon energy range of FEL-1 and FEL-2 and the main parameters of these ranges are summarized here.


FEL photons are delivered to the LDM end-station through the PADReS (Photon Analysis, Delivery, and REduction System) section of the FERMI FEL. A layout is sketched on the following Figure:



PADReS covers the section between the exit of the FEL and the beamlines it currently serves. A set of plane mirrors (PM1a and PM1b serving FEL-1; PM2a serving FEL-2) is installed in the so-called safety hutch. Each FEL source has its own beam-diagnostics and beam-conditioning instruments, such as the intensity and beam position monitors, beam defining apertures, and gas absorber. At the end of the safety hutch the two photon beam paths enter the PM1b chamber, where one or the other can be selected.
 
Outside the safety hutch, the energy spectrometer (PRESTO) records the spectrum of each pulse by diffracting and detecting 1–2% of the total intensity [4], while delivering the essentially unperturbed beam downstream. A split-and-delay line is installed after the spectrometer, for FEL pump – FEL probe experiments. The instrument is based on wavefront splitting and subsequent recombination of the incoming radiation on the edges of grazing-incidence plane mirrors, and can introduce a [-2,+30] ps-delay at all wavelengths.

The 3-way switching mirror chamber selects which beamline is in use. The plane switching mirror (SW) serving LDM deflects the beam in the horizontal direction, and has a dual coating (graphite and iridium, each covering half the width of the mirror), intended to maximise the reflectivity for FEL-1 and FEL-2, respectively. Downstream of the switching mirror, the beam undergoes three further reflections: on a vertical deflecting mirror (VD) and on two Kirkpatrick-Baez (K-B) mirrors [5].

The FEL light co-propagates with a certain amount of seed laser light, with a wavelength in the range 230–260 nm, and this light may interfere with the experiment. To suppress it, there are three retractable aluminium filters of thicknesses 200, 400 and 800 nm in the PADReS system.
The light enters the experimental station horizontally, where it can be routinely focused to a spot of about 30 μm diameter. Substantially better focusing has also been achieved when more time and/or other instrumentation was available.

Focusing performance

Calculations of the expected spot size and shape have been carried out for both sources (FEL-1 and FEL-2) in their whole wavelength range, using the codes SHADOW [6], based on ray tracing, and WISE [7], based on physical optics.
Calculated spot profiles in the case of smallest spot size (4 μm × 6 μm FWHM for FEL-1 and 3μm × 5μm FWHM for FEL-2) are represented in Figure. Here the effect of the non-ideal mirror shape can be seen as a broadening of the spots and the presence of some diffraction peaks.
The additional techniques have been adopted by PADReS to optimize focusing of the photon beam according to the needs of the user. For studies requiring microfocusing, the most informative diagnostic is a wavefront sensor (WFS): with it the focus down to an average spot size of 6.5μm (FWHM) has been achieved and it’s in a good agreement with the WISE simulations (see Figure) While for a larger spot size, above ≈20μm (FWHM), a simpler fluorescent screen (YAG, YAP, or phosphor) inserted at the nominal focal plane is used.



Simulated focal spots for the LDM beamline in the case of ideal (red) and real (blue) mirrors; WISE program was used. The intensity profiles are calculated for FEL-1 at 30nm and FEL-2 at 4 nm, and are displayed along the vertical and the horizontal directions. The diffraction effect is due to the finite size of the mirrors. The smallest achievable spot sizes (FWHM) are predicted to be ≈4μm×6μm for FEL-1 and≈3μm×5μm for FEL-2.

Beamline transmission

Due to the beamline geometry (five mirrors reflecting in the horizontal plane, two in the vertical one for FEL-1; four and two for FEL-2) the vertical and horizontal polarizations are not equally transmitted to the LDM endstation, thus the ellipticity varies across the whole wavelength range. The linear horizontal polarization suffers greater losses than the vertical one. While this effect is negligible below 40 nm, above this wavelength it rapidly increases. As an example the calculated overall transmission of the beamline (excluding geometrical losses) is shown in Figure. The choice of APPLE-2 undulators for FERMI allows a smooth control of the polarization of the emitted radiation: i.e., one can set the FEL polarization to be elliptical upstream of the transport optics, in order to have circular polarization radiation at the endstation [3].
The LDM beamline transmission is also affected by some geometric losses due to the finite size of the mirrors. For example, the geometrical acceptance at different wavelengths is reported in the following Table.

Wavelenght (nm)	FEL	Θrms (μrad)	Geometrical acceptance (%)	Total transmission (%)
65	1	81.3	41.3	22.7
52	1	65	54.5	31.6
43	1	53.8	68.1	38.8
32	1	40	86.9	51.3
20	1	25	99.3	56.6
20	2	30	97	55.3
<10	2	15	100	65.0

Calculated transmission of the LDM beamline optics for the two FEL sources and the two linear polarizations (FEL-1: full (vertical) and dotted (horizontal) red curves; FEL-2: full and dotted blue curves). The geometrical losses have not been included in the graph.


[1] Allaria, E. et al., New J. Phys. 12(7), 075002, (2010).
[2] Allaria, E. et al., Nature Photon. 6(10), 699, (2012).
[3] Allaria, E. et al., Phys. Rev. X 4, 041040, (2014)
[4] Svetina, C.et al., Proc. SPIE, 8139, 81390J, (2011)
[5] Svetina C et al, Proc. SPIE 8503, 850302, (2012)
[6] Sanchez del Rio et al., Journal of Synchrotron Radiation, 18 (5), 708, (2011).
[7] Raimondi, L. et al., Nucl. Instrum. Methods A, 710, 131, (2013).