The material of the vacuum windows was selected based on two criteria: transmittance and susceptibility to the high-intensity UV radiation. The experiments conducted at BESSY, Germany, show different degrees of radiation damage (colour centre formation) in cases of various commonly used window materials [MgF2, CaF2, SrF2, BaF2, Suprasil quartz HER102 and SU311, and sapphire with crystal orientation along the c-axis (0001) to substantially minimize its birefringence] which were exposed to the same level of light intensity (DIP 3.1B beamline; G. Reichardt, personal communication). During the commissioning of module A, a vacuum window was placed before the entrance slit to allow the alignments of the optical elements inside the UHV monochromator. The CaF2 window was damaged within an hour of irradiation whereas the fused silica, although less transparent than CaF2, lasted longer. This is because the footprint was different as the CaF2 window flange (CF40) was closer to the entrance slit than that of the fused quartz flange (CF100). The sapphire window for module B shows three irradiated areas with different amounts of colour centre formation (right side of the figure). After about six months of use a clear area is selected by moving the UHV XY stage to which the vacuum window is attached. This device is a key and novel technical feature of B23. For both modules, before the monochromator, a shutter (Fig. 1) is operating in the open position only during measurements, avoiding unnecessary irradiation and hence damage of the optic elements (vacuum windows and monochromator's gratings and mirrors).