EIS-TIMER setup
The installation of the EIS-TIMER photon transport system is almost completed and integrated with the experimental chamber. We commissioned a first pair of FEL pump branch-lines (crossing angle 18.4o). An optical-UV probing laser is available and can be used to collect TG and CRS data. The implementation of the FEL probing is foreseen in 2016. User operation will start on 2017.
Basic scheme of the instrument
Sketch of the optical layout of EIS-TIMER (top view)
The beam exiting from the photon analysis delivery and reduction system (PADReS; not shown) is deflected towards the EIS-TIMER beamline by a switching plane mirror (SM). Then the beam impinges into a plane mirror (PM1) working as a wavefront-division beam splitter, i.e. it intercepts just half of the beam diameter. One half of the beam is reflected upwards while the other half can propagate freely. The latter beam is further split, in the horizontal plane, by a second wavefront-division beamsplitter (PM2). The two FEL beams emerging from PM2 are used as pump pulses (red lines in the above picture) and are focused at the sample position by a pair of toroidal mirrors (TM's). Four pairs of such mirrors can be inserted and removed from the beam paths in order to change the crossing angle (2θ), and hence the transient grating vector:
k=(4π/λFEL)sin(2θ/2),
where λFEL is the pump wavelength. In the present configuration the possible values of 2θ: 18.4o, 27.6o, 79o and 105.4o. The time delay between the crossed FEL pulses can be adjusted in the ~7 ps range (see bottom-left sketch). In case of double-pulse operation, this also permit to easily delay the two train of two pulses in order to overlap the first pulse coming from one side with the second one coming from the other side in order to perform coherent Raman scattering (CRS) experiments. The third FEL probing pulses comes from the beam reflected upwards by PM1, which is made coplanar to the FEL pump beams by two additional plane mirrors PM3 and PM4 (not shown) and then goes through a delay line equipped by 4 multilayer mirrors (ML) at 45o incidence and 1 inch diameter. At present there are 4 sets of 4 MLs (designed to reflect the FEL radiation at λpr=17.85, 13.3, 6.72 and 3.17 nm) that can be interchanged and select the wavelength of the probing photons; other wavelengths might be possible upon request or the users can provide their own set of ML mirrors. The maximum time delay depends on the chosen value of 2θ and vary from ~3.5 ns (2θ=18.4o) to ~1.5 ns (2θ=105.4o). The intensity of the probe pulse emerging from the delay line is reduced (ideally) by a factor 10 to 100 and it is focused onto the sample by other 4 (insertable) toroidal mirrors placed in a way such to met the TG phase matching conditions (Bragg diffraction) for λFEL=3λpr, so that the 3rd harmonic of the FEL emission can be used as a probe. The FEL optical layout is designed to provide focal spots of about 200x100 um2 and a spatial profile that somehow optimize the overlap conditions (see bottom-right sketch). The combination of such (2θ,λFEL)-values permits to map the 0.03-1.1 nm-1 range.
IMPORTANT NOTE: the full set of wavelengths cannot be exploited in a single run, since λFEL=3λpr=53.55, 39.9 and 20.16 nm are in the range of the FEL1 source while λFEL=3λpr=20.16 and 9.51 nm are in the range of the FEL2 source, and both sources are not available for a single experiment (typically one week long). An alternative solution, presently under evaluation, is to use only the FEL2 source and exploit the radiation of the first stage to cover the wavelength range of FEL1. In this case the probe radiation would be the output of the second stage of FEL2, which, as a further advantage, is much more brilliant than the 3rd harmonic.
Optical probe option
As an alternative to the FEL probing, the TG can be probed by an ultrafast optical pulse (Δt~100 fs, down to ~40 fs in the next future) at wavelength λlaser=785, 393 or 261 nm (see bottom-left sketch); the installation of OPA and/or NOPA devices for probing pulses tunable in the UV-visible range are planned and might be provided in 2017 upon request from users. The angle of incidence of the optical pulse with respect to the bisector of the crossed FEL beam (θB) can be tuned in the ~40o-70o range (NOTE: in this case 2θ-values of 79o and 105.4o are not available) or set at 0o, different angles are possible upon request. A breadboard with a quite large space (~0.5 m2) is installed near the end-station to host optical setups requested by users (see bottom-right layout). At present the TG signal is detected in trasmission geometry (NOTE: this requires samples transparent to the probe pulse) by an in-vacuum Peltier cooled CCD detector (PI-MTE, Princeton Instruments). Single elements detectors are also availble, as well as solid state filters, slits and a focusing lens to reduce the diameter of the signal beam at the detector(s). In the next future TG detection in reflection geometry will be implemented (see bottom-left sketch) in order to probe samples opaque to the probe. Multiple probing pulses, alternative ways to probe the TG (e.g. by using high-harmonic generation devices) and heterodyne detection are under evaluation, proposals and suggestions are very welcome!
IMPORTANT NOTE: using the optical laser as a probe the phase matching for TG (Bragg diffraction) is forbidden for FEL wavelenght shorter than ~45 nm, even exploiting the smallest value of 2θ (18.4o) and the larger incident angle / shorter wavelength (θB ~ 70o / λlaser ~ 260 nm) of the probing laser. This limits the maximum grating wavevector to k~0.05 nm-1. On the other hand, using mini-Timer this limitation is relaxed in light of the much lower 2θ-values (order of a couple of degrees) achievable by this setup. This allows the fulfilment of the phase matching basically in the whole range of FERMI (i.e. down to λFEL~4 nm), still being the grating vector lower than k~0.05 nm-1. The main advantage of using the optical probe instead of the FEL probe is the possibility to fully exploit the wavelength tunability of FERMI, since in this case the limitation of using just some selected λFEL-values (determined by the ML of the FEL delay line) is removed. TG and CRS experiment with wavelength tunability are hence possible and may allow to resonantly excite a large array of samples.
Hints for proposals
- If you want to use the mini-TIMER setup, you have to ask for the DiPROI beamline (not EIS-TIMER!) and to state in the field "other requirements" that you need mini-TIMER.
- During the allocation phase we may evaluate the possibility to schedule the experiments on EIS-TIMER in place of mini-TIMER or vice-versa. If you need to use a specific instrument please stress the motivations on the proposal.
- The "standard duration" of an experiment is 16 shifts (1 shift = 8 hours), we reccomend to try to fit your experimental plan in 16 shifts. The facility usually adds other 4 shifts for FEL and beamline preparation just before each user experiment (in some cases we may add more preparation shifts), so if you ask for 16 shifts your experiment will likely be 20 shifts long. Please consider that the time needed to align mini-TIMER or EIS-TIMER is longer than 2 shifts, and any changes in the experimental geometry (FEL crossing angle, angle of incidence of the probing laser, etc.) will take time...
- It is not possible to operate FEL1 and FEL2 sources in the same beamtime, If you need wavelengths from both sources please write two proposals. Keep in mind that in this case the two experiments cannot be scheduled one after the other (most likely there will be some to several weeks in between). To avoid this one may evaluate the possibility to use the first stage of FEL2 to cover the wavelength range of FEL1 (please contact the beamline scientist for more info).
- Up to now we only used narrow bandwidth and vertically polarized FEL light (the beams cross in the horizontal plane). Narrow bandwidth implies a modest reduction in the maximum energy per pulse, while vertical polarization imposes constraints in the wavelength tunability. Consider that TG and CRS data can be collected at low pulse energy, e.g. we observed a decent TG signal from Si3N4 (at mini-TIMER, FEL wavelength ~27 nm) with ~0.5 uJ at the source.
- The ML of the EIS-TIMER delay line reflect only vertically polarized light, if you need the EUV probing clearly state in the proposal that the probe beam has to be vertically polarized.
- If you need two-color FEL emission (e.g. for CRS) state that clearly in the proposal and consider that the total energy radiated in the two pulses (I1+12) is lower than in single pulse mode. Also keep in mind that using our split-delay-recombination system (see the Research page for some more details) half of total emitted energy (I1+12) goes into the pre- and post-pulses. The time separation between the two-color pulses is typically in the 0.3-0.5 ps range while the maximum relative separation in the photon energy, Δω/ω=(ωFEL1-ωFEL2)/ωFEL1, is typically 0.5% of the central wavelength, e.g. 500 meV for ωFEL1=100 eV. Larger values of Δω/ω, e.g. up to 1%, are possible in some cases, if needed please ask the beamline scientist.
- Always ask for the user laser, since it is needed to adjust the timing between the FEL pulses (for this activity we prefer the fundamental harmonic: wavelength 785 nm). At EIS-TIMER there is quite a lot of space in the breadboard and around the chamber to implement additional optical setups, eventually proposed by users. The space available for additional optical setups at mini-TIMER is more limited.
- We used focal spot size in the range 100-200 x 50-100 μm (Hor x Vert). Using mini-TIMER it is possible to adjust the focal spot size and to use the AC/DC (auto-correlator/delay-creator) system, at EIS-TIMER not.
- Summary of our experience in TG/CRS: to date we carried out TG measurements on SiO2, Si3N4, diamond and BaF2 and CRS measurements on diamond and Si3N4. We used FEL wavelengths of 52 nm (EIS-TIMER) and 32 or 27 nm (mini-TIMER). In general we noticed sample damage for pulse energies larger than ~40 uJ at the source, in these condtions you burn very nice permanent gratings on your sample. We typically collected TG data with pulse energies at the source in the 2-5 uJ range, roughly twice in the case of CRS (remember that in our CRS scheme half of the total pulse energy goes into the pre- and post-pulses). To have TG data with a signal-to-noise of about 1000 we need to integrate ~1 minute per time-delay point; in CRS the signal-to-noise is of about 100 in similar conditions (same integration time and twice the total pulse energy). After some improvements in the setup and alignment procedures the overall throughput of FEL-based TG measurements has improved with respect to our first experience, where we needed ~7 minute per time-delay point and ~15 uJ at the source to have comparable signal-to-noise. It is likely that further upgrades of the setup (planned for 2016) would further improve the throughput, in particular for CRS.
