The European XFEL is a hard X-ray free-electron laser (FEL) based on a high-electron-energy superconducting linear accelerator. The superconducting technology allows for the acceleration of many electron bunches within one radio-frequency pulse of the accelerating voltage and, in turn, for the generation of a large number of hard X-ray pulses. We report on the performance of the European XFEL accelerator with up to 5,000 electron bunches per second and demonstrating a full energy of 17.5 GeV. Feedback mechanisms enable stabilization of the electron beam delivery at the FEL undulator in space and time. The measured FEL gain curve at 9.3 keV is in good agreement with predictions for saturated FEL radiation. Hard X-ray lasing was achieved between 7 keV and 14 keV with pulse energies of up to 2.0 mJ. Using the high repetition rate, an FEL beam with 6 W average power was created. The first operation of the European X-ray free-electron laser facility accelerator based on superconducting technology is reported. The maximum electron energy is 17.5 GeV. A laser average power of 6 W is achieved at a photon energy of 9.3 keV.

Coupling between the transverse plains leads to an in-crease of the horizontal and vertical electron beam emittances. The coupling can be measured with dedicated multi quadrupole scans while the correlations of the beam are ob-served on a screen. In this paper we show the results from first coupling studies in the European XFEL injector. for coupling Additional information can be found paper, we present measurements of the transverse in the European XFEL injector.

The SwissFEL Injector Test Facility operated at the Paul Scherrer Institute between 2010 and 2014, serving as a pilot plant and testbed for the development and realization of SwissFEL, the X-ray Free-Electron Laser facility under construction at the same institute. The test facility consisted of a laser-driven rf electron gun followed by an S-band booster linac, a magnetic bunch compression chicane and a diagnostic section including a transverse deflecting rf cavity. It delivered electron bunches of up to 200 pC charge and up to 250 MeV beam energy at a repetition rate of 10 Hz. The measurements performed at the test facility not only demonstrated the beam parameters required to drive the first stage of an FEL facility, but also led to significant advances in instrumentation technologies, beam characterization methods and the generation, transport and compression of ultra-low-emittance beams. We give a comprehensive overview of the commissioning experience of the principal subsystems and the beam physics measurements performed during the operation of the test facility, including the results of the test of an in-vacuum undulator prototype generating radiation in the vacuum ultraviolet and optical range.

Minimizing the dilution of the electron beam emittance is crucial for the performance of accelerators, in particular for free electron laser facilities, where the length of the machine and the efficiency of the lasing process depend on it. Measurements performed at the SwissFEL Injector Test Facility revealed an increase in slice emittance after compressing the bunch even for moderate compression factors. The phenomenon was experimentally studied by characterizing the dependence of the effect on beam and machine parameters relevant for the bunch compression. The reproduction of these measurements in simulation required the use of a 3D beam dynamics model along the bunch compressor that includes coherent synchrotron radiation. Our investigations identified transverse effects, such as coherent synchrotron radiation and transverse space charge as the sources of the observed emittance dilution, excluding other effects, such as chromatic effects on single slices or spurious dispersion. We also present studies, both experimental and simulation based, on the effect of the optics mismatch of the slices on the variation of the slice emittance along the bunch. After a corresponding reoptimization of the beam optics in the test facility we reached slice emittances below 200 nm for the central slices along the longitudinal dimension with a moderate increase up to 300 nm in the head and tail for a compression factor of 7.5 and a bunch charge of 200 pC, equivalent to a final current of 150 A, at about 230 MeV energy.

Total beam delivery time to user stations is a key parameter for FEL user facilities. Therefore downtime due to RF issues, among other things, should be minimized. Specifically in case of a RF failure machine operation and beam delivery should be maintained as long as the next scheduled maintainance day. This is achieved by increasing the power in all remaining klystrons to recover the lost beam energy. These modification of the beam energy profile along the machine induces an optics perturbation which is typically compensated by a rescaling of the quadrupole magnet gradients to maintain a constant focusing strength. However, we would like to resume operation at the next macro-pulse after a RF event. While, in general, the RF systems can handle such changes the magnets can not. In this paper we will explore optics perturbations for the case that we do not change the magnets at all, to estimate the feasibility of fast beam recovery after klystron failure. In addition corrections to the RF setup are calculated with the goal of avoiding changes in the bunch compression dynamics of the machine. BEAM ENERGY MANAGEMENT FOR THE EUROPEAN XFEL Superconducting technology used at the European XFEL allows for RF pulses as long as 600 μs supporting bunch trains with an internal repetition rate of up to 4.5 MHz. These pulses are, which are refereed to as macro-pulses are triggered with 10 Hz. The European XFEL is driven in total by 26 1.3 GHz multi-beam klystrons [1] [2]. These RF stations are distributed along European XFEL as shown in Fig. 1. The energy gain ∆E, number of RF stations N , and individual voltages per klystron ∆E/N and cavity V are summarized in Table 1. Each klystron in the Linac 1-3 sections drives four accelerator cryo-modules consisting of eight cavities with a total energy gain up to 755 MeV. Design gradient of the niobium cavities is 23.6 MV/m. Since the assembly of the cryo-modules is work in progress we do not have final numbers on the actual available gradient. After final testing and should the situation arise re-treatment of all modules we assume an available gradient of 23.6 MV/m with an average overhead of 10% [3]. Linac 3 is configured to achieve nominal final beam energy of 17.5 GeV at the nominal gradient of 23.6 MV/m using 20 instead of 21 RF stations as beam energy reserve. In the following the name klystron refers to the full RF station including modulator, pulse cables, pulse transformers, klystrons, waveguides, down to the cavities, and failures in each of these components are refereed to as klystron failure. From the point-of-view of electron beam energy management the machine is conveniently separated into three parts. First the Injector and Linac 1 section. In this region of the machine each section is essentially driven by one klystron. Klystron failures in this part are fatal and can not be compensated, immediate repair is required to resume operation. The second part is Linac 2. A reduction of acceleration voltage can be recovered by reserves in Linac 3. This Linac 2 however is upstream of the last bunch compressor chicane. Voltage changes effectively modify the energy chirp at BC2 and therefore the final longitudinal beam profile. In addition to energy profile reorganization the off-crest phases needs modification to maintain the final current profile. Linac 3, the main linac, is the last part. Here the majority of the beam energy is generated and here beam energy variations are corrected. The nominal energy gain of Linac 3 is 15.1 GeV. Since we only rely on 20 instead of 21 klystron stations and assume an 10% energy overhead the total voltage capacity of Linac 3 is 17.4 GeV . This additional energy reserve of about 2.3 GeV can be used to compensate the outage of about three klystron stations. Table 1: XFEL Energy Gain Configuration Linac ∆E N ∆E/N V Section [GeV] [MeV] [MV/m] Injector 0.13 1 130 16.5 Linac 1 0.57 1 570 17.8 Linac 2 1.7 3 567 17.7 Linac 3 15.1 21/20 719/755 22.5/23.6 MAIN LINAC ENERGY MANAGEMENT To redistribute the energy gain along we propose an iterative procedure. We start with an index set I which includes all klystrons used for energy correction. Typically I contains all stations except the failed one. The voltages of modules not in I ∆Vj / ∈I are not necessarily set to zero, to allow modeling of reduced gradients in individual stations, e.g. detuned cavities within a module or reduced voltage operations as quench prevention. The voltage of each module in operation is modified according to: ∆V ′ i = Enominal − ∑ j / ∈I ∆Vj |I| wi 〈wi〉 , i ∈ I. (1) The positive weight factors wi are chosen to set priorities according to the performance and reliability of the individual RF stations. After voltage scaling according to Eq. 1 each ∆Vi is compared with the individual maximum. If the maximum is exceeded it is set to this maximum and the station THP002 Proceedings of FEL2014, Basel, Switzerland ISBN 978-3-95450-133-5 672 C op yr ig ht © 20 14 C C -B Y3. 0 an d by th e re sp ec tiv e au th or s Electron Bunch Generation and Manipulation GUN ACC1 ACC39 Linac 1 Linac 2 Linac 3 BC0 BC1 BC2 LH 130MeV 700MeV 2.4GeV 17.5GeV 1x4x8 3x4x8 21(20)x4x8 1x1x8 1x1x8 Figure 1: Overview of the XFEL linac and bunch compressor sections. The beam energy at different positions along the machine is given together with the number of RF stations per linac section. For instance Linac 2 consists of 3 klystrons driving 4 modules each containing 8 cavities. Linac 3 is consists of 21 klystron stations. removed from the set I to have these stations not modified in further iterations. Such iterations continue until the nominal energy gain is retained or all available structures are set to maximum voltage. Examples of such energy profile corrections are shown in Fig. 2, 3, and 4. As mentioned earlier 0 10 20 30 40 50 60 70 80 90 0 5 10 15 module no. E [ G e V ] no spare / klystron 10 off 0 10 20 30 40 50 60 70 80 90 −5 0 5 module no. re la ti v e e n e rg y d e v ia ti o n [ % ] nominal uncorrected failure corrected failure Figure 2: Example of energy correction in Linac 3 of XFEL. In the upper plot the nominal energy profile (black solid), reduced energy after RF failure in station 10 (blue dashed), and the corrected energy profile (red solid) are shown. The relative deviation, the corresponding quadrupole field deviation, along the linac indicated by the module number is shown in the lower plot. In this example no dedicated spare is used, all cavities nominally are operated at 23.6 MV/m · 20/21 =22.5 MV/m. the main linac is driven by 21 RF stations while only 20 are required for nominal operation. Basically two options can be considered to use this reserve, either all stations are operated at 20/21 of the nominal gradient or one RF stations is not used in nominal operation and activated as needed. In the latter case this "spare" can be located at different positions along the linac. In the following we consider four scenarios, all stations are in operation, the most downstream station, the middle station or the klystron in the beginning of the main linac are deactivated. Comparing this first two options in Fig. 2 and Fig. 3 we see that in the first case we have an energy deviation at all positions while in the second case the deviation is somewhat localized. The exact shape of the energy deviation depends on the position of the broken RF station. An overview of the energy 0 10 20 30 40 50 60 70 80 90 0 5 10 15 module no. E [ G e V ] spare / klystron 10 off 0 10 20 30 40 50 60 70 80 90 −8 −6 −4 −2 0 module no. re la ti v e e n e rg y d e v ia ti o n [ % ] nominal uncorrected failure corrected failure Figure 3: As in Fig. 2. In this example a reserve klystron station is located at the end of the main linac. 0 10 20 30 40 50 60 70 80 90 0 5 10 15 module no. E [ G e V ] spare / klystron 10 off and klystron 16 at 60% 0 10 20 30 40 50 60 70 80 90 −10 −5 0 5 module no. re la ti v e e n e rg y d e v ia ti o n [ % ] nominal uncorrected failure corrected failure Figure 4: As in Fig. 3. In this example RF station 10 is off and station 16 is set to 60% of nominal gradient. deviations along the linac for the different scenarios is given in Fig. 5. The impact on the beam optics mismatch amplitude BMAG = ξ + √ ξ2 − 1 with ξ = 1/2 · (βγ0 − 2αα0 + γβ0) at the end of the linac with respect to the deign optics is summarized in Fig. 6. Overall optics mismatch amplitude is optimized in the case with all klystrons in operation and the "spare in the middle" case. In all these situations the maximum deviation from the design optics occurs if the first RF station is out of operation. In general this behavior is expected. As shown Proceedings of FEL2014, Basel, Switzerland THP002 Electron Bunch Generation and Manipulation ISBN 978-3-95450-133-5 673 C op yr ig ht © 20 14 C C -B Y3. 0 an d by th e re sp ec tiv e au th or s

The Singular Value Decomposition (SVD) method has been applied to the SwissFEL Injector Test Facility to identify and better expose the various relationships among the possible jitter sources affecting the longitudinal phase space distribution and the longitudinal diagnostic elements that measure them. To this end, several longitudinal tracking simulations have been run using the Litrack code. In these simulations the RF and laser jitter sources are varied one-byone within a range spanning several times their measured stability. The particle distributions have been dumped close to the diagnostic locations and the measured quantities analyzed. A matrix has been built by linearly fitting the response of each measured quantity to each jitter source. This response matrix is normalized to the stability of the jitter source and the instrumentation accuracy, and it is inverted and analyzed using SVD. From the eigenvalues and eigenvectors the sensitivity of the diagnostics to the jitters can be evaluated.

The SwissFEL compact accelerator design is based on extremely low emittance electron beam from an RF photoinjector. Proper temporal and spatial shaping of the photocathode drive laser is employed to reduce the space charge emittance contribution. However, the ultimate limit for the beam quality is the thermal emittance, which depends on the excess energy of the emitted photoelectrons. By varying the photocathode laser wavelength it is possible to reduce the thermal emittance. For this purpose, we applied a tunable Ti:sapphire laser and an optical parametric amplifier which allow to scan the wavelength between 250 and 305 nm. The system permits to study the thermal emittance and the quantum efficiency evolution as function of the laser wavelength for the copper photocathode in the RF gun of the SwissFEL injector test facility. The results are presented and discussed. INTRODUCTION AND MOTIVATIONS The Paul Scherrer Institute (PSI) is building an X-ray Free Electron Laser (FEL) user facility, which aims to deliver ultrashort coherent photon pulses with wavelengths ranging between 0.1 and 0.7 nm by the year 2017 [1]. For cost and space reasons the driving accelerator is foreseen with relatively modest final energy, thus calling for very low emittance. In preparation of SwissFEL, PSI is commissioning a 250 MeV photo-injector (SITF), which intends to demonstrate the generation of high-brightness electron beams and serves as a realistic test bed for crucial components for SwissFEL [2]. Modern linear accelerators demonstrated that it is feasible to preserve the electron beam emittance throughout acceleration. It becomes therefore important to generate the electron bunch at the source with the lowest possible emittance. Its growth due to the linear space charge forces is effectively counteracted by emittance compensation scheme. The photocathode drive lasers employ typically spatial and temporal pulse shaping in order to compensate the emittance dilution due to nonlinear space charge effect. Therefore the thermal or intrinsic emittance becomes a realistic limit for the beam quality. This parameter is a measure of the temperature of the electrons emitted from the cathode and it depends on the excess of energy of the photoelectrons in vacuum. Thermal emittance is function of the cathode material and surface quality, the accelerating electric field and laser wavelength. The value of the intrinsic emittance is linked directly to the quantum efficiency (QE) of the photocathode (number of emitted electrons per incident photons). In the presented work we characterize the intrinsic emittance and the QE in RF gun while varying the laser wavelength. Similar studies are reported for photocathodes in DC gun [3]. The intrinsic emittance, εin can be written as [4]: € ε in ω ( ) =σL ω −Φeff 3mc 2 (1) with ω the laser wavelength, ω the photon energy, σL the rms laser spot size and Φeff the effective work function of the copper cathode including the Schottky effect. The Schottky term accounts for the reduction of potential barrier due to the applied electric field on the cathode surface. The total emittance can be reduced by adapting the laser photon energy to the net work function of the cathode. A decrease in quantum efficiency (QE) is expected when the laser photon energy approaches the effective work function. The QE can be expressed as [4]: € QE ω ( )≈ K ⋅ ω −Φeff ( ) 2 (2) K takes into account the reflection of the laser at the cathode surface and the probability of the emission process. From equations 1 and 2 it is clear that lower εin can be obtained at the price of also lower QE. For the design of high brightness accelerator a trade-off between the maximum acceptable intrinsic emittance and the quantum efficiency need to be established. The lower QE calls for higher energy and more complex drive laser with consequent worsening of system stability and ability control the photon beam tridimensional shape.

Experimental and simulation results of an electron gun test facility, based on pulsed diode acceleration followed by a two-cell rf cavity at 1.5 GHz, are presented here. The main features of this diode-rf combination are: a high peak gradient in the diode (up to 100 MV/m) obtained without breakdown conditioning, a cathode shape providing an electrostatic focusing, and an in-vacuum pulsed solenoid to focus the electron beam between the diode and the rf cavity. Although the test stand was initially developed for testing field emitter arrays cathodes, it became also interesting to explore the limits of this electron gun with metallic photocathodes illuminated by laser pulses. The ultimate goal of this test facility is to fulfill the requirements of the SwissFEL project of Paul Scherrer Institute [B.D. Patterson et al., New J. Phys. 12, 035012 (2010)]; a projected normalized emittance below 0.4 mu m for a charge of 200 pC and a bunch length of less than 10 ps (rms). A normalized projected emittance of 0.23 mu m with 13 pC has been measured at 5 MeV using a Gaussian laser longitudinal intensity distribution on the photocathode. Good agreements with simulations have been obtained for different electron bunch charge and diode geometries. Emittance measurements at a bunch charge below 1 pC were performed for different laser spot sizes in agreement with intrinsic emittance theory [e.g. 0.54 mu m/mm of laser spot size (rms) for Cu at 274 nm]. Finally, a projected emittance of 1.25 + / - 0.2 mu m was measured with 200 pC and 100 MV/m diode gradient.

A screen monitor containing OTR foils and scintillator crystals has been designed to measure the transverse profile of electron bunches in the SwissFEL. In conjunction with quadrupole magnets in FODO cells and a transverse deflecting structure, the screen monitors will be used to measure transverse and longitudinal phase space projections of the electron pulses in the 250 MeV Injector. Tomographic methods will be used to reconstruct the phase space distributions.