SPIE Membership Get updates from SPIE Newsroom
  • Newsroom Home
  • Astronomy
  • Biomedical Optics & Medical Imaging
  • Defense & Security
  • Electronic Imaging & Signal Processing
  • Illumination & Displays
  • Lasers & Sources
  • Micro/Nano Lithography
  • Nanotechnology
  • Optical Design & Engineering
  • Optoelectronics & Communications
  • Remote Sensing
  • Sensing & Measurement
  • Solar & Alternative Energy
  • Sign up for Newsroom E-Alerts
  • Information for:

SPIE Photonics West 2019 | Register Today

SPIE Defense + Commercial Sensing 2019 | Call for Papers



Print PageEmail PageView PDF

Lasers & Sources

The second generation of soft-x-ray free-electron lasers

The implementation of a seeding scheme has been used to improve both the temporal and spectral characteristics of emissions.
15 August 2007, SPIE Newsroom. DOI: 10.1117/2.1200708.0840

Free electron lasers (FEL) emitting hard and soft x-ray wavelengths are designed and built at various places around the world. At DESY (the German Electron Synchrotron) in Hamburg, the first FEL generating wavelengths down to 20nm is already providing services to outside users. The new facilities allow experiments with x-rays of unprecedented peak intensity, time resolution, and transverse coherence. The laser operation of these machines is based on a process called self-amplified stimulated emission (SASE) and starts out as density fluctuations in the working medium: a highly relativistic electron beam travelling through a spatially periodic magnetic field array, called an undulator. The lack of mirrors with high reflectance for x-rays prevent the use of cavities, which are commonly used for lasers in the visible and near ultraviolet. For this reason, the laser amplification is driven to saturation within a single pass of the electron beam through the undulator. As a result, the output of the SASE FELs consists of several modes and shows a highly statistical, spiky structure.

Seeding with an external laser is one way to improve the poor longitudinal coherence of the SASE-FEL's output. In a process called high-gain harmonic generation (HGHG),1 the seeding wavelength is reduced by a factor three or five. This has been demonstrated by Yu and coworkers at the Brookhaven National Laboratory for a seed wavelength of 800nm.2 In a future light source, a cascaded arrangement of up to four HGHG stages allows the conversion of seed wavelengths of 230nm down to 1.2nm.3 The benefits of the seed process on the output is shown in Figure 1, where the expected temporal pulse structure of a SASE-FEL is compared to a HGHG-FEL. The HGHG device has a superior temporal and spectral behaviour as compared to the SASE-FEL, and it is expected that many experiments can be performed without the need of further pulse shaping or monochromatization. Nevertheless, SASE-FELs deliver significantly higher peak powers, which will be of benefit to some experiments.

Figure 1. Output of SASE (top) and HGHG (bottom) FELs calculated with the GENESIS simulator.4

Figure 2. Schematic of the high-gain harmonic generation principle with modulator, dispersive section, and radiator.

The setup for a single HGHG-stage is shown in Figure 2. Electron bunches, with energies in the GeV range and an energy spread below 1%, are generated in a superconducting linear accelerator at repetition rates of up to 25kHz (not shown). Each bunch enters the undulator, here called modulator, with a magnetic period of λU, where it is overlapped with the short-pulse seeding laser with wavelength λL. The laser modulates the electron energy with the periodicity of the laser wavelength if the resonance condition

is fulfilled. Here γ denotes the electron Lorentz factor and K the undulator parameter.

The resonance condition can be tuned by varying the electron beam energy but, in most cases, it is preferred to change the K-parameter (1 < K < 15). This is accomplished by a simple variation of the vertical separation between the upper and lower magnetic structure of the modulator.

After the modulator, the electron beam enters the dispersive section, where its energy modulation is transformed into a density modulation of the same periodicity, and the properties of the seed pulse are imprinted onto the electron beam. The subsequent radiator, a device similar to the modulator, is tuned to a specific higher harmonic, h, of the density modulation: it emits, therefore, at a fraction of the seed wavelength. The output of this radiator can in turn be used to seed the next HGHG stage. By cascading up to four such stages, the initial laser wavelength can be down-converted to the desired target wavelength. A final amplifier is used to bring the amplification process into saturation at the final wavelength.

Unfortunately, the interaction between the laser seed field and the electron beam degrades the electron bunch quality by increasing its energy spread, and the part of the electron bunch that was involved can no longer be used in subsequent HGHG stages. In the ‘fresh bunch’ technique, subsequent HGHG stages use an, until then, unused part of the electron bunch.5 This is achieved by delaying the electron bunch with respect to the seed pulse in a magnetic chicane between two HGHG stages. Seeding in subsequent stages is done at undisturbed parts of the electron bunch, which means that the bunch must be significantly longer (ca. 700fs, flat top) than the initial seeding pulse (ca. 20fs, rms). This seeding scheme ensures an inherent temporal correlation between x-ray pulse and initial seed laser. It eliminates the jitter that the electron beam gains in the linear accelerator and makes it possible to perform pump probe experiments with both pulses.

After the successful operation of the HGHG scheme at Brookhaven, the STARS project at BESSY will demonstrate the feasibility of a cascaded HGHG-FEL with two stages down to wavelength of 40nm. In the next step, a user facility for the wavelength range from 50nm to 1.2nm is planned at the BESSY site.

Rolf Follath
Berlin, Germany