Desy deal by ihuangpingba

VIEWS: 0 PAGES: 15

									   Tuning and characterization of short electron
       and FEL radiation pulses at FLASH
        during shifts 19(a)-21(m).01.2011

             E. Schneidmiller and M. Yurkov (SASE & MCP)
C. Behrens, W. Decking, H. Delsim , T. Limberg, R. Kammering (rf & LOLA)
             N. Guerassimova and R. Treusch (PGM & GMD)
                                Characterization techniques
                                                                            2




        • Electron pulse: LOLA (pulse shape), toroids (charge), pyro
          detectors (signal related to bunch shape and charge).

        • Photon pulse: Pulse energy (GMD and MCP), measurements
          of statistical fluctuations (MCP), spectral measurements
          (PGM).

        • A lot of data has been recorded. Here we present only brief on-
          line analysis and preliminary conclusions.




DESY, Beam Dynamics Meeting, January 24, 2011
E.A. Schneidmiller, M.V. Yurkov
                    Statistical fluctuation method
                                                                        3




                                                     Ackermann et al.,
                                                     NaturePhoton.1(2007)336




DESY, Beam Dynamics Meeting, January 24, 2011
E.A. Schneidmiller, M.V. Yurkov
                     Statistical fluctuation method
                                                                               4



• Fluctuations / instability of the machine / electron beam parameters also
  contribute to the fluctuations of the radiation pulse energy.
• Then one needs a lot of statistics to eliminate those contributions (gating);
  measurement may take from few minutes to one hour; data analysis takes a
  while and needs experts
• At saturation the pulse is usually longer than in exponential gain regime; so,
  the method gives a lower estimate for pulse duration at saturation
• Very useful in combination with other methods (single-shot spectra etc.) and
  start-to-end-simulations
• Method was used at TTF1 and FLASH, recently also at LCLS




DESY, Beam Dynamics Meeting, January 24, 2011
E.A. Schneidmiller, M.V. Yurkov
                     Single-short spectra
                                                                             5



• In an ideal case (monochromatic electron beam) and linear mode of SASE
  operation average number of spikes in spectrum is nearly the same as the
  number of spikes (wavepackets) in the time domain.
• Width of the spike in the spectrum is inversely proportional to the pulse
  length.
• Thus, qualitative analysis of spectra measured in the linear regime may give
  a hint for qualitative estimation of pulse length and number of modes in the
  radiation pulse.




DESY, Beam Dynamics Meeting, January 24, 2011
E.A. Schneidmiller, M.V. Yurkov
                     Results: pulse energy and number of modes
                                                                                    6

• We tuned SASE @ 14 nm to maximum radiation energy with different charges 150
  pC, 250 pC, and 500 pC.
• Pulse energy level at the full undulator length was:
•                            150 pC             25-35 uJ
                             250 pC             35 uJ
                             500 pC             >200 uJ
• Then SASE process has been suppressed in the undulator modules 5 and 6 (by
  means of orbit kick) in order to deal with linear regime of SASE FEL operation.
• Characterizations for the number of modes has been performed with statistical
  measurements using MCP07 detector. Measured number of modes in the linear
  regime:




DESY, Beam Dynamics Meeting, January 24, 2011
E.A. Schneidmiller, M.V. Yurkov
                     Results: number of modes and pulse duration
                                                                                            7



• Measurements of fluctuations in the end of the linear regime gives us an estimate for
  the number of the radiation modes (spikes) M in the pulse.
• This allows to estimate radiation pulse length as T_rad ~ M x coherence time. We
  estimate coherence time as 5 fs (similar to our estimate in 2007 Nature Photonics
  paper).
• Pulse lengthening in the nonlinear regime is defined by two effects. The first one is
  slippage estimated to be about 10 fs when saturation occurs in the middle of the 5th
  undulator module. The second effect relates to the lasing to saturation of the low-
  current tails of the bunch. Pulse lengthening in this case depends on the shape of the
  tails. For gaussian-like bunch shape we estimate pulse lengthening by 30% to 50%.
  Thus, we have the following results:

   150 pC                          sigma = 60%   M = 2.8   T_rad ~ 15 fs (+10 fs + 7 fs)
   250 pC                          sigma = 59%   M=3       T_rad ~ 15 fs (+10 fs + 7 fs)
   500 pC                          sigma = 29%   M = 12    T_rad ~ 60 fs (+10 fs + 30 fs)

   Here numbers in brackets refer to the pulse lengthening in the nonlinear regime.
   Our experience tells us that practical accuracy of this estimate is about factor of 2.
DESY, Beam Dynamics Meeting, January 24, 2011
E.A. Schneidmiller, M.V. Yurkov
                     Spectra @ 150 pC, linear regime
                                                                                           8



• Statistical measurements agree qualitatively with spectral measurements .

• 150 pC                           sigma = 60%   M = 2.8   T_rad ~ 15 fs (+10 fs + 7 fs)




DESY, Beam Dynamics Meeting, January 24, 2011
E.A. Schneidmiller, M.V. Yurkov
                     LOLA images @ 150 pC
                                                9




DESY, Beam Dynamics Meeting, January 24, 2011
E.A. Schneidmiller, M.V. Yurkov
                     Spectra @ 250 pC, linear regime
                                                                                        10



• 250 pC                           sigma = 59 %   M=3   T_rad ~ 15 fs (+10 fs + 7 fs)




• Not a good agreement. Drift of the machine parameters? – There was a big time
  interval between spectral and statistical measurements.
DESY, Beam Dynamics Meeting, January 24, 2011
E.A. Schneidmiller, M.V. Yurkov
                     LOLA images @ 250 pC
                                                11




DESY, Beam Dynamics Meeting, January 24, 2011
E.A. Schneidmiller, M.V. Yurkov
                     Spectra @500 pC, linear regime
                                                                                           12



• 500 pC                           sigma = 29%   M = 12   T_rad ~ 60 fs (+10 fs + 30 fs)




DESY, Beam Dynamics Meeting, January 24, 2011
E.A. Schneidmiller, M.V. Yurkov
                     LOLA images @ 500 pC
                                                13




DESY, Beam Dynamics Meeting, January 24, 2011
E.A. Schneidmiller, M.V. Yurkov
                     How to tune SASE with short pulse
                                                                                                  14

I. Simplified procedure (1 shift):
a) Choose small charge (between 100 and 200 pC).
b) Tune gun+laser (incl. iris), rf parameters, optics and orbit for maximum pulse energy.
c) With high probability the pulses are short. If users complain, take single-shot spectra.

If estimated pulse length is too big, go to
II. More thorough procedure (2 shifts):
a) Choose small charge (between 100 and 200 pC).
b) Set up compression (if possible with the help of experts), check phase space with LOLA.
Important indication is energy chirp due to space charge (opposite to that from RF).
c) Tune SASE as in I.b).
d) Do LOLA measurement.
e) Take single-shot spectra (optionally also statistics in linear regime).
f) Iterate if necessary.

At least, our trial attempts resulted in short pulse length. Characterization of electron pulse shape
with LOLA resulted in a confined and nearly symmetric shape. Final result proving ultra-short
pulse duration is spectral measurement showing small number of spikes.


DESY, Beam Dynamics Meeting, January 24, 2011
E.A. Schneidmiller, M.V. Yurkov
                         Summary
                                                                                      15




  • Tuning of short pulses of about coherence length is possible with linearized
    compression scheme.
  • For radiation wavelength of 14 nm short pulse mode of operation is achieved in the
    range of bunch charges 150 pC – 250 pC. Even without thorough optimization
    saturation pulse energy is 25 – 35 uJ – similar to that reachable with the roll-off
    compression scheme during past years.
  • Stability of the SASE level in the saturation was nearly the same as during typical
    user runs with the roll-off compression scheme (about 35% at 150 pC). It can be
    improved significantly (down to 25%) by stabilization of the bunch charge – it
    fluctuated a lot during our measurements, about 7 % rms at 150 pC.
  • Readings of pyro detectors were too noisy at small charge. Dedicated tuning of pyro
    detectors for small charges will help a lot.
  • We should find more dedicated time to pass procedure for the short-pulse-SASE
    tuning and characterization jointly with accelerator and photon experts for different
    wavelengths (especially long). Goal: tuning dedicated regimes for users
    requesting short pulses at specific wavelengths.

DESY, Beam Dynamics Meeting, January 24, 2011
E.A. Schneidmiller, M.V. Yurkov

								
To top