Performance of a first generation X-band photoelectron rf gun

Building more compact accelerators to deliver high brightness electron beams for the generation of high flux, highly coherent radiation is a priority for the photon science community. A relatively straightforward reduction in footprint can be achieved by using high-gradient X-band (11.4 GHz) rf technology. To this end, an X-band injector consisting of a 5.5 cell rf gun and a 1-m long linac has been commissioned at SLAC. It delivers an 85 MeV electron beam with peak brightness somewhat better than that achieved in S-band photoinjectors, such as the one developed for the Linac Coherent Light Source (LCLS). The X-band rf gun operates with up to a $200\mathrm{MV}/\mathrm{m}$ peak field on the cathode, and has been used to produce bunches of a few pC to 1.2 nC in charge. Notably, bunch lengths as short as 120 fs rms have been measured for charges of 5 pC ($\sim 3\times {10}^7$ electrons), and normalized transverse emittances as small as 0.22 mm-mrad have been measured for this same charge level. Bunch lengths as short as 400 (250) fs rms have been achieved for electron bunches of 100 (20) pC with transverse normalized emittances of 0.7 (0.35) mm-mrad. We report on the performance and the lessons learned from the operation and optimization of this first generation X-band gun.

This article appears in the following collection:

Figure 1

Mark-1 5.6 cell X-band gun. (a) Photo of the X-band gun: an S-band gun cathode is shown on the right for comparison, (b) CAD model of the gun, (c) cross section of the gun (d) the pi-mode longitudinal field (${\mathrm{E}}_{\mathrm{z}}$) along the axis.

Figure 2

X-Band test area beamline. The 1.05 m long copper accelerator structure is seen in the middle of the photo, and one of three phase shifters is visible on the wall. The apparent curvature of the beamline is an artefact of the wide-view camera lens.

Figure 3

Schematic of the XTA. PS1, 2, and 3 are rf phase shifters, YAGS is an yttrium aluminum garnet screen, OTRS is an optical transition radiation screen, TCAV is an X-band deflecting cavity (powered by a different RF source) and RMO is the RF master oscillator.

Figure 4

Three high power phase shifters: the top two (copper cylinders with bellows on each end) are used to adjust the power going to the gun, and the third phase shifter (below the other two) is installed on the branch line that feeds the linac structure (in the photo, it is obscured by a metal support plate).

Figure 5

Schematic of XTA laser system.

Figure 6

(a) Evolution of $\gamma$ with time for a single electron in a $200\mathrm{MV}/\mathrm{m}$ X-band gun (blue) and a $120\mathrm{MV}/\mathrm{m}$ S-band gun (magenta) for different initial rf phases, (b) corresponding longitudinal electric field seen by the electron and (c) distance from cathode as a function of time.

Figure 7

RMS bunch length versus laser-to-rf phase for different laser pulse durations and spot sizes, but a fixed 30 fC charge (i.e., $2\times {10}^5$ electrons) calculated at a distance 0.2 m from the cathode.

Figure 8

Results from charge parameter scans as a function of the laser spot size radius for 20 pC bunches where the blue crosses are for X-band ($200\mathrm{MV}/\mathrm{m}$ cathode field) and the red circles are for S-band ($120\mathrm{MV}/\mathrm{m}$ cathode field): (a) minimum 95% transverse emittance, (b) minimum bunch length, (c) maximum peak brightness.

Figure 9

Results from 100 pC bunch charge parameter scans as a function of the laser spot size radius where the blue crosses are for X-band ($200\mathrm{MV}/\mathrm{m}$ cathode field) and the red circles are for S-band ($120\mathrm{MV}/\mathrm{m}$ cathode field}: (a) minimum 95% transverse emittance, (b) minimum bunch length and (c) maximum peak brightness.

Figure 10

Cathode surface images taken through the extraction mirror (opposite to the injection mirror) using a camera just outside of the vacuum chamber. (a) This image was taken using visible light after the gun was rf conditioned but before the first photoelectron beam was generated. The surface already appears damaged at this stage. (b) Cathode surface imaged with visible light after 2 years of operation of the gun. The center circle of bright spots fills the entire 7 mm diameter cathode. (c) Zoomed area of the cathode image shown in (b).

Figure 11

Integrated dark current as a function of cathode gradient with 147 ns long rf pulses measured 0.56 m from the cathode.

Figure 12

Dark current as a function of gun temperature at two different SLED-II power levels with 147 ns pulses.

Figure 13

Calculated single electron energy from the X-band gun versus the cathode rf phase at the laser arrival time for various cathode gradients. Zero phase corresponds to the rf zero-crossing.

Figure 14

Bunch energy versus rf-to-laser phase measured after the gun at different cathode gradients using either the horizontal (data points with crosses) or vertical corrector magnets (data points with circles). The gradient is based on the SLED-II power level with the expectation that 150 MW, corresponding to 15.3 MW at the gun input, yields a $190\mathrm{MV}/\mathrm{m}$ cathode field (as discussed below).

Figure 15

Simulated RTOA versus rf phase relative to the laser arrival for different cathode gradients.

Figure 16

Measured relative time of arrival (converted to phase at 11.424 GHz) versus rf phase relative to the laser arrival for different SLED-II power levels. The RF pulse length was 140 ns.

Figure 17

(a) RTOA measurements at 150 MW compared with simulations at gradients of $190\mathrm{MV}/\mathrm{m}$ and $200\mathrm{MV}/\mathrm{m}$, (b) zoom-in of the data near the RTOA minimum. RMS fluctuations were calculated using the 50 sample points.

Figure 18

Simulated bunch lengths 0.2 m from the cathode versus bunch charge. Also shown are power law fits to these results.

Figure 19

Simulated bunch lengthening (a) in the drift region after the gun for different bunch charges and densities assuming a laser pulse having a 0.5 mm spot radius and 100 fs FWHM Gaussian longitudinal distribution. Two charge densities were simulated by changing the gun solenoid focusing strength. (b) The corresponding rms transverse beam sizes.

Figure 20

(a) Streaked beam profile, and (b) beam profile with the deflector off.

Figure 21

Bunch lengths measured at the transverse deflector versus bunch charge, and simulated data as described in the text. Two additional simulated data points, in purple, show the bunch length when a 20-cm long RF compressor is used.

Figure 22

Example of the signal (in orange) from the pyro-detector.

Figure 23

Variation of the pyro-detector signal with the square of the bunch charge.

Figure 24

Electron beam profile on the OTR screen for a bunch with low charge: (a) image on the screen and (b) Gaussian fits.

Figure 25

Beam size versus quadrupole strength: (a) rms beam size obtained from Gaussian fits, (b) rms beam size obtained using a 5% area cut. The blue (red) curves are fits to the horizontal (vertical) data. The rms emittances are 0.32 mm-mrad and 0.28 mm-mrad, respectively, in the horizontal and vertical plane for the rms fits. They are 0.36 mm-mrad and 0.28 mm-mrad, respectively, in the horizontal and vertical plane for the Gaussian fits.

Figure 26

Emittance calculated from the horizontal data (rms with area cut) shown in Fig. 28. after deconvolving the optics system resolution: (a) beam size versus quadrupole strength fits with the data corrected for various resolution contributions and (b) the resulting emittances versus the resolution.

Figure 27

Quadrupole scans with 30 pC bunch charge: (blue) horizontal plane and (red) vertical plane. The fits give a normalized horizontal emittance of 0.47 mm-mrad and a normalized vertical emittance of 0.43 mm-mrad.

Figure 28

Emittances computed from the 30 pC data as a function of the truncation level used in computing the rms bunch sizes. The Gaussian fit results are plotted at 0% truncation. The blue points are from the horizontal scans and the red points are from the vertical scans.

Figure 29

Quadrupole scans and examples of bunch profiles for 100 pC bunches. The resulting normalized emittances are 1.09 mm-mrad and 0.81 mm-mrad, respectively, from the fits to the horizontal data in plot (c), and the vertical emittance is 0.66 mm-mrad from the fit to the data in plot (f).

Figure 30

Emittance analysis of the 100 pC data in Fig. 29 as a function of truncation level used in computing the rms bunch sizes. The Gaussian fit results are plotted at 0% truncation. The blue points are from the horizontal scans and the red points are from the vertical scans.

Figure 31

Typical QE map of the cathode. The length scale at the cathode was determined from the mirror actuator calibration and its distance from the cathode.

Figure 32

Transverse stability of the electron beam measured on 500 consecutive shots (during a 50 second period with a 10 Hz pulse rate). (a) Vertical, in red, and horizontal, in blue, rms beam sizes, (b) horizontal position and (c) vertical position.

Figure 33

(a) Simulated trajectories along beamline for various changes in the rf field amplitude—the reference trajectory corresponds to a $200\mathrm{MV}/\mathrm{m}$ cathode field, a laser offset of $300\mu \mathrm{m}$ and an $88\mathrm{MV}/\mathrm{m}$ acceleration gradient. (b–d) Zoomed-in views of the trajectories at two z locations and the beam energy variation just after the gun.

Figure 34

Energy stability measured from the beam centroid jitter on the spectrometer screen: (a) 50 samples for each spectrometer magnet setting and (b) energy jitter relative to the nominal 80 MeV beam energy.