1. Beam Loss Issues for High-Power, High-Brightness Electron Injectors Bruce Dunham Cornell University December 18, 2012

2. What is important for running high currents? Halo is a major problem (tuning, radiation shielding and machine protection) Beam dump monitoring and protection Fast shutdown – want to block the laser before anything else trips... Catching transients (due to FE, ions, scattering,...) for troubleshooting RF trips (mostly due to coupler vacuum) Feedback for bunch charge, laser position and beam orbit Current measurement Measurements of RF response to the beam, HOM’s Monitoring HV power supply ripple and frequency response Vacuum monitoring, fast and slow Personnel protection Overall machine stability Bruce Dunham Cornell University High Current Operation and Diagnostics

3. field emission from the cathode field emission from the gun electrodes discharges from the gun insulator stray light reaching the cathode (big problem for high QE cathodes) room lights, scattered laser light x-rays/UV light from SRF cavities x-rays/UV from gun electrode discharges field emission from SRF cavities space charge aberrations non-uniform laser which makes long tails in time or space ghost pulses from the laser, cathode response time too long which produces tails in time electron scattering Causes of Halo Bruce Dunham Cornell University

4. Dynamic Range and Halo 100 mA 100  A 100 nA100 pA Design current: Non-intercepting or minimally intercepting diagnostics Phase space measurements: Non-intercepting or fully intercepting diagnostics Beam setup: viewers, cameras Halo, cathode lifetime: viewers, PMT’s 10 3 10 6 to 10 9 Bruce Dunham Cornell University

5. Halo/Beam Loss Measurement We are very interested in measuring halo, but our methods are crude and qualitative Measurements

6. Halo Measurement A viewer with a hole for imaging halo Simple! ~10 5 range But a little dangerous for high currents... Arrays of PMTs and GM tubes Bruce Dunham Cornell University Can be confusing as we can get radiation from the dump, which is close by

7. Bruce Dunham Cornell University To ‘calibrate’ the beam loss due to radiation measurements – steered a 50 pA, 5 MeV beam onto a beam pipe Radiation Gamma probe 50 pA/5 MeV produces about 10 mSv/hr at an external gamma probe For the best beam setup, we typically observe from 0 to 40 mSv/hr along the length of the beamline. From this, we estimate a total beam loss of ~1 nA out of 50 mA, or 2x10 -8 50 pA

8. Bruce Dunham Cornell University Vacuum pressure near the gun is one of the most sensitive measurements of halo (and cathode lifetime) Extractor gauge for very low pressures. Working on a fast (sub-msec) response ion gauge to look for transients Granville-Phillips now has a new type of residual gas analyzer with very fast scan times (10s of ms) Vacuum Diagnostics

9. A large quadrant detector in front of the beam dump is used to monitor ‘halo’, but is mainly for machine protection Large BeO viewscreen at the dump location Beam image beyond the raster during calibration A bpm located after the raster for continuous monitoring A large quadrant detector before the dump ensures that the beam cannot get too big Beam Dump Monitoring Bruce Dunham Cornell University

10. Controlling ‘stray’ light from the laser Any photons reaching the active area of the cathode (other than the desired laser beam), cause unwanted electrons. Both transverse and longitudinal (time) must be considered. Stray Light

11. Halo – Vacuum Laser Mirror Vac. mirror: ~5% in halo Normal mirror: ~0.1% in halo Image on the cathode using normal dielectric mirror Our final laser mirror scattered ~50x more light compared to dielectric mirrors (which we cannot use). A new mirror with 2 nm-rms surface roughness was installed to fix the problem. Still see light 20 meters away from the gun. Image on the cathode using coated metal mirror Following work by a group at DESY Bruce Dunham, Cornell University

12. Bruce Dunham Cornell University Laser Topics to consider: Since the cathode is not perfectly smooth, the scattered light can bounce of metal surfaces and reach the cathode – most likely at the wrong phase. This may become worse with time – cathode damage from ions and heat tends to roughen the surface. We have observed laser light 20 meters away through the beam pipe. ‘Ghost’ pulses (in time) that arrive between the main laser pulses will cause halo. It is critical to have a laser using a second-harmonic generation crystal to produce the final wavelength. The non-linear process suppresses the ‘ghost’ pulses. Cathode masking (next slide) also it useful to eliminate transverse laser halo.

13. Bruce Dunham Cornell University Non-recoverable QE damage on GaAs at high current– can’t be recovered by heat treatment and reactivation Cause of Damage? Ion Backbombardment Ion implantations Rise in vacuum pressure Field emission/arcing Cathode Masking

14. Bruce Dunham Cornell University Cathode Masking We can make any pattern on GaAs cathodes. This one has four active areas away from the center. With the laser directed onto one spot, we are able to observe 3 faint halo spots 20 meters downstream (using the viewer with a hole in it).

15. Bruce Dunham Cornell University Active area (4 spots) Discoloration due to damage GaAs Laser location

16. Bruce Dunham Cornell University This is the masking geometry we use for all cathodes now – both GaAs and alkali cathodes. We could not run high current for extended times without this geometry.

17. Bruce Dunham Cornell University CsK2Sb, 50 nm pk-pk

18. Bruce Dunham Cornell University We have to tune the optics to reduce beam loss. Unfortunately, this moves away from the minimum emittance optics. Use the first two solenoids to minimize beam loss between the gun and the cryomodule, and at the exit of the cryomodule. Use the four quadrupoles after the cryomodule to minimize the beam envelop size between the cryomodule and the dump. The beam is very large due to space charge, making things difficult. We have been running at lower gun voltages (250 to 350 kV) to reduce field emission, but eventually we will need to run at 400+ kV to keep the beam size smaller. Beam Optics Issues

19. Bruce Dunham Cornell University Other Sources of Halo All sources of light (visible, UV or x-rays) that reach the cathode can produce halo. At UV and x-ray wavelengths, the QE will be much higher than for visible light (~100%) Cathode roughness – the more a cathode is used, it may become rougher due to various mechanisms. Sharp points can cause field emission electrons, which can strike the vacuum chamber, increase the pressure, and produce light Electrode field-emission – same results Insulator charge build-up – stray electrons that land on the insulator will eventually discharge, producing UV light

20. Bruce Dunham Cornell University Other Sources of Halo Buncher tuners – the mechanical frequency tuners on the buncher cavity can generate particles when it moves, increasing the pressure and potentially producing RF multipactoring Couplers – arcs and field emission can produce light Cavities – field emission and multipactoring can produce light and particles We believe that many of these processes produce large bursts of photoemitted electrons at the cathode, and lead to trips of the machine protection system. Each trip produces visible damage on the cathode surface (see next page). There may also be a positive feedback mechanism which leads to bursts of current.

21. Bruce Dunham Cornell UniversityB. Dunham Cornell University Front surface of the cathode (CsK 2 Sb on Si) after use. Back surface of the cathode puck after use. Cathode Damage Active area is offset from the center Ion damage limited to the central area

22. Bruce Dunham Cornell University Cathode trajectories Initial Simulations

23. Bruce Dunham Cornell University Conclusion We have operated up to 65 mA so far... Careful attention to removing known sources of halo Estimate beam loss to be 10 -8 to 10 -7 range at 50 mA Better diagnostics needed! Further work on modeling coming soon

24. This work is supported by the National Science Foundation grant DMR-0807731 Support Bruce Dunham Cornell University