1. Preparation/Characterization of Atomically Flat and Clean Mo (100) Surfaces and Thermal Emittance/Response Time Measurements of Cs 3 Sb Photocathodes Matt Nichols Advisor: Luca Cultrera

2. Introduction Energy Recovery LINAC: DC gun operates on laser based photoemission of electrons Performance of ERL related to brightness of electron beam Beam brightness is limited by intrinsic properties of the photocathode in the DC gun Want 100 mA average current and 2 ps bunch length To achieve this, we need cathodes with high QE (in visible), small response time, low thermal emittance, and long lifetime

3. Introduction Cont… Need QE> 5% (in visible), response time < 1 ps, and MTE of ~120 meV GaAs photocathodes work well, but have a short lifetime (hours) and need pressure of 10 -11 Torr CsK 2 Sb photocathodes have recently shown similar success but with added effect of longer lifetime CsK 2 Sb cathodes grown on Si (100) substrates here at Cornell have shown QEs that are a factor of ~2 lower than expected (a) Quantum Efficiency vs. Wavelength for the CsK 2 Sb photocathode

4. Substrate-Cathode Interface The substrate may actually play a more important role in photocathode properties than originally thought Mismatch of the Si and CsK 2 Sb crystal structures may lead to various crystal defects: line defects, plane defects, etc… These defects could lower the mean free path of excited electrons by acting as recombination centers Lower mean free path means lower QE It has been proposed that using Mo (100) (Molybdenum) substrates could improve the crystalline lattice mismatch to as low as 3% Mo is a body centered cubic crystal with a lattice parameter of 3.14 Å The lattice parameter of CsK 2 Sb is 8.61 Å

5. Substrate-Cathode Interface Cont… Matching the lattice parameters of the substrate and film could reduce the density of defects A smaller number of defects should theoretically increase the mean free path of electrons A larger MFP means more electrons can escape which leads to an improved QE This project aims to test this hypothesis by preparing clean and atomically flat Mo (100) surfaces for deposition of CsK 2 Sb thin films Film-substrate boundary 45 deg rotation of film with respect to substrate

6. Project Summary Assembled UHV system for surface diagnostics and Mo sample prep Wrote software to operate LEED/AES system devices and save data Did preliminary system tests Prepared Mo substrates Analyzed effect of preparation process on substrate and effect of substrate on photocathode The vacuum system before the rebuilding began

7. Experimental Setup/Preparation System already contained LEED/AES optics, 300 L/S ion pump, Ar leak valve, and RGA Things added to system: -Heater/bellows and power supply -TSP and TSP controller -Load lock system -Cold cathode gauge on load lock -Gate valve -New view ports -Hanger assembly (thermocouple, cradle, etc…) -Camera and mount -Ar ion gun -Measurement devices (pico-ammeter, thermocouple box, ADC/DAC) Performed 48 hour bake out at 150 C

8. LEED/AES System Software

9. LEED/AES Software Software used to monitor various parameters: hanger temperature; chamber pressure; current through puck; heater current, power, and voltage; partial pressures and residual gas mass spectra This involves measuring analog output signals of thermocouple box, cold cathode gauge, and pico- ammeter with an ADC Also used to operate heater power supply, RGA, and camera RGA software edited to allow for total pressure measurements and abiltiy to turn on/off CEM at will

10. Preliminary Tests

11. Thermocouple Calibration We need to measure the temperature of the puck/substrate during annealing Must calibrate the hanger thermocouple in order to find the temp of the puck Attached a thermocouple to a Mo test puck Increased heater current in small steps and measured temperature of both thermocouples after equilibration at each step Found an average calibration factor of about 2.7 Calibration factor may be off due to puck not fitting in hanger/thermocouple breaking

12. LEED LEED: Low-Energy Electron Diffraction Useful for determining 2-D surface structure since electrons do not penetrate deeply Diffraction condition depends on reciprocal crystal lattice and is very sensitive to condition of first ~4 monolayers of crystal LEED will be used as a binary test to determine whether or not the Mo (100) surface is clean and flat Unit Cell

13. Si Crystal Cleaving and LEED Test Tested LEED using a Si crystal tip surrounded by a BeO screen BeO screen used originally to line up beam, but BeO built up charge which caused diffraction patterns to vary with time Removed BeO screen and cleaved Si crystal on magnetic arm Tested LEED on cleaved crystal and obtained tie-dye patterns This indicated that there may be some residual magnetic/electric field problems This test warranted the need to test LEED on a different type of crystal, GaAs, which had previously produced succesful LEED patterns

14. Experimental Issues Tested LEED on clean, well-defined crystals: freshly cleaved GaAs and Si (with and without BeO screen) Expected typical LEED pattern (several bright intense spots) Obtained tie-dye like pattern which changed with both time and incident electron energy Tie-dye phenomenon may be due to presence of residual magnetic fields (YAG screen test, and hall probe test of LEED optics and puck hanger) May also have charge building up on LEED screen insulators or view ports (a) GaAs LEED pattern(b) GaAs LEED pattern (c) Si LEED pattern(d) Si LEED pattern

15. GaAs Cleaving and LEED Test Before performing GaAs test, the LEED optics and the hanger were degaussed, though strong fields (5-15 gauss) were still present, even after degaussing Cleaved GaAs crystal in same manner as for Si to provide a clean, well-defined crystal surface This method had produced a good LEED pattern on another system Obtained tie-dye pattern once again

16. Ar Ion Sputtering Needed to test the abilities of the ion gun Used an oxidized GaAs wafer which allowed us to clearly see the effects of the sputtering Sputter gun requires system be backfilled with Ar to 5e-05 Torr (can be time consuming) Sputtering occured at near normal incidence (alligned by eye) for approximately 20 minutes in one spot (also time consuming) Sputtering parameters were 20 mA emission current and 2 kV filament voltage

17. Mo Sample Preparation Prepared 2 Mo substrates from 99.9% pure Mo foil Cut Mo foil (~2 mm thick) to disk shape Hand polished disk using 30 μm diamond suspension followed by 15 μm sand paper 4 or 5 deep scratches remained which couldn’t be removed Moved to 400 grade silicon carbide sand paper (with water) followed by 600 grade, then 9 µm, 6 µm, 3 µm, 1 µm, 0.25 µm, and 0.1 µm diamond suspension pastes (in that order) Between each polishing, the surface was cleaned using coconut-oil soap and isopronal Following polishing, the two substrates were taken to AFM to analyze RMS roughness

18. Mo Surface Roughness (Pre-Annealing) (a) Surface of hand polished Mo substrate (Luca). 1 micron scale, RMS Roughness: 2.27 nm (b) Surface of hand polished Mo substrate (Luca). 1 micron scale, RMS Roughness: 2.64 nm (c) Surface of hand polished Mo substrate (Luca). 10 micron scale, RMS Roughness: 5.32 nm (d) Surface of hand polished Mo substrate (Matt). 10 micron scale, RMS Roughness: 18.95 nm (e) Surface of hand polished Mo substrate (Matt). 1 micron scale, RMS Roughness: 3.52 nm

19. Thermal Annealing and Ion Sputtering After AFM imaging, samples underwent a sputtering/annealing process The flatter substrate underwent a 3-stage process and the other underwent a 2-stage process 1 stage of the process consisted of at least 20 minutes of ion sputtering in the same location followed by annealing for at least 30 minutes at temperatures >870 C Reached a peak temperature of 950-1000 C on the flatter puck Found that currents through the heater of >12 A cause enough heat to cause accumulation of melted indium solder which holds puck in place

20. Results

21. SEM/EDX Measurements

22. SEM/EDX Measurements Cont…

23. More SEM The EDX spectra indicate that carbon is in fact present on the surface: diamond from the suspensions may have lodged itself in the Mo surface There are other elements which may have come from dust or polishing (soap, suspensions, etc…) There are also clearly still large pits, holes, and scratches in the surface, even after annealing

24. AFM Measurements After Annealing (a) Surface of hand polished Mo substrate (Luca). 10 micron scale, RMS Roughness: 7.43 nm (b) Surface of hand polished Mo substrate (Luca). 10 micron scale, RMS Roughness: 7.46 nm (c) Surface of hand polished Mo substrate (Luca). 50 micron scale, RMS Roughness: 4.05 um

25. Spectral Response of CsK 2 Sb Photocathode with Mo (100) Substrate

26. Thermal Emittance and Response Time Measurements Emittance measured using beam profile after solenoid and solenoid transfer matrix Used 0.5mm, 1.0mm, and 1.5mm laser aperature sizes and 473 nm, 532 nm, and 405 nm laser wavelengths Response time measured using deflector cavity in ERL injector Measured <1ps for the Cs 3 Sb response time (limited by resolution of equipment) These measurements resemble closely those of CsK 2 Sb (to within error bars)

27. Future Work The improved QE obtained using the Mo substrate warrants a more detailed investigation into the preparation of the Mo substrates for the multi-alkali photocathodes and their overall effect on the cathode’s performance A bi-alkali photocathode on a Mo substrate still needs to be tested in the DC gun (need to fix the melted indium solder problem) A more detailed investigation must be undergone to fix the problems associated with LEED

28. Acknowledgements Thanks to Luca Cultrera, Ivan Bazarov, Tobey Moore, Joseph Conway, Richard Merluzzi, John Dobbins, Karl Smolenski, Siddharth Karkare, Adam Bartnik, and the whole ERL injector team Thanks also to the NSF and DOE http://www.wellesley.edu/Chemistry/Chris/LEED.ht ml http://www.wellesley.edu/Chemistry/Chris/LEED.ht ml http://www.lns.cornell.edu/~ib38/research.html