High-brightness electron beam evolution following the laser-based cleaning of a photocathode

Laser-based techniques have been widely used for cleaning metal photocathodes to increase quantum efficiency (QE). However, the impact of laser cleaning on cathode uniformity and thereby on electron beam quality are less understood. We are evaluating whether this technique can be applied to revive photocathodes used for high-brightness electron sources in advanced x-ray free electron laser (FEL) facilities, such as the Linac Coherent Light Source (LCLS) at the SLAC National Accelerator Laboratory. The laser-based cleaning was applied to two separate areas of the current LCLS photocathode on July 4 and July 26, 2011, respectively. The QE was increased by 8-10 times upon the laser cleaning. Since the cleaning, routine operation has exhibited a slow evolution of the QE improvement and comparatively rapid improvement of transverse emittance, with a factor-of-3 QE enhancement over five months, and a significant emittance improvement over the initial 2-3 weeks following the cleaning. Currently, the QE of the LCLS photocathode is holding constant at about 1.2  10  4 , with a normalized injector emittance of about 0.3 µm for a 150-pC bunch charge. With the proper procedures, the laser cleaning technique appears to be a viable tool to revive the LCLS photocathode. We present observations and analyses for the QE and emittance evolution in time following the laser-based cleaning of the LCLS photocathode, and comparison to the previous studies, the measured thermal emittance versus the QE and comparison to the

the drive laser on the cathode had to be moved frequently to find new high-QE spots. This movement and subsequent retuning of the photo injector occupied significant LCLS machine time, and only a limited number of laser locations on the cathode could deliver the desired low emittance electron beam for reasonably good FEL performance. The second cathode was then replaced by a third one in May 2011, but the initial QE of this third cathode was only ~510 -6 , insufficient for user operations. Eventually, laser-based cleaning was initiated on the third photocathode, in order to boost the QE. Previous cleaning attempts for the third cathode, using in-situ hydrogen plasma cleaning [3], failed to achieve adequate QE improvement. Laser-based cleaning techniques have been used in the photo injector community for many years on metal cathodes, such as copper and Mg, to enhance QE [4][5][6]. A high-intensity laser beam, interacting with the cathode, may ablate the cathode surface and/or remove contamination, thereby resulting in a QE increase. However, the impact of laser cleaning on cathode uniformity and electron beam emittance are unknown at present. We evaluated whether this technique could be used to revive the LCLS photocathode for x-ray FEL facilities, which have stringent requirements on the beam emittance as well as the QE. Laser-based cleaning for the LCLS photocathode was successfully performed in July 2011, and we observed an evolution of the QE and emittance following the cleaning. This paper will first introduce parameters and procedures used for the laser-based cleaning of the LCLS copper photocathode. Then, the observed evolution of the QE and emittance following the laser-based cleaning and corresponding analyses are presented. Finally, we present the measured LCLS thermal emittance for different QE.

II. LASER-BASED CLEANING PARAMETERS AND PROCEDURES
The applied laser fluence is a key parameter in the laser-based cleaning of metal cathodes. The fluence of the refocused UV drive laser (253 nm) needs to be properly chosen so that the laser can effectively remove surface contamination to enhance the QE, but will not destroy the cathode surface quality or change the surface morphology.
For this application to the LCLS copper cathode, the laser fluence used for laser cleaning was determined by the "vacuum activity" in the photocathode RF gun [5]. In other words, the applied laser fluence (laser energy for a given laser spot size) had to be gradually increased until a change in vacuum pressure in the RF gun was observed. In the LCLS gun system, the nearest vacuum gauge to monitor the gun vacuum is located at a nearby RF-feed waveguide, as shown in Figure 1 [7]. The cold cathode ion gauge on the waveguide is about 50 cm away from the cathode.
Estimate shows the vacuum pressure on the cathode is 1.3-1.5 times higher than the ion gauge [8]. A pressure rise of ~0.510 -10 Torr was observed in the LCLS gun waveguide when the pulse energy of the laser illuminating the photocathode was increased to 17 J with a 30 m rms spot size. With this vacuum activity, removal of cathode surface contamination was expected. The laser was then rastered in a 2D grid across the cathode to clean the surface, using a 30 m step size in x and y. Figure 2 shows the typical vacuum activity in the gun waveguide during the laser cleaning. The gun-waveguide vacuum increased to about 7.510 -10 Torr from the base value of 710 -10 Torr. Note that the base vacuum on the gun waveguide was elevated to 710 -10 Torr, from the typical steady-state vacuum of 310 -10 Torr (with the gun RF off), due to a few previous hydrogen cleanings and laser fluence-determination testing prior to the formal laser cleaning. After the first run of the cleaning, the QE increased to ~110 -5 from an original value of 610 -6 . Two more runs followed, with laser pulse energy slightly increased to about 20 J, maintaining a 30-m rms spot size. The cleanings enhanced the QE to ~410 -5 , 7-8 times before the laser cleaning. Table 1 gives the major parameters used for cleaning the LCLS cathode. The RF power for the gun was always turned off during the laser cleaning process. All QE data were measured at a 30 laser launch phase from zero-crossing, with 115 MV/m of peak gun electric field, using a 1 mm diameter drive laser spot on the cathode. During the QE measurements, the laser energy on the cathode was varied to produce a constant 150 pC bunch charge.

III. EVOLUTION OF QE AND EMITTANCE FOLLOWING CATHODE LASER CLEANING
Two separate square areas on the LCLS cathode (2 mm  2 mm each) were processed by the laser-based cleaning technique on July 4 and July 26, 2011, using the focused drive laser beam. The focus size was  x =30 m rms. Figure

A. QE evolution
The QE measured immediately after the cleaning process was about 410 -5 in the areas exposed to the laser cleaning. Figure  Note that the LCLS machine is always operated at a 120 Hz repetition rate for user operations. The figure shows that over time the QE increased by a factor-of-3 and reached about 1.2x10 -4 after about 5 months of operation. The gun waveguide vacuum also improved, from 810 -10 Torr to 6.510 -10 Torr during this period. The QE for the unusedbut-laser-cleaned spots in the area B is also improved following the cleaning on July 26, 2011. About 6 weeks later, the QE was increased to 610 -5 from 510 -5 . The QE was mapped again on February 2, 2012, eight months following the cleaning ( Figure 5). We measured the QE of the area B, (Figure 3b), by moving the drive laser beam from -1.5 mm to +1.5 mm in the x-plane for different y-locations, -2.5 mm, -2.0 mm, and -3.0 mm, respectively.
Because a 1-mm-diameter laser spot size was used to measure the QE, the full laser beam was located within the cleaned area for a laser central x-location ranging from -0.5 mm to +0.5 mm ( Figure 5). For a central x-location beyond +1.5 mm or -1.5 mm, the full laser beam diameter was located completely outside the cleaned area. The data in Figure 5 show that the QE for the spots within the cleaned area had also increased from 510 -5 to ~1.310 -4 . The data shown in both Figures 4 and 5 suggest that the QE improvement in both areas might be related to the gun vacuum. Figure 5 also shows that the QE data for the un-cleaned areas, beyond +1.5 mm or -1.5 mm for laser central x-location, was still within the 10 -6 scale. Figure 6 shows x/y=0/0 area was still at a very low level, although the overall gun vacuum had continuously improved. Table 2 summarizes the QE evolution for the areas A, B and non-laser-cleaning area.

B. Discussion of the QE evolution and comparison to previous studies
Although detailed surface and material science studies for the third LCLS cathode are still pending, we assume that the cathode surface exposed to laser cleaning still retained contaminants, which were pumped out over time, causing a slow increase in the QE. However, the contaminants on the un-cleaned surface remain unchanged, and appear strongly bound to the surface, and are not removed as a result of vacuum improvement. Cathode R&D programs to further understand the detailed surface and material processes that take place during the laser cleaning are under way at the SLAC [9].
The first LCLS cathode was also processed by laser cleaning for a few times in 2007 and 2008 but using different procedures compared to the one for the third LCLS cathode. The QE of the first cathode decayed quickly during the operations following the cleaning. During the laser cleaning process in 2007, the vacuum measured on the gun-waveguide increased to about 110 -8 Torr from the base vacuum of 110 -9 Torr. The first cathode was laser cleaned again in 2008. The gun vacuum rise during the cleaning in 2008 was at least 2-3 times higher than during laser cleaning of the third cathode (~0.510 -10 Torr). The laser fluence for cleaning the first cathode was at least twice for the third cathode. A small vacuum leak in the waveguide for the first accelerator section following the gun system was observed during the early days of the LCLS operations, which caused an additional gas load to the cathode. Similar phenomena for cleaning Mg cathodes were also observed [6]. Upon the laser cleaning, the QE of the Mg cathode improved two orders of magnitudes [6] against about one order for the third LCLS cathode. The laser fluence and/or laser exposed time for the cleaning [6] were much higher than for the current third LCLS photocathode. During operation of the Mg cathode the QE following the laser cleaning did not decay during the first three months of operation but did not further increase as we observed for the third LCLS cathode. The comparison of the previous cleaning results to the third LCLS cathode illustrate that the laser fluence and laser exposed time for the cleaning need be properly chosen to have a good QE evolution during operations following the laser cleaning.

C. Emittance evolution
The LCLS injector emittance measurements are made using a quadrupole scan. After acceleration of the electron beam to 135 MeV, the beam is intercepted by a 1-m thick aluminum screen. Here, the transverse electron beam size is measured using optical transition radiation (OTR) from the screen, which is imaged onto a digital camera.
The strength of an upstream quadrupole is varied over several settings while the horizontal beam size is measured on the OTR screen. Figure 7 shows Summary of the emittance evolution of the spots in the areas A and B following the cleanings and comparison to the simulations is given in Table 3. The emittance improvement, compared with the value measured immediately after cleaning, is attributed to an improved, more-uniform QE emission. Figure 8 shows a comparison of the QE emission measured immediately after cleaning and few months following the cleaning. The continuous RF conditioning and the formation of residual gas layers on the cathodes surface during routine user operations may smooth-out a non-uniform surface created by the laser cleaning. Gaussian-cut laser spatial profile, rather than a pseudo-uniform profile [11].

IV. CONCLUSION
The laser cleaning technique has been successfully applied to the LCLS injector, for cleaning the copper photocathode. The QE was enhanced by 8-10 times upon the laser cleanings. Since the laser cleaning was performed on the LCLS cathode, routine operations have shown a slow improvement of the QE and comparatively rapid improvement of the transverse emittance, with a factor-of-3 QE enhancement over five months and a significant emittance improvement over the initial 2-3 weeks following the cleaning. Currently, the LCLS photocathode QE is holding constant at about 1.210 4 , with a normalized injector emittance of about 0.3 µm for a 150-pC bunch charge. Similar evolutions of both QE and emittance for the two separate areas exposed to the laser cleaning are observed. Discussions on the QE evolution and comparison to previous studies are presented. With the proper procedures, the laser cleaning technique appears to be a viable tool to revive the LCLS photocathodes for x-ray FEL operations. In addition, measurements show that LCLS cathode with different QE (up to a factor of 2) has similar thermal emittances, which suggests that cathode surface contamination impacting QE may not modify the work function, and thereby the thermal emittance.

V. ACKNOWLEDGEMENTS
We would like to thank the dedicated LCLS physicists and operations team for their contributions and strong support. We particularly thank Vitaly Yakimenko (BNL) for providing the laser-cleaning recipe and helpful advice, and Marcelo Ferreira for the estimate of vacuum on the cathode with the vacuum data on the ion gauge. This work is supported by DOE under contract No. DE-AC02-76SF00515.