Performance of laser patterned copper plasmonic photocathodes

We study ultrafast laser surface nanopatterning as an alternative to improve the photo-emissive properties of metallic photocathodes. By tailoring the physical dimensions of these surface nanostructures, one can localize the optical field intensity and exploit plasmonic effects occurring in such nanostructures. As a result, this surface nanopatterning technique can become a great tool for improving metallic photocathodes photoemission behavior enabling their use for next generation high brightness electron sources. Our goal is to investigate such surface-plasmon assisted photoemission processes with a view on simplifying the photocathode production at CERN while extending the lifetime of existing photoinjectors. The performance of two different femtosecond laser nanopatterned plasmonic photocathodes was analyzed by measuring the quantum yield with a 65kV DC electron gun utilizing 266nm laser excitation generated by a nanosecond laser with 5ns pulse duration and 10Hz repetition rate. By comparing the electron emission of the copper surface nanostructured areas with that of a flat area, our results suggest quantum yield enhancements of up to a factor of 5.


Introduction
In the field of particle accelerator physics, the use of plasmonic nanostructures is extremely interesting and constitutes a pathway to more efficient photocathodes by optimizing the photoemission process either via increased light absorption or tailored electric field-enhancement.Plasmonic nanostructures have already been routinely utilized for electron beam production with specific phase space properties.Nanostars [1], nanorods [2], carbon nanotubes [3] and other multi-resonant particles [4] have shown exciting routes for spatiotemporal photocurrent control over femtosecond (fs) laser pulses thanks to frequency and polarization selective excitation of the fabricated plasmonic hot spots/nanotips.Even plasmonic lenses through bulls-eye designs have been demonstrated for spatiotemporal confinement of energetic electron pulses [5].These electron emitters constitute a great way to generate ultrafast electron pulses with tailored properties and nanometer-scale density modulations of the electron beam which can be of great benefit in next-generation coherent X-FELs and ultrafast electron diffraction studies.
Most of these studies utilize IR fs-pulses and are tested for rather low peak current generation but are not a suitable option when a high average current needs to be extracted from the photocathode (need for µA-mA range).Advanced biomedical accelerator applications such as FLASH therapy for cancer treatment [6], energy recovery linacs [7], inverse Compton scattering sources [8] or new generation x-ray sources based on free electron lasers, rely on the use of photoinjectors as the electron beam source and require high peak and average current [9].
Photocathodes used nowadays for such applications are metals (such as Cu, Mg or Nb) or high quantum-efficiency positive-electron-affinity semiconductors like alkali-antimonides [10].Metallic photocathodes tend to have relatively low QE (values in the 10 −4 -10 −5 region when using UV-photons above work function) but are standardly the material of choice for most high peak brightness photoinjector guns due to their prompt response, low vacuum requirements, ruggedness and long lifetime.Therefore, new routes for enhancing metallic photocathodes QE will suppose a great advance for electron beam based applications.
In this work, we propose to use plasmonic nanostructures for enhancing directly the singlephoton absorption photoemission process using UV laser pulses over metal's work function (the standard source of photocathode excitation in most of state-of-the-art electron accelerator facilities).By exploiting the field-enhancement generated from the interaction of the UV photons with resonant nanostructures, we are able to improve the measured quantum yield (QY) of two distinct plasmonic copper photocathodes with a DC-electron gun.

Surface nanopatterning
In order to generate such field-enhancements under UV laser irradiation, spherical nanoparticles (NPs) and/or periodic patterns with a sub-wavelength size are needed.However, it is known that the manufacturing of robust and reproducible structures in a size range smaller than 200 nm can be challenging [11].
To address this challenge, we take advantage of the direct ablation process induced by ultrashort pulsed lasers (same standard lasers used for photocathode irradiation) when the laser fluence is tuned slightly above the material's ablation threshold leading to different types of controllable size nanostructures.
The use of ultrafast lasers (fs and ps) for surface nanostructuring typically results into two main types of nanoscale features.First, the well-known ripple-like Laser Induced Periodical Surface Structures (LIPSS) with controllable sub-wavelength geometries depending on parameters such as irradiation wavelength, fluence, number of pulses or wavefront [12].Second, the condensation of individual NPs and larger particle clusters covering the laser-treated surface arising from the rapid expansion and cooling of the ablation plume generated after the absorption of high-intensity ultra-short laser pulses [13,14].
Interestingly, depending on the material and laser pulse duration, the use of fluences slightly above the material's ablation threshold can lead to a preferential generation of NPs with nanometric size and suppression of larger size clusters deposited at the surface which often arise at higher fluences [15,16].These NPs can present sizes in the range of few tens to hundreds of nanometers, which result to be perfect to tailor the plasmonic response of UV-illuminated photocathodes.
Sample photocathodes were fabricated utilizing two distinct fs-laser systems tuning the laser fluence to be slightly above the photocathodes ablation threshold aiming at generating soft nanopatterning.Figure 1 shows that the fabrication process led to a combination of LIPSS nanopatterns covered by numerous NPs.
The photocathode nanopatterned with a fs-laser with 800 nm wavelength, 150 fs pulse duration, 1 kHz frequency, focused spot size of 5.5 µm and scanned at 70 µm/s led to the nanopatterned areas showed in Fig. 1(a, c) and is identified as Cathode A. The photocathode nanopatterned with a fs-laser with 515 nm wavelength, 260 fs pulse duration 100 kHz frequency, focused spot size of 28 µm and scanned at 10 mm/s led to the nanopatterned areas showed in Fig. 1(b, d) and is identified as Cathode B. The resulting LIPSS nanopatterns presented a spatial periodicity of 550 nm and 380 nm respectively.However, alongside with the formation of the nanoripples, NPs of considerably smaller size were observed in the SEM images.In this regard, the dominant NPs size measured over the nanopatterned areas presented an average diameter of 95 ±27nm and 64 ±17nm for each case, as shown in the insets of Fig. 1(c, d).
The insets of Fig. 1(a, b) show optical pictures of the final produced photocathodes and it is important to remark that in the case of Cathode A (inset Fig. 1a) only the central area of the photocathode was nanopatterned with the purpose of having flat areas to directly compare the effect of the nanostructures.

Surface composition analysis An X-ray photoelectron spectroscopy (XPS) analysis of the surface composition of each photocathode was carried out inside the laser nanopatterned areas and in the non-treated areas (i.e. polished areas). The results presented in Table 1, indicate that the photocathode's surfaces presented residual Cs and Te traces which may have originated from the bake-out process used to achieve ultra-high vacuum conditions in the DC-GUN chamber housing the photocathodes reaching temperatures up to 250 degrees.
The difference in Cs and Te composition between polished and nano-patterned areas however is small, indicating that the residual deposition of Cs and Te was homogeneous along the photocathodes surface and therefore allowing a relative comparison of the effect of the nanopatterns in the photoemission behavior during testing with the DC-GUN.

Results
The photoemission experiments were conducted inside a DC electron gun with an online monitored constant base pressure of less than 10 −11 mbar and a supplied voltage of 65 kV.The drive UV laser with λ= 266nm, 5 ns pulse duration and 10 Hz repetition rate was directed to the photocathodes surface via in-vacuum mirror with an incidence angle of 5 degrees, and focused to a gaussian beam spot size of approximately 2 mm diameter at the photocathodes surface.Finally, the electron beam current was online monitored by a wall-current monitor and a faraday cup.For a complete QY assessment, we scanned the UV laser beam in x-and y-direction across the whole photocathodes surface and measured the generated current for different values of the UV laser pulse energy (online monitored by sampling the beam).The results are presented in Fig. 2.
As expected after the XPS analysis results, the measured QY for the polished copper areas was slightly higher than what has been reported for copper where the QY reported values are most of the times in the range of 1*10 −5 -5*10 −5 for UV laser irradiation over the work function [17,18].Figure 2(a, c) shows that for polished copper areas in Cathode A the measured QY was 1*10 −4 .This result can be explained by the detected Cs and Te traces deposited during the heating of the DC-GUN preparation chamber.
In contrast, the QY measured from the nanostructured area of the same Cathode A was determined to be approximately 5*10 −4 , showing an almost 5 times enhancement clearly appreciated in the QY map presented in Fig. 2a.According to the XPS measurements, it seems unlikely that the QY increase comes from different Cs % measured in each area since the difference is very small.In addition, since the LIPSS nanopatterns periodicity (550 nm) is larger than the drive UV laser wavelength, they are not expected to provide any significant plasmonic field enhancement.Therefore, the main contributors for this enhancement are expected to be the previously characterized spherical nanoparticles with a measured diameter 95 ±27nm.
Cathode B QY analysis is shown in Fig. 2(b, d).In this case, the whole photocathode area was nanopatterned with a checkerboard design that can be observed in the inset of Fig. 1b.The results for Cathode B show that the maximum measured QY (2.5*10 −3 ) is approximately 25 times higher than that of the polished areas of Cathode A. This promising result however needs to be taken with precaution since as shown in the XPS analysis, Cathode B presented a higher Cs content which might explain the higher QY increase.Apart from that, as shown in Fig. 1d, the NPs present in this cathode show a dominant size distribution with 64 ±17nm diameter which can lead to higher local field-enhancements.
In the performed measurements, a local variation of the QY corresponding to the variation of polished or nanostructured regions could not be identified (see Fig. 2).The first possibility to

Figure 1 .
Figure 1.(a, b) SEM images of the nanopatterned areas with dimensions of 2x2mm, filled with LIPSS nanopatterns covered by NPs on Cathode A and B respectively.Inset pictures show the actual plugs prior introduction into the DC-GUN (c, d) Higher magnification SEM images of Cathode A and B respectively, with insets showing the measured NPs size distribution.

Figure 2 .
Figure 2. (a) Quantum yield map obtained for Cathode A using a pulse energy of 13.6µJ(b) Quantum yield map obtained for Cathode B using a pulse energy of 5.7µJ.(c)Electron beam charge as function of the UV laser energy delivered to Cathode A at the nanostructured area and at the polished area, corresponding to coordinates (0, 0) and (0,4) in the 3D map respectively.(d) Electron beam charge as function of the UV laser energy delivered to Cathode B at the point of maximum and minimum QY observed in the 3D map, corresponding to coordinates (0, 4.5) and (-2, 0) respectively.