Introduction

The short and ultrashort laser pulse irradiation of copper surfaces allows generating geometrical, chemical, structural, and morphological surface modifications. Laser ablation processes inducing copper surface modification were reported for pulsed irradiation with different pulse duration and laser wavelength [1, 3, 5, 15, 18, 20,21,22, 26, 28, 29, 31, 33, 41, 42, 44]. Laser ablation was studied for ns- [3, 14, 18, 23, 33, 37], ps- [5, 13, 15, 20, 22, 23, 29, 41, 42, 44] and fs-laser irradiation [15, 20, 26, 28, 29, 41, 42, 44] with wavelengths from ultraviolet to infrared. The ablation was furthermore investigated as a function of several other laser parameters such as the wavelength λ [33], the pulse duration ∆tp [15, 29, 41], the number of laser pulses N [5], the laser pulse energy E [33], the repetition rate f [20, 28], as well as multi-pulse irradiation (burst) [4, 44]. Usually, a complex correlation of the laser processing parameters and the resulting ablation characteristics, such as the ablation volume Vab, the shape of the formed crater, the crater width d and depth ∆z, are observed. In general, the ablated volume increases with higher laser pulse energy [29]. Typically the ablation process is more effective in case of UV irradiation compared to IR laser sources [33]. Fan et al. [11] reported experiments on the irradiation of copper with an IR ps-laser, and found at suitable laser parameters the formation of a nanostructured surface induced by a redeposition process. The size and number of the formed nanoparticles are dependent on the laser processing parameters, especially the scanning speed [11]. The related ablation processes were also studied for different gas environments [3, 31] and for additional external magnetic fields [23]. Also, the pulsed laser deposition (PLD) of copper surfaces at ambient pressure results in the formation of a chain-shaped nanostructure at the copper surface [12]. Besides the characterization of the ablation structures by surface sensitive techniques, the dynamics of the ablation process was also studied by shadowgraphy [23]. The ablation process was described in a thermodynamic model. At ultrashort laser pulses, a two-temperature model can be used to describe the laser-solid interaction [1, 6]. A hybrid simulation, that couples the finite difference method to solve the heat transfer equations with molecular dynamics calculations, allows a good description of the ablation mechanism [27]. The ablation process of metal surfaces results, in addition to ablation pits, for suitable laser parameters in the formation of a nanostructured surface [31] with interesting and adjustable physical quantities, such as low reflectivity [31] and low secondary electron yield (SEY) [2, 7, 8, 14, 32, 34,35,36,37, 39, 40] enabling multiple applications [2, 11, 12, 39]. SEY is the physical quantity which governs multipacting in radiofrequency devices, charging phenomena in satellites and electron cloud in particle accelerators. In particular, in cryogenic section of accelerators, as for the Large Hadron Collider (LHC), it is responsible of a large fraction of the beam-induced heat-load to be dissipated by the cryogenic plant. Effects of laser-treatment on SEY was mainly studied at λ = 532 nm [7, 8, 35]. In this study, the UV ps-laser treatment of copper surfaces was investigated with the aim to produce nanostructured surfaces with reduced secondary electron yield (SEY) to extend the knowledge of laser-induced nanostructure formation process and the laser-induced SEY reduction. The focus of the performed experiments is the analysis of correlations between surface morphology and the SEY.

Experimental Details

Polycrystalline copper samples (oxygen-free electronic grade – OFE, size: 20 × 20 × 1 mm3) were degreased by a commercial detergent and subsequently passivated in a chromic acid solution. These samples were mounted onto an x-y-stage of a laser workstation from 3D-Micromac. The sample surface was irradiated in air by a Nd:YVO4 laser with a pulse duration of Δtp = 12 ps, a repetition rate of frep = 100 kHz, and a wavelength of λ = 355 nm. A galvo scanner is advancing the laser spot across the sample surface at a velocity v. This scanning velocity was varied from 1 to 200 mm/s. The laser beam was focused onto the sample surface by a f-theta lens having a focal length of 103 mm. The Gaussian laser beam spot radius ω (14.6 ± 1.5) μm was determined by Liu’s method [19]. The irradiation scheme is based on the creation of parallel lines with a spacing Δy (see Fig. 1). First, single non-overlapping lines (Δy > > 2ω) were produced to characterize the ablation behaviour. In a second step, large areas were laser-textured with a partial overlap of the scan lines with Δy = 10 μm. The number of accumulated laser pulses on the same surface point for non-overlapping line experiments can be estimated by:

  • $${N}_l=\frac{frep\cdot 2\omega }{\mathbf{v}}$$
    (1)
Fig. 1
figure 1

Schematic illustration of the experimental setup for UV ps-laser treatment

Full size image

and for overlapping lines: by

  • $${N}_a=\frac{frep\cdot {\pi \omega}^2}{\mathbf{v}\cdot \varDelta \mathrm{y}}$$
    (2)

The surface morphology of the laser-treated copper surfaces was analysed using optical microscopy (OM), scanning electron microscopy (SEM) and white light interferometry (WLI). For trench profile analysis of the non-overlapping lines, the irradiated line patterns were mechanically wipe-cleaned using propanol-immersed cotton and subsequently underwent ultrasonication in propanol to remove any re-deposited material and particles inside the trenches. For cross section preparation, large-area machined samples were embedded in a transparent resin, ground with SiC papers down to grit P1200 and subsequently polished in sequence with 3 μm and 1 μm diamond suspensions, respectively. These cross sections were characterized using a digital microscope (KEYENCE VHX 6000). The secondary electron yield (SEY) was measured for primary electron energies between 50 and 1800 eV on 20 × 20 mm2 homogenously laser-treated samples. The details of the utilized setup and the implemented experimental conditions for SEY analyses are described in Ref. [24].

Results and Discussion

Characterization of the laser ablation

The single line laser scribing across the copper surface (see Fig. 1) leads to the formation of trenches with a truncated “Gaussian-shaped” cross section. If the line distance Δy is much larger than the Gaussian radius ω, interaction phenomena within the laser process due to redeposited material from previous trench ablation of the nearby line do not occur. The laser-induced trenches were analysed by white light interferometry (WLI). First, the trench depth Δz and trench width d (defined in Fig. 1) were determined in dependence of the laser power P and the scanning speed v.

Both, trench depth and width are increasing for increasing laser power P and decreasing for increasing velocity v, as visualized in Fig. 2(a-d). The increase of the trench depth Δz with increasing laser power P confirms earlier experimental studies [6, 33]. Moreover, the trench depth Δz increases with decreasing scanning velocity, in agreement with other reports on laser ablation of copper [9]. The increase of the trench width at decreasing scanning speed and increasing number of laser pulses N is induced by incubation effects [16], i.e. a reduction of the ablation threshold Pth for increasing number of laser pulses N on the same surface location. For extremely low scanning speeds, thermal accumulation [25] may also be relevant.

Fig. 2
figure 2

Depth Δz ((a), (b)) and width d ((c), (d)) of single trenches, as determined by WLI measurements, in dependence of the laser power P and the scanning velocity v (black dots). The red lines are fits to the data using the model described by Eqs. 5 and 7. (e) Laser power dependence of the ablation volume Vab per laser pulse as determined from WLI profiling within cleaned trenches at a scanning speed v = 200 mm/s (red line: estimated ablation volume Vab)

Full size image

The laser power distribution for a Gaussian laser beam profile is given by:

$$\mathbf{P}\left(\mathbf{r}\right)={\mathbf{P}}_{\mathbf{0}}\bullet {\boldsymbol{e}}^{-{~}^{\mathbf{2}{\boldsymbol{r}}^{\mathbf{2}}}\!\left/ \!\!{~}_{{\boldsymbol{\omega}}^{\mathbf{2}}}\right.}$$
(3)

with the Gaussian radius ω, the spatial distance r to the centre of the laser spot, and the maximum laser power P0 in the laser spot centre. The accumulated number of laser pulses N is dependent on the scanning velocity v, the repetition rate frep, and the Gaussian radius ω (Eqs. 1 and 2). This incubation effect can be empirically described by [10]:

$${\boldsymbol{P}}_{\boldsymbol{th}}\left(\boldsymbol{N}\right)={\boldsymbol{P}}_{\boldsymbol{th}}\left(\mathbf{1}\right)\cdot {\boldsymbol{N}}^{\boldsymbol{S}-\mathbf{1}}\to {\boldsymbol{P}}_{\boldsymbol{th}}\left(\mathbf{v}\right)={\boldsymbol{P}}_{\boldsymbol{th}}\left(\mathbf{1}\right)\cdot {\left(\frac{frep\cdot 2\omega }{\mathbf{v}}\right)}^{S-1}$$
(4)

With S being the incubation coefficient. After substitution of Eq. 4 into Eq. 3, the trench width d can be calculated by:

$$\boldsymbol{d}\left(\mathbf{v},\mathbf{P}\right)=\boldsymbol{\omega} \sqrt{\mathbf{2}\cdot \boldsymbol{\ln}\left(\frac{\boldsymbol{P}}{{\boldsymbol{P}}_{\boldsymbol{th}}\left(\mathbf{1}\right)\cdot {\left(\frac{frep\cdot 2\omega }{v}\right)}^{S-1}}\right)}$$
(5)

Based on Eq. 5, the experimentally determined dependencies of the trench width on the velocity v and the laser power P were fitted (red lines in Fig. 2(c),(d)). The obtained fit parameters are summarized in Table 1. Assuming that the ablation depth Δz is given by the effective penetration depth (\({\alpha}_{eff}^{-1}\)) of the laser light according to the Beer-Lambert law, and considering that the ablation threshold (Pth) must be exceeded, the ablation depth can be estimated by:

$$\mathbf{P}\left(\mathbf{z}\right)={\mathbf{P}}_{\mathbf{0}}\bullet {{\boldsymbol{e}}^{-}}^{{\boldsymbol{\alpha}}_{\boldsymbol{e}\boldsymbol{ff}}\bullet \boldsymbol{z}}\to \Delta \mathbf{z}\left(\mathbf{P}\right)=\frac{\mathbf{1}}{{\boldsymbol{\alpha}}_{\boldsymbol{e}\boldsymbol{ff}}}\boldsymbol{\ln}\left(\frac{\boldsymbol{P}}{{\boldsymbol{P}}_{\boldsymbol{th}}}\right)$$
(6)
Table 1 Summary of the parameters determined by iterative fitting Eqs. 5 and 7 to the experimental results in Fig. 2
Full size table

The substitution of Eq. 4 into Eq. 6 allows the description of the trench depth Δz in dependence of the processing parameters (scanning speed v and laser power P):

$$\Delta \mathbf{z}\left(\mathbf{v},\mathbf{P}\right)=\frac{\mathbf{1}}{{\boldsymbol{\alpha}}_{\boldsymbol{eff}}}\boldsymbol{\ln}\left(\frac{\boldsymbol{P}}{{\boldsymbol{P}}_{\boldsymbol{th}}\left(\mathbf{1}\right)\bullet {\left(\frac{{\boldsymbol{f}}_{\boldsymbol{rep}}\bullet \mathbf{2}\boldsymbol{\omega }}{\mathbf{v}}\right)}^{\boldsymbol{S}-\mathbf{1}}}\right)\bullet \kern0.5em \left({\boldsymbol{f}}_{\boldsymbol{rep}}\bullet \mathbf{2}\boldsymbol{\omega } \right)/\mathbf{v}$$
(7)

Based on Eq. 7, the dependencies of the trench depth on the velocity v and the laser power P were fitted (red lines in Fig. 2 (a), (b), fit parameters listed in Table 1). In agreement with the increase of the trench depth and width at increasing laser power, the ablation volume increases as well. Figure 2(e) presents the determined ablation volume Vab in dependence of the laser power P at v = 200 mm/s. In addition, the ablation volume per laser pulse can be described by Vab = d * Δz * ω / N with d, Δz, and N being calculated from Eqs. 5 to 7, and Eq. 1, respectively, and using the obtained parameters from the fit procedure that are summarized in Table 1. Furthermore, the dependence of the ablation volume on the laser power allows the estimation of the ablation threshold, which is ~0.088 mW at v = 200 mm/s (corresponds to N = 14.6). The value agrees with the ablation threshold calculated by Eq. 4 of Pth (v = 200 mm/s) = 0.082.

This semi-empirical approach based on Eqs. 5 and 7 allows a good description of the experimental results. The determined absorption coefficient αeff (Table 1) is of the same order of magnitude as the linear absorption coefficient of copper, which is 6.8·107 m−1 [17]. The estimated ablation threshold of 0.13 W corresponds to a laser fluence threshold of ~0.4 J/cm2. The value is of the same order of magnitude as the published value for 532 nm, which is ~0.24 J/cm2 [2].

The incubation factor S of 0.83 is almost similar to the reported value of S = 0.7 using a IR ps-laser [43].

Two-dimensional laser irradiation

The laser irradiation of the copper surface utilizing overlapping lines of small distance (Δy is about the size of the Gaussian radius) allows a large area modification of the surface. In Fig. 3(a), a photograph of laser-treated sample is shown. For each sample, 64 regions were modified using different laser treatment parameters. The laser power was varied from 1.23 to 0.01 W (from top to bottom) and the scanning velocity from 1 to 200 mm/s (from left to right).

Fig. 3
figure 3

(a) Photographs of laser-treated Cu samples irradiated with different scanning velocities v [mm/s] = 1, 2, 5, 10, 20, 50, 100, 200 (from left to right) and laser power at 355 nm P [W] = 1.23, 1.19, 1.06, 0.84, 0.53, 0.24, 0.06, 0.01 (from top to bottom). (b), (c) Selected scanning electron micrographs at different magnification of a surface created at P = 1.23 W and v = 1 mm/s, Δy = 10 μm. (d), (e) Optical Micrographs of the near-surface cross section of the same sample

Full size image

The photograph (Fig. 3(a)) reveals the modification of the optical appearance for different treatment parameters that spans from weak colouring, changed reflectivity to completely dark surfaces. In particular, dark surfaces (low reflectivity) can be created at high laser power P and low scanning velocity v. Figure 3(b-e) includes SEM images and optical cross section images of the region marked in Fig. 3(b), that was irradiated with a scanning speed of v = 1 mm/s, a laser power of P = 1.23 W and a line distance of Δy = 10 μm. The nanostructured morphology of the surface is visible in the SEM image at higher magnification, which is due to redeposition of nanoparticles on top of the micrometer-sized textured copper sample, as seen in the cross sections. The surface appears homogeneous, and the top-view SEM measurements do not clearly show the underlying trench-topography. In Figs. 4 and 5, the surface topography as well as the cross section of selected samples dependent on the laser parameter (laser power P and scanning speed v) are summarized for the complete explored parameter range.

Fig. 4
figure 4

SEM images of laser-treated copper surfaces (λ = 355 nm, f = 100 kHz, Δy = 10 μm) at different laser power P (grey: calculated laser fluence Φ) and scanning speed v (grey: calculated number of laser pulses N based on Eq. 2). Based on the different shapes of the formed surface structures, the complete parameter field can be separated in four different zones with different surface morphology. Exemplary SEM images of the four processing parameter ranges are marked with (a–d) and are shown magnified in Fig. 5

Full size image
Fig. 5
figure 5

SEM top view and optical cross-sectional images of the four classified surface modification types. The colour of the frames is related to the modification type (see Fig. 5). (a) P = 400 mW, v = 1 mm/s (red), (b) 760 mW, 20 mm/s (yellow), (c) 140 mW, 1 mm/s (green), (d) 760 mW, 200 mm/s (blue). The white scale bar corresponds to length of 500 nm which is constant for all 4 SEM images, while the black scale bars correspond to a length of 100 μm (the optical micrographs were measured at different magnification)

Full size image

Based on the shape of the nanostructures at the copper surface, the laser parameter field can be subdivided into four groups showing a particular type of morphology (see Fig. 4 surrounded by colour lines):

  • Compact & dense, flower-head-like nanostructures (P ≥ 400 mW, v ≤ 10 mm/s, marked red)

  • Non-compact filament-shaped nanostructures (P ≤ 140 mW, v ≤ 2 mm/s), marked green)

  • Molten and resolidified spherical structures (v ≥ 10 mm/s, size <1 μm, marked yellow)

  • Minor modified surface (P ≤ 760 mW, v ≥ 5 mm/s, size <150 nm, marked blue)

Selected cross sections were prepared in the different modification zones. Optical images of the cross sections for the four different laser parameter groups are shown in Fig. 5. The cross sections below the SEM images reveal the microscopic substructure of the samples as created by the laser–material interaction.

Within the red-marked laser parameter range, the irradiation results in the formation of compact and dense, flower head-like surface nanostructures as seen in Fig. 5(a). A similar structure is shown in Fig. 3(d) and (e) for the sample processed using a laser power of 1230 mW and a scanning speed of 1 mm/s, where the respective cross section is shown for two different magnifications. These cross-section images indicate a modification depth m of approximately 211 μm that is covered by an additional redeposited nanostructured layer of ~28 μm thickness. The reduction of the laser power to 400 mW (v = 1 mm/s, see Figs. 4 and 5(a), red-marked parameter range) results in a reduction of the texture depth to ~108 μm. In this case, the redeposited layer had a thickness of ~31 μm.

For the red-marked laser parameter range, the laser power is significantly higher than the ablation threshold and a distinctly higher ablation volume is expected, i.e. a higher number of particles are available for the redeposition process. The resulting texture depth and the ablated volume both increase for a higher laser power or when reducing the scanning speed.

Beside the ablation volume, the amount of the redeposited material on laser processed surface is additionally influenced by secondary effects. The laser ablation process results in the formation of an expanding laser plume with a high local pressure inside, which results in an expansion of the plume, and consequently defines the formation, scattering as well as deposition of the particles [30].

Furthermore, the small scanning velocity causes an additional laser irradiation of the redeposited material and results in a modification of it. This additional laser tempering process results most likely in the formation of the existing compact nanostructures at the surface. Similar compact flower head-like copper nanostructures were also reported by Calatroni et al. [6] using a ps-laser with a wavelength of λ = 532 nm and a pulse duration of Δtp = 10 ps. In addition, also the utilisation of UV and IR ns laser radiation results in the formation of compact nanostructures [14, 36, 37].

It seems that the resulting nanostructures appear independent of the used laser wavelength and pulse duration. This confirms the conclusion that the nanostructure formation on top of the ablated grooves is explained by a redeposition process. However, the proportion and the morphology of the redeposited material depends on the characteristics of the ablation process, which is defined by the laser processing parameters.

The reduction of the laser power (green-marked parameter range) results in the formation of non-compact filament-shaped nanostructures. At a laser power of 140 mW and v = 1 mm/s (see Fig. 6(c)) the texture depth m reduces to ~20 μm. The distance of the periodic textures corresponds to the scanning scheme and the line distance of 10 μm. For the green-marked parameter range, the laser power is close to the ablation thresholds. An ablation threshold of ~30 mW at 1 mm/s can be estimated (based on Table 1 and Eq. 4). The redeposition process results here in the formation of non-compact nanostructures. Similar structures were found at pulsed laser deposition (PLD) of copper using picosecond laser irradiation in air. The PLD induced copper particles presented a high tendency to agglomerate and form chain-like structures [13].

Fig. 6
figure 6

(Left) Optical micrograph of the laser-treated copper surface (positions 1–4 mark the parameters for which the surface topography and cross sections are shown in Fig. 6). (Right) SEM images of the same laser treated copper sample with λ = 355 nm, f = 100 kHz, Δy = 10 μm at 5 keV primary electron energy

Full size image

The utilisation of higher scanning speed (yellow laser parameter range) results in the formation of molten and resolidified spherical structures (see Figs. 4 and 5(b)). At a scanning speed of v = 20 mm/s and a laser power of P = 760 mW (see Fig. 5 (b)), the resolidified structures exhibit an average texture depth of m ~ 16 μm. For low scanning speed v and low laser power P (blue laser parameter range) only minor surface modifications occur, consistent with a low texture depth. For example, at a laser power of P = 760 mW and a scanning speed of v = 200 mm/s, a texture depth of m = 8 μm was found. For the yellow and blue laser parameter ranges, the ablation volume is small and melting processes mainly dominate the surface modification process. The surface tension and the laser-induced temperature gradient result in a mass flow of the liquid before cool-down and results in solidified micro- and nanostructures on the copper surface.

The laser texturing using different ablation parameters clearly causes macroscopic topography effects (grooves and textures) and microscopic near-surface modifications, as well as redeposited layers. Such complex multiscale surface modifications induce changes of the material properties, as for instance the optical properties (see Fig. 6 (left)). For high laser power and low scanning speed (red zone), the reflectivity in the visible wavelength range is significantly lower than the reflectivity of the pristine surface. The low optical reflectivity (see Fig. 6 (left)) correlates with the formation of surface nanostructures (see Fig. 4).

The SEM images shown in Fig. 6 (right) are obtained using the same SEM conditions, i.e. the emissivity of secondary electrons for the primary excitation energy Ep of 5 keV is represented by the grey level of the image. The figure indicates that the high-Ep (5 keV) SEY is dependent on the laser-treatment parameters. The copper surface irradiated at high laser power and low scanning speed (red zone) has a lower SEY at 5 keV than the untreated surface. In contrast, at low laser power and high scanning speed, the SEY of the laser-treated surface exhibits a higher value than the untreated surface at Ep = 5 keV. In addition, the SEY was determined in dependence on the primary electron energy Ep in the range between 50 and 1800 eV for selected large-area (18 × 18 mm2) samples as shown in Fig. 7(a). The colours of the SEY graphs correspond to the identified four different types of surface topographies. The shape of the SEY curves of the laser-treated samples is very different from the air exposed untreated Cu surface [24], which exhibits a SEY maximum δmax of ~2.18 Ep = 280 eV. The most significant difference after laser-treatment is the much higher primary electron energy of the SEY maximum δmax; in most cases δmax was found at the highest accessible energy of 1800 eV in the experiment. Figure 7(a) indicates that the SEY maximum is reduced for increasing laser intensities. On the other hand, for low laser intensities (e.g. high scan speeds) the SEY at 1800 eV can exceed that of pristine air exposed Cu, a fact which is in agreement with the observation in secondary electron emission at 5 keV as shown in Fig. 6 (right). In particular, the compact flower head-like nanostructures formed at high laser fluence (Fig. 5(a) (red)) exhibit a SEY maximum well below 1 in the full accessible Ep range. At P ~ 1 W and v = 1 mm/s, a SEY maximum of 0.8 was obtained. The SEY maximum can be reduced to δmax = 0.7 at P ~ 2.46 W and v = 2 mm/s for the implemented UV laser. It has to be mentioned that the utilisation of UV and IR ns laser radiation allows the fabrication of similar compact nanostructures with a δmax below unity [14, 36].

Fig. 7
figure 7

a Energy dependence of the secondary electron yield (SEY) for laser-treated copper surfaces (λ = 355 nm, f = 100 kHz, Δy = 10 μm) at different scanning speed v and laser power P. b and c Secondary electron yield maximum δmax dependent on the accumulated laser fluence and the estimated ablation depth Δz (based on Eq. 7), respectively. The red lines are the results of data fitting using Eqs. 8 and 9, respectively

Full size image

The non-compact nanostructures (Fig. 5 (b) (green)) exhibit a SEY maximum around unity (δmax = 0.96 at P ~ 0.13 W and v = 1 mm/s). The re-solidified droplet structures (6(c) (yellow)) exhibit a δmax above 1 (δmax = 1.06 (1.35) for P ~ 1 (0.13) W and v = 20 mm/s.

Similar to the optical appearance, the secondary electron yield maximum δmax is also directly dependent on the accumulated laser fluence Φacc (Φacc = P·N/(f·π·ω2)): as shown in Fig. 7(b) it decreased for increasing Φacc. The dependence can be described by means of the following empirical equation:

$${\delta}_{max}\ \left({\Phi}_{acc}\right)={a}_{\Phi}-{b}_{\Phi}\bullet \mathit{\ln}\left(\frac{\Phi_{acc}-{c}_{\Phi}}{1\ J/ cm2}\right)$$
(8)

which resulted in aΦ = 1.66, bΦ = 0.101, cΦ = 40.62 J/cm2. The experimental results suggest that there is a correlation between δmax and the ablation volume and the ablation depth, respectively. In Fig. 7(b), the SEY dependence on the ablation depth Δz (calculated by Eq. 7 under assumption of the number of laser pulses Na for overlapping lines Eq. 2) is shown. In agreement with the accumulated laser fluence dependency, the secondary electron yield maximum decreased at increasing ablation depth and the behaviour can be described by:

$${\delta}_{max}\ \left(\Delta z\right)={a}_z-{b}_z\bullet \mathit{\ln}\left(\frac{\Delta z-{c}_z}{1\ \upmu m}\right),$$
(9)

where the experimental data and the corresponding fit are plotted in Fig. 78 (b) with the obtained fit parameters az = 1.27, bz = 0.09, cz = 2.01 μm. It can be assumed that the determined ablation depth Δz is a measure for the experimentally determined modification depth m. Thus, the results suggest that the increase of the modification depth leads to a reduction of the maximum secondary electron yield. In this experimental setup, a lowest limit for the maximum secondary electron yield of about 0.7 was determined.

Similar effects were found for the SEY of laser-induced microscopic trenches. Bajek et al. [2] reported the SEY maximum δmax of nanostructure-covered trenches and cleaned trenches. The δmax of nanostructure-covered trenches was smaller than that of the cleaned trenches. The results suggest that the nanostructures are more important for SEY reduction than the microscopic trenches. The ablation process creates a nanostructured surface with deep trenches within the bulk material. Consequently, the laser parameters and especially the accumulated laser fluence have impact on the resulting sample topography and surface roughness as also reported by Gaidys et al. [13]. Both features, the micro-trenches and nanostructures, enhance inelastic scattering of the incoming electrons and in that sense enable an efficient way of energy dissipation that reduces the probability of secondary electron emission.

The experimental finding that δmax decreases with increasing accumulated laser fluence was also observed when irradiating copper surfaces with other laser sources. Valizadeh et al. reported that the reduction of the scanning speed or line spacing, corresponding to an increase of the accumulated laser fluence, leads to a reduction of δmax [37, 38].

Conclusions

Polycrystalline copper surfaces were textured by ps-laser irradiation at λ = 355 nm, f = 100 kHz, and Δtp = 12 ps. The copper ablation process was studied in dependence of the laser power and scanning velocity and an ablation threshold of 0.13 W and 0.4 J/cm2 was determined, respectively. An incubation coefficient of 0.17 was estimated. Test structures consisting of line arrays with a distance of 10 μm were examined to evaluate correlations between surface morphology and the secondary electron yield (SEY). The achieved surface morphologies were classified into four different categories, which form dependent on the laser processing parameters: (1) compact flower head-like nanostructures, (2) non-compact nanostructures, (3) melting structures and (4) minor surface modification. The nanostructure formation is the result of combined processes including laser ablation from the copper surface (microstructure formation), redeposition of ablated products (porous nanostructure formation) and modification/tempering of the redeposited material (modification of the redeposited nanostructures) due to locally enhanced pressure conditions (ablation pressure). The experimental results suggest that the combination of a high ablation volume (high available particle density for the redeposition process), the tempering process and a locally elevated pressure enable the formation of the compact flower head-like nanostructures (at high laser power and low scanning speed) resulting in surfaces with very low SEY maximum as low as δ = 0.7.

In general, the laser-ablation induced morphological and topographical changes have significant impact on the SEY in dependence of the primary electron energy as well as on the SEY maximum. The ablation process creates a nanostructured surface with deep trenches within the bulk material. Both features enhance inelastic scattering of the incoming electrons and in that sense enable an efficient way of energy dissipation that reduces the probability of secondary electron generation and emission. The reduction of the SEY to below unity is required to efficiently mitigate electron cloud formation in particle accelerators.