2. Materials and Methods
A schematic drawing of the laser setup and raster scan strategy is depicted in Figure 1
a. For the sake of clarity, the material reported in the Results section is aluminum. In order to verify whether the proposed ultrafast LASS is generic, a few other materials were also chosen for the same test, such as copper and titanium alloy Ti6Al4V, and comparable results were obtained from these materials, too. Nonetheless, a similar test on stainless steel 316L did not yield satisfactory results, suggesting a certain material dependency and limitation of this process. The targets were prepared by conventional metallography procedures with a surface roughness Ra = 20 nm. The micromachining was carried out using a fiber fs laser system (Tangerine HP, Amplitude Systems, Pessac, France). The laser has a central wavelength of 1030 nm with a pulse duration of 300 fs, and a tunable repetition rate from single shot to two MHz. The linearly polarized laser pulses were attenuated, sent through a Galvano scanner, and focused through a 100-mm telecentric f-theta lens. The focused laser spot exhibits a Gaussian profile and the spot diameter (at 1/e2
) measures 2ω0
= 22 µm. A raster scanning strategy was used. The overlap ratio of 0.91 between successive laser pulses and successive scan tracks was kept constant for all of the experiments. This relatively high overlap ratio was intended for a homogeneous energy deposition [13
]. The laser fluence quoted in this paper is the peak fluence F = 2ε
, with ε
being the laser pulse energy.
The topographical analysis was performed using an optical microscope (OM, Zeiss, Oberkochen, Germany) and scanning electron microscope (SEM, FEI Europe B.V., Eindhoven, Netherlands). Cross-section samples for OM were prepared by wheel-saw halving, cold-setting resin mounting, and mechanical polishing. Cross-section samples for SEM inspection were made using a dual beam system SEM and Focused Ion Beam (FIB, FEI Helios Nanolab 600i dualbeam workstation). Energy-dispersive X-ray spectroscopy (EDX, Brucker, Bremen, Germany) was used to evaluate the surface chemistry change. A Vikers microindentation (Cetim, Saint-Etienne, France) was also performed to access hardness, with a loading force of two N. All of the hardness values quoted in the text were averaged values from five individual measurements. For the wetting test, a three-µL water droplet was deposited on the surface by a microsyringe, and the side-on view was captured by a dedicated camera. The contact angle was deduced from the registered images using software Digidrop (version 13.06.3.12GB, GBX, Romans sur Isère, France).
The lower-right part of the schematic drawing shown in Figure 1
a can be used to illustrate the concept of the LASS by intense ultrafast laser irradiation (roughly defined as fluence >5~10 times the ablation threshold). Then, the material experiences a photo excitation, electron–phonon non-equilibration-enhanced electron heat conduction before equilibrium with the lattice, and then undergoes spallation and phase explosion [14
]. Various species, depending on the laser fluence, consisting of anything from liquid layers and/or large droplets to a mixture of vapor-phase atom clusters and droplets etc., are ejected as a consequence of stress release. Under certain circumstances (a more dedicated discussion is in the following paragraphs), the ejected species land on the adjacent site, adhere to the vicinity of ablation site, and form an accumulation. Such a material redistribution is somehow similar to the work of Temmler, et al. [15
] (although the mechanism governing the material displacement is totally different). By raster scanning a small area on a substrate, the accumulation continues on and eventually forms a three-dimensional structure by itself, which will be termed as additive structures (AS), as indicated by the white bump in Figure 1
a. An AS on aluminum sample was fabricated in this way, and its SEM micrograph is given in Figure 1
b. The material to form the AS was sourced from the ablation groove, as marked in Figure 1
b, at a laser fluence of 18 J/cm2
, and 500-kHz repetition rate. The lateral dimension of the AS is about half of the groove width, and the longitudinal dimension is approximately 50 µm, which is better seen from the mechanical cross-section presented in the inset.
In order to characterize the AS, we performed microstructural and chemical analysis. The SEM micrograph in Figure 1
c shows a FIB prepared cross-section from the substrate–AS junction region. Laminated structures are observed on the AS, the formation of which is due to ablative and accumulative processes. The cross-section elaborates also that the cohesion between the AS and substrate is at the microstructural level. A borderline is evident at the junction, but no major voids nor pores are located there, which implies a robust conjugation between the substrate and the AS. The AS is relatively dense. There are some sub-µm sized voids in the bulk region of the AS, but they are sparse. Nonetheless, it is evident that some micropores are located at the outermost surface region and the lower outer surface region. Those micropores located in the outermost surface region could be potentially eliminated by a substractive ablation process step, which is discussed later in this communication. An EDX spectrum was acquired from different locations across the entire AS, and compared with the spectrum obtained from the substrate. In the AS, a small increase of oxygen and nitrogen is deduced from their characteristic peaks’ intensity in the spectra. This chemical composition change may be due to an oxidation/nitration or air trapping (due to the presence of the sub-µm voids) within the AS during the process.
As mentioned in Section 2
, the AS can be achieved on aluminum but not stainless steel; a typical comparison can be viewed in Figure 2
a,b, with the cross-sections of these two materials treated at a fluence of 18 J/cm2
and a 250-kHz repetition rate. The laser processing is shown to be strongly material-dependent in the ~100 kHz repetition rate regime [16
], resulting either in material accumulation on the side of the ablation groove in the aluminum case, or to significant roughness development in the case of stainless steel. The laser plasma hydrodynamics and ablation particles shielding may be taken account of in these observations.
The interaction of ultrafast laser at fluence exceeding the ablation threshold with metal targets is associated with a large variety of physical phenomena, taking place from femtosecond up to microsecond timescales [14
]. At 250 kHz (a pulse-to-pulse interval of four microseconds), the pulse-by-pulse thermal accumulation alone does not play a major role, as the irradiated surfaces have enough time to cool down during the interval. Besides thermal accumulation, the laser ablation by high-intensity pulses is accompanied by plasma plume expansion away from the metal target [18
]. Figure 2
c depicts the initiation of ablation simulated by two-dimensional (2D) compressible Navier–Stokes equations [2
], taking account of Gaussian pulse energy deposition on the surface, lattice heating, and hydrodynamic movement, which causes strong pressure gradients and material removal. The lowest density corresponds to hot laser-induced plasma, the highest density corresponds to unaffected ablation area, and the intermediate density corresponds to a liquid state. The central part of the plume leaves the surface and expands longitudinally above the surface; however, the corners of the plasma plume remain less hot, may re-solidify before the arrival of the next pulse, and then form a redistributed aluminum layer above the initial level. In this case, the consequent pulse-by-pulse ablation with a scan from left to right leads to a pronounced microstructure on the left side of the laser-processed area, as displayed in Figure 2
a. Other possible mechanisms of material re-deposition include Marangoni melt flow and recoil vapor pressure [15
]. However, these effects fail to explain why the accumulated material is observed above the ablated area, but not outside it.
Another characteristic feature of the ultrashort laser ablation of metal targets is the formation of nanoparticles of sizes up to r = 100 nm in plasma plume [21
]. The species are known to be moving much slower than the plasma plume, with the velocities slightly exceeding 100 m/s [22
]. The presence of slow moving nanoparticles generated by a laser pulse may influence the following laser pulse interaction with the surface. To investigate the effects of light absorption on the surface, with and without transmission through nanoparticles, we use an ablation model, including Maxwell equations [26
] coupled with electron-ion heat transfer equations [19
]. The spherical nanoparticles with a radius r = 50 nm and a concentration of C = 50 µm−2
are distributed randomly in an ellipsoid-centered region at distance of 50 µm above the initial surface level, with a radius Rz = 10 µm in the propagation direction, and Rx = 20 µm in the transverse direction. The nanoparticles are assumed to be created by the first laser pulse. The numerical results show the surface profile after N = one and N = five pulse irradiation with and without taking account of the energy loss in transmission through nanoparticles in Figure 2
d. Although only a slight difference is seen after the first pulse irradiation, the inhomogeneous distribution of the energy due to the presence of nanoparticles seriously degrades the ablation quality after five consecutive pulses: the ablation profile is now asymmetric, with roughness features up to a few hundred nm. It is worth mentioning that owing to the low reflectivity of stainless steel (R = 0.56) compared to that of aluminum (R = 0.96), the greater transmission of laser pulse through steel nanoparticles would cause a stronger surface roughness. The effects of the interaction between the nanoparticles and laser-induced plasma created by the next pulse may as well lead to uncontrolled material deposition inside the crater. Both phenomena may have an impact on the quality and/or surface morphology [22
]. The difference in ablated volume and depth is revealed by simulations, taking into account the presence of nanoparticles in Figure 2
d already after five pulses of irradiation. Thermal effects might be at the origin of further ablation crater degradation, as the diffusivity of stainless steel
is the lattice thermal conductivity,
is the lattice heat capacity, and ρ is the lattice density), which is six times lower than the diffusivity of aluminum
. For aluminum, the material accumulates only at the corners of the laser-processed area, leaving the central part inside the ablation crater free from micro-debris [16
], and facilitating the proposed technique of additive surface structuring.
The application of such AS for surface functionalization is multifold. One of the simplest is implemented by building two basic AS blocks against one to another, and is reported in Figure 3
a. The inset is a mechanical cut cross-section revealing the negatively inclining feature of the sidewalls of the AS. Such geometry is otherwise termed the re-entrant shape, and is capable of rendering a surface liquid repellent property [27
]. In order to test this functionality, a large surface of a few mm2
area was fabricated with repeating “re-entrant” pattern periodically (Figure 3
b). The surface was hydrophilic and anisotropic right after the LASS process, but turned to isotropic and hydrophobic after a thermal treatment at 150 °C for two hours: a droplet could sit on the AS surface without spreading, as shown in the inset. Such surface morphology, once capped with a low surface energy coating, might also exhibit oleophobic characteristics [28
]; experiments on coating preparation and oleophobicity test are currently underway. It is also worth noting that upscaling such ultrafast LASS to an even larger surface, in the order of cm2
, is fairly straightforward nowadays, as it is based on standard ultrafast laser beam shaping, scanning, and processing tools [3
Furthermore, an additional merit offered by the ultrafast LASS process is worth illustrating. Given that the ultrafast laser is also a high-precision micromachining tool, additional laser processing steps can be added after the AS is fabricated, and the surface finishing of the AS can be improved by using the very same laser source. For instance, the edges of AS can be sharpened by ultrafast laser ablation, as shown in Figure 4
. The four sub-mm sized pillars were fabricated by the LASS process (by circulating the laser beam from the center to outer region), and they had a mushroom-shaped geometry right after the process (a side-on view of the mushroom-shaped AS is given as an inset in Figure 4
a). The cap part of the “mushroom” is trimmed to a pillar shape, as shown in Figure 4
a, after an additional ultrafast laser substractive scan along the contour of the AS at a reduced laser fluence (7 J/cm2
). The sharp edge is better viewed in Figure 4
b, where the sample is tilted by a nearly 90° angle. The pillars in this particular case are 250 µm in diameter and 100-µm high. Pillars with height (the elevation from the substrate surface to the peak of the pillars) up to approximately 300 µm and small diameters can also be made by this strategy; the side-on view is given in the Figure 4
b inset. The large surface production of the AS, which is similar to those in Figure 4
a, is also easily implemented, and an example is presented in Figure 4
c. The inset in Figure 4
c is a wetting test on the surface with a large array of sub-mm sized pillars, where a superhydrophobicity contact angle of 150 °C is observed (after a thermal treatment at 150 °C for two hours). Mechanical property-wise, the microindentation test revealed a decreased hardness at the surface of the AS, a Vickers hardness value at two N load HV2N
= 230 at the AS, compared to a Vickers HV2N
= 252 of the substrate. The reduction of the hardness is thought to be related to the presence of the micropores and reduced density in the AS compared to bulk aluminum. It is obvious that the substractive process step can be applied to remove the previously mentioned micropores lodged in the outermost surface region. If this surface porous layer were removed, the microhardness would be further improved. A more comprehensive study on this matter and process optimization will surely be meaningful toward further exploring the value of this additive and subtractive surface structuring approach.