Experimental Investigations on Laser Ablation of Aluminum in Sub-Picosecond Regimes

: Due to high power and ultrashort pulses, femtosecond lasers excel at (but are not limited to) processing materials whose thicknesses are less than 500 microns. Numerous experiments and theoretical analyses testify to the fact that there are solid grounds for the applications of ultrafast laser micromachining. However, with high costs and complexity of these devices, a sub-picosecond laser that might be an alternative when it comes to various micromachining applications, such as patterns and masks in thin metal foils, micro-nozzles, thermo-detectors, MEMS (micro electro-mechanical systems), sensors, etc. Furthermore, the investigation of sub-picosecond laser interactions with matter could provide more knowledge on the ablation mechanisms and experimental veriﬁcation of existing models for ultrashort pulse regimes. In this article, we present the research on sub-picosecond laser interactions with thin aluminum foil under various laser pulse parameters. Research was conducted with two types of ultrafast lasers: a prototype sub-picosecond Yb:KYW laser (650 fs) and a commercially available femtosecond Ti:S laser (35 fs). The results show how the variables such as pulse width, energy, frequency, wavelength and irradiation time a ﬀ ect the micromachining process.


Introduction
A scientific goal of this study is a better understanding of the physical phenomena and identification of ablation mechanisms in sub-picosecond regimes. A practical goal is to explore and enable sub-picosecond (i.e., in the range of several hundred of femtoseconds) lasers as an alternative to both femtosecond and picosecond systems in high precision micromachining processes.
Even though there are already femtosecond industrial lasers able to work for production environments, given the full advantages of material micromachining without the necessity of post-processing tools, they are not considered the first choice for many manufacturers. Still, the complex and unstable nature of femtosecond systems undermines their reliability and efficiency in commercial use, in contrast to the picosecond laser systems [1].
An effective micromachining process depends on the laser beam parameters, such as pulse duration, wavelength, energy, frequency or irradiation time, as well as target material properties [2]. These parameters affect the course of laser ablation, which engages different mechanisms than longer, i.e., picosecond and nanosecond, pulses [3,4]. We found this region between picosecond and femtosecond laser pulses interesting for further investigation, since both short and long pulse interactions with matter may appear.
Most of commercial femtosecond lasers offer the pulse length of tens of femtoseconds. In our research, we used a popular 35 fs Ti:S laser and a 650 fs prototype Yb:KYW laser. Both offered adjustable wavelength (UV/VIS/IR) and frequency. The pulse energy ranges were picked in the manner to make

Theoretical Background
The discovery of the chirped pulse amplification (CPA) method [6] and, in consequence, ultrafast lasers, opened a new field for both fundamental and applied research. Ablation that occurs during the athermal state of the sample in the dynamics of the pulse absorption, commonly known as cold ablation, became one of the most promising applications of ultrafast lasers.
The process occurs within a time range of tens of nanoseconds and differs significantly for long and ultrashort laser pulses. Most of the experiments are carried out with Gaussian beams and just above the ablation threshold. The ablation process involves stages such as: energy absorption, energy transfer to the lattice and material removal. After the pulse energy is deposed into the material, the excitation of electrons from the valance to the conduction band occurs, as well as the free carrier absorption. With significant beam intensity, either multiphoton or avalanche ionization appears. Nonlinear absorption is therefore a dominant mechanism in ultrashort pulse regimes [7].
The occurrence of the complex processes during laser interactions with matter is almost simultaneous, which makes it difficult to determine the contribution of each one. Depending on the material, on a timescale from hundreds of fs to a few ps, the energy transfer from electrons to the lattice takes place. Hence, the non-equilibrium state of the electrons and the lattice are best portrayed by a two-temperature model [8]. The transfer of energy results in an accelerated thermal or non-thermal melting [9]. Since the mass transport happens in much longer timescales than the non-thermal or thermal melting, the melted parts of material are left at a hot, near solid state. Subsequently, the hydrodynamic expansion of the ablated material begins, on a timescale of a few hundred ps after the initial excitation [10].
Despite the numerous investigations on the fundamentals of laser material removal, those in the femtosecond regime are still argued. To calculate the estimated ablation threshold in this study, several theoretical models were considered [11][12][13]. We found the following solution proposed by Gamaly et al. to be the most adequate: According to Equation (1), in case of metals, the ablation threshold's F th definition is based on the condition that the electron energy must reach the value equal to the sum of the atomic binding energy ε b and the work function ε esc ; η a is electron density [11]. By this definition, the threshold values for the aluminum samples used in our research have been calculated for each wavelength λ. The results are presented in Table 1.

Experimental Setup
The micromachining of aluminum samples was conducted with two different lasers, a sub-picosecond Yb:KYW (Lab-designed at the Institute of Physical Chemistry PAS, Warsaw, Poland, 2010) and a femtosecond Ti:S laser (Spectra-Physics, Santa Clara, CA, USA, 2012). The parameters are presented in Table 2. For the first laser, 1030 nm and 345 nm wavelengths were made available for the micromachining processes. For the second one, three different wavelengths from the UV, VIS and IR range were picked, as well as being used for the calculations in Table 1. Figure 1 shows that, as the laser beam passes through the optical collimator and focusing system, consisting of mirrors and lenses adequate for the applied laser wavelength, it is then transferred to the high-precision positioning system, based on XY linear motors. [14] Appl. Sci. 2020, 10 For the first laser, 1030 nm and 345 nm wavelengths were made available for the micromachining processes. For the second one, three different wavelengths from the UV, VIS and IR range were picked, as well as being used for the calculations in Table 1. Figure 1. shows that, as the laser beam passes through the optical collimator and focusing system, consisting of mirrors and lenses adequate for the applied laser wavelength, it is then transferred to the high-precision positioning system, based on XY linear motors. [14]

Results and Discussion
As it was previously stated, ablation mechanisms strongly depend on the laser beam parameters [2]. The methodology of the conducted experiments is presented in Table 3. Table 3. The summary of the experiment cases.

Type of Machining Parameter Change Results
Scribing (1 cm line) Pulse energy 7.5-57.5 µJ Figure 2 Prolonged irradiation Time 1-15 s Figure 3 Scribing ( Scribing with increasing pulse energy Wavelength 266-800 nm Figure 6 Pulse duration: 35 fs, 650 fs Figure 7 In the materials undergoing laser micromachining, there are various markers that can be connected to the dynamical mechanisms and the ablation threshold. Among them, there are crater

Results and Discussion
As it was previously stated, ablation mechanisms strongly depend on the laser beam parameters [2]. The methodology of the conducted experiments is presented in Table 3.
In the materials undergoing laser micromachining, there are various markers that can be connected to the dynamical mechanisms and the ablation threshold. Among them, there are crater depth and width, local changes in material structure, cracks, surface debris, and colour change caused by melting, to mention a few. To analyse some of those features, we used a Nikon Eclipse metallographic microscope (E100, Nikon Corporation, Tokyo, Japan, 2006) and a 3D Hirox KH8700 microscope (Hirox Co. Ltd., Tokyo, Japan, 2011). Post-machining crater dimensions were measured and the heat-affected changes on the surface of the samples were observed. The 3D analysis was applied to examine a cross-section of profiles and material structure changes. The representative results are presented below (Figures 2-5). Figure 2 presents a 0.406 mm thick aluminum sample machined with the sub-picosecond laser. Each 1 cm line was scribed with different laser pulse energy. It was observed that crater width decreases with decreasing energy. There were no visible effects of the heat affected zone and no debris on the surface, even for much higher energies than the ablation threshold. Line edges were straight and clean, except for energies exceeding 37.5 µJ. The best results were obtained for 7.5-17.5 µJ. Crater depths obtained from 3D analysis varied from 2 to 5 µm and were decreasing with pulse energy as well. All samples were machined with 25 mm/s beam transition speed.    Figure 4 shows the crater width decreasing with the increasing laser frequency. Although higher frequency results in tight overlapping of the laser pulses (the beam transition speed was constant, at 25 mm/s), the pulse energy decreases, resulting in a thinner line. However, the frequency change strongly affects the crater profile. Crater depths obtained from 3D analysis varied from 10 to 30 µm ( Figure 5).

Type of Machining Parameter Change Results
Scribing (1 cm line) Pulse energy 7.5-57.5 µJ Figure 2 Prolonged irradiation Time 1-15 s Figure 3 Scribing ( Scribing with increasing pulse energy Wavelength 266-800 nm Figure 6 Pulse duration: 35 fs, 650 fs Figure 7 Appl. Sci. 2020, 10, 8883 5 of 8 Figure 4 shows the crater width decreasing with the increasing laser frequency. Although higher frequency results in tight overlapping of the laser pulses (the beam transition speed was constant, at 25 mm/s), the pulse energy decreases, resulting in a thinner line. However, the frequency change strongly affects the crater profile. Crater depths obtained from 3D analysis varied from 10 to 30 µm ( Figure 5).    Each 1 cm line was scribed with different laser pulse energy. It was observed that crater width decreases with decreasing energy. There were no visible effects of the heat affected zone and no debris on the surface, even for much higher energies than the ablation threshold. Line edges were straight and clean, except for energies exceeding 37.5 µJ. The best results were obtained for 7.5-17.5 µJ. Crater depths obtained from 3D analysis varied from 2 to 5 µm and were decreasing with pulse energy as well. All samples were machined with 25 mm/s beam transition speed. the effect of the wavelength change was different. For aluminum, the material reflectivity is very high (above 90%) in UV-VIS, with a dip at 700-900 nm, and then increases with the wavelength in the far infrared range [15]. It can be predicted that the laser beam will be best absorbed in the near infrared window, and not in the UV range. This effect was confirmed for the Yb:KYW sub-picosecond laser, whereas for the femtosecond Ti:S laser, machining with 266 nm and 800 nm beam resulted in a similar crater size. The VIS range was poorly absorbed by aluminum sample; however, better absorption occurred for 532 nm in shorter, fs pulses, than for 345 nm in longer, sub-ps pulses. Appl. Sci. 2020, 10, x FOR PEER REVIEW 7 of 9 Figure 6. Ablation dependence on laser fluence for different wavelengths. Figure 7 shows a comparison between the results obtained for femtosecond and sub-picosecond laser. Vertical lines represent theoretical ablation thresholds, as calculated in Table 1. Other beam parameters were picked as close as achievable for both lasers. Pulse energy varied from 20 µ J to 80 µ J and the wavelength was in the IR range.
For the sub-picosecond laser, the experimental ablation threshold was closer to the theoretical predictions; however, for both lasers, the calculated values were lower than the experimental results. Despite the fact that the ranges of crater dimensions were similar in both cases, there is a significant difference in the results obtained close to the ablation threshold. The sub-picosecond characteristic is more steep with lower fluence. For the femtosecond laser, once a certain energy level is achieved, the curve begins to flatten.

Conclusions
With both crater width and depth increasing with the laser fluence, the increase in the heat affected zone was small or non-existent in the sub-picosecond regime. The aluminum sample has a high heat transfer coefficient, which could be considered in favor of the minimized heat effects. Considering future applications, for other materials, e.g., plastics, further shortening of the laser pulse might be recommended. The proposed Equation (1) for the ablation threshold can be applied in both (femto and sub-picosecond) cases, with better predictability in sub-picosecond regimes. The influence of laser wavelength is less significant in femtosecond regimes than it is in sub-picosecond regimes. This is possibly due to the involvement of non-linear absorption mechanisms (e.g., multiphoton absorption) occurring with ultrashort laser pulses. However, in our study, we focused on macroscopic results of the laser-matter interactions, without considering the specific heat transfer mechanisms. A further investigation on that matter is needed, including a broader variety of materials.
An upper limit for the highest usable frequency is given by the interaction of the generated plasma with succeeding laser pulses, since this will distort or shield the laser beam [16]. A successful application of the pulse repetition rate in the range of 100-400 kHz in our research enables the  The purpose of the experiment was to overheat the sample and enable the heat transfer to the area surrounding the target. Despite the prolonged time of laser-matter interaction and high pulse energy, no melting or significant color change (i.e., yellow to blue rings assigned to high temperature metal tempering) were noticed. The crater diameter was insignificantly increasing with the pulse energy. Figure 4 shows the crater width decreasing with the increasing laser frequency. Although higher frequency results in tight overlapping of the laser pulses (the beam transition speed was constant, at 25 mm/s), the pulse energy decreases, resulting in a thinner line. However, the frequency change strongly affects the crater profile. Crater depths obtained from 3D analysis varied from 10 to 30 µm ( Figure 5). Figure 6 presents the results obtained with a Ti:S laser, using different wavelengths and pulse energy. Comparable to the Yb:KYW laser, crater width decreases with decreasing pulse energy. Still, the effect of the wavelength change was different. For aluminum, the material reflectivity is very high (above 90%) in UV-VIS, with a dip at 700-900 nm, and then increases with the wavelength in the far infrared range [15]. It can be predicted that the laser beam will be best absorbed in the near infrared window, and not in the UV range. This effect was confirmed for the Yb:KYW sub-picosecond laser, whereas for the femtosecond Ti:S laser, machining with 266 nm and 800 nm beam resulted in a similar crater size. The VIS range was poorly absorbed by aluminum sample; however, better absorption occurred for 532 nm in shorter, fs pulses, than for 345 nm in longer, sub-ps pulses. Figure 7 shows a comparison between the results obtained for femtosecond and sub-picosecond laser. Vertical lines represent theoretical ablation thresholds, as calculated in Table 1. Other beam parameters were picked as close as achievable for both lasers. Pulse energy varied from 20 µJ to 80 µJ and the wavelength was in the IR range.
For the sub-picosecond laser, the experimental ablation threshold was closer to the theoretical predictions; however, for both lasers, the calculated values were lower than the experimental results. Despite the fact that the ranges of crater dimensions were similar in both cases, there is a significant difference in the results obtained close to the ablation threshold. The sub-picosecond characteristic is more steep with lower fluence. For the femtosecond laser, once a certain energy level is achieved, the curve begins to flatten.

Conclusions
With both crater width and depth increasing with the laser fluence, the increase in the heat affected zone was small or non-existent in the sub-picosecond regime. The aluminum sample has a high heat transfer coefficient, which could be considered in favor of the minimized heat effects. Considering future applications, for other materials, e.g., plastics, further shortening of the laser pulse might be recommended. The proposed Equation (1) for the ablation threshold can be applied in both (femto and sub-picosecond) cases, with better predictability in sub-picosecond regimes. The influence of laser wavelength is less significant in femtosecond regimes than it is in sub-picosecond regimes. This is possibly due to the involvement of non-linear absorption mechanisms (e.g., multiphoton absorption) occurring with ultrashort laser pulses. However, in our study, we focused on macroscopic results of the laser-matter interactions, without considering the specific heat transfer mechanisms. A further investigation on that matter is needed, including a broader variety of materials.
An upper limit for the highest usable frequency is given by the interaction of the generated plasma with succeeding laser pulses, since this will distort or shield the laser beam [16]. A successful application of the pulse repetition rate in the range of 100-400 kHz in our research enables the ultrashort lasers to meet industrial demand when it comes to processing speed. Considering all of the abovementioned observations, a sub-picosecond laser would allow better parameter control in micromachining processes and more predictable results, while maintaining the advantages of the femtosecond laser micromachining.
The obtained results allow identification of some of the ablation mechanisms in the sub-picosecond regime and show their dependence on laser pulse parameters. Therefore, they make it possible to verify the existing models and theories and increase the applicability of ultrafast lasers in commercial micro and nanomachining. A sub-picosecond laser proved to be a sufficient alternative to the micromachining of aluminum and possibly of other metals as well.