Open Access
Add to BibSonomy
Abstract
We report on crater formation, line scribing and cavity milling experiments on Silicon, Copper, Aluminum and stainless steel with GHz bursts of femtosecond pulses. The intra-burst repetition rate has been varied between 0.88 and 3.52 GHz, the number of pulses per burst between 50 and 3200, the burst fluence between 8 and 80 J/cm2. For these experiments, a 100-W femtosecond GHz-burst laser has been developed on an industrial laser basis, delivering a total burst energy up to 1 mJ at 100 kHz, with an adjustable number of pulses per burst. The results highlight the conditions to obtain high-ablation efficiency, show how to optimize the machining quality and point out the burst duration as the relevant parameter for femtosecond GHz machining.
© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Femtosecond laser micromachining is nowadays well established in industrial laser processing for a huge variety of applications [1] thanks to its machining accuracy down to the sub-micrometric level and the availability of compact and reliable industrial femtosecond laser systems. The drawback of relatively long processing times is overcome by the development of high-power and high-energy femtosecond laser systems allowing for parallelizing the machining processes using beam splitting devices such as spatial light modulators [2,3] or diffractive optical elements [4]. Another strategy consists in using lasers delivering the same energy at a higher repetition rate to speed up the machining process, which imposes the use of high-speed beam deflection devices such as polygonal scanners [5].
In this context, an increase of the femtosecond ablation efficiency may open new possibilities. Nowadays, femtosecond lasers can be operated in a mode delivering single pulses temporally dived in trains of high-repetition rate sub-pulses, called bursts. In this case, the sub-pulses have an energy corresponding to the single-pulse energy divided by the number of pulses within the burst. The fluence of each burst sub-pulse can be settled near the optimal fluence, defined as the fluence at which the ablation rate is the highest [6]. An increased ablation volume per pulse has been reported [7–9] but not exceeding a 10-30% efficiency increase compared to single-pulse ablation. In such configurations, the choice of the laser parameters (number of pulses per burst, delay between pulses in the burst, fluence per pulse) will strongly influence the efficiency of the process. In addition, thermal accumulation can also occur with damages affecting the machining quality, and a plasma shielding effect can lower the pulse absorption.
Very recently, a new kind of processing applying so-called GHz-bursts, i.e. pulse trains of femtosecond pulses with repetition rates in the GHz range within the pulse train, was demonstrated and showed exciting results in material processing featuring very high ablation rates [10–14]. However, also inferior ablation rates for high-speed bursts [15] and thermally damaged surfaces have been published [12–14] with different laser and processing parameters. Thus, the benefits or drawbacks of this GHz-burst technique are controversially discussed at the moment. Indeed, the results depend strongly on the applied laser parameters and demand a deeper understanding to provide an insight into the influence of the laser processing parameters on ablation rates and machining quality.
In this work, we present a thorough study of GHz-burst processing with respect to drilling, cutting and milling applications over a very large range of processing parameters including the total burst fluence, the intra-burst repetition rate, the total burst duration, and the number of pulses within the burst. The experiments were carried out on four different materials: Silicon, Copper, Aluminum, and stainless steel. Our results allow for an interpretation of general trends in GHz-burst femtosecond laser machining and give an overview on advantageous and disadvantageous configurations. Moreover, we discuss the influence of the different parameters on the ablation rate of drilling, cutting, and milling, as well as on the machining quality and surface homogeneity.
2. Experimental setup
2.1. Laser system
The laser system is based on a commercial laser (Tangor from Amplitude) and consists of a bulk oscillator, fiber-based components for burst handling and pre-amplification and a bulk main amplifier as depicted in Fig. 1. The solid-state oscillator is passively mode-locked and generates soliton-like pulses of 310 fs pulse duration at 1030 nm center wavelength. The oscillator cavity length is fixed in order to reach a pulse repetition rate of 0.88 GHz. The laser beam is fiber-coupled into a standard polarization-maintaining single-mode fiber (PM 980). To access to higher repetition rates with the same oscillator, fiber-based optional delay lines can be introduced at this level of the laser system by a pair of 50/50 couplers to achieve 1.76 GHz and by an additional one for 3.52 GHz. The femtosecond-pulses are then temporally stretched with a Chirped Fiber Bragg Grating (CFBG) for further amplification. Bursts are obtained by picking series of pulses from the original pulse train with an Acousto-Optic Modulator (AOM). Further direct amplification of these bursts would cause an inhomogeneity of the pulse energy within the burst due to gain depletion. To pre-compensate this effect, an energy distribution adapter is applied on the burst by the AOM, driven by an Arbitrary Waveform Generator (AWG). However, the rise and fall-times of the AOM impose an energy inhomogeneity for the first and last pulses of the burst that cannot be fixed in the same way. These pulses are removed by an Electro-Optic Modulator, presenting shorter rise and fall times than the AOM. These two components together, with the optional delay lines, allow to vary the pulse repetition rate, the burst repetition rate, the number of pulses per burst and the energy distribution inside the burst over a wide range. After the fiber-based pre-amplification chain, the GHz-bursts are amplified up to more than 100 W in an Yb-doped crystal-based amplifier. After the grating-based compressor, we obtain 100 W of average power. In this study, the burst repetition rate is fixed at 100 kHz, corresponding to a maximum burst energy of 1 mJ.
Fig. 1. Schematic of the GHz laser source. Col: fiber collimator, AOM: acousto-optic modulator, AWG: arbitrary waveform generator, EOM: electro-optic modulator and LD: pump laser diodes.
2.1.1 Experimental protocol
In order to study the processing conditions with respect to drilling, cutting and milling, we perform machining experiments on crater formation, line scribing and cavity formation as illustrated on Fig. 2. The laser beam seeds a two-axis Galvo scanner IntelliScan-14 III from Scanlab and is further focused on the sample with a 100-mm f(theta) lens. The incoming beam-diameter can be adjusted with a set of lenses. We obtain a 1/e2 beam diameter of 38.3 µm at the focal point for the experiments at 0.88 GHz repetition rate, and of 39.5 µm for the experiments at 1.76 and 3.52 GHz repetition rate, respectively. For line scribing, we have considered ten levels of burst fluence between 8 and 81 J/cm2 and eight levels of scanning speed (0.1, 0.25, 0.50, 1.0, 2.0, 3.0, 4.0 and 5.0 m/s), corresponding to overlaps from 97,4% to 0% (no overlap, individual craters). For each tested condition, we applied 1 and 30 passes along 2.5-mm lines. The cavities are obtained by applying series of lines, keeping the same overlap on the X and Y axes (fixed overlap of 47.8% at 2.0 m/s). The same ten levels of burst fluence have also been tested for two numbers of passes, 1 and 15, creating cavities of (2 × 1) mm2. All experiments were performed on 500 µm-thick polished samples of four different materials: Silicon, Copper, Aluminum and stainless steel (316L). The tested burst configurations are summarized in Table 1.
Fig. 2. (a) Schematic drawing of the experimental setup for micromachining experiments, and (b) illustration of overlap and scanning speed (vscan) for crater formation, line scribing and cavity milling, respectively.
Table 1. Maximal specific ablation rate Eff (mm3/min/W) and corresponding burst fluence F (J/cm2) for investigated laser parameters: RR: intra burst repetition rates (GHz), Nb: number of pulses in the burst, T: burst duration (ns).
2.2. Measure and analysis
The profile of each crater was measured with a Leica DCM 3D confocal microscope (objective x50). The ablated volume is obtained by averaging the measured volume of three craters (volume below z=0) for each burst configuration. The maximal precision for the z-position is 0.2 µm and the accuracy of the actual position of the z=0 plane is limited by the initial Sa roughness of the sample. Therefore, we assume an error of +/- (0.2 µm + Sa roughness) on the z position of each measured point.
For line scribing, the profile was measured with the same confocal microscope (objective x50). Then, the 3D profile is divided into a series of 2D sections for which the depth and width of the line, the ablated section (below z=0) and the section of redeposited material (above z=0) are measured. We considered the same error for the trench measures as for the craters. The characteristics of the scribing trenches are calculated by averaging multiple sections along the line and the specific ablation rate in mm3/(min/W) is computed. For deep lines, only a few points can be measured inside the trench, making it impossible to calculate accurate values of the profile. These measurements are then rejected.
For cavity analysis, the same confocal microscope is used to measure the bottom surface roughness of the cavities (Sa, arithmetical mean height). The Sa roughness of the samples before machining were respectively 0.15 µm for Silicon, 0.35 for Copper, 0.33 for Aluminum and 0.54 for stainless steel. The depth of the machined cavities is measured with a Mitutoyo MF-B1010D microscope (objective x20), to retrieve the specific ablation rate. Following the specifications of this microscope, we assume an error of +/- 5.0 µm on depth measurement.
3. Results
3.1. Ablation efficiency
First, we present a comparison of ablation efficiencies for crater formation, line scribing and cavity milling. We have performed experiments for different burst durations, total burst fluences, intra-burst repetition rates and number of pulses within the burst. The different specific ablation rates for varying total burst fluences are depicted in Fig. 3 for Silicon, in the case of an intra-burst repetition rate of 0.88 GHz and 200 pulses per burst (corresponding to a total burst duration of 228 ns). The burst-to-burst overlap was fixed at 48% for both line scribing and cavity milling. In Fig. 3, the specific ablation rates show the highest values for crater formation (circles) which are slightly exceeding the ones for line scribing (lines), whereas the ones for cavity milling (squares) have lower values.
Fig. 3. Specific ablation rate of Silicon as a function of burst fluence obtained with bursts of 200 pulses at 0.88 GHz intra-burst repetition rate, corresponding to a total burst duration of 228 ns, for crater formation (open circles), line scribing (lines) and cavity milling (squares). The burst-to-burst overlap was fixed at 48% for lines and cavities. The dashed lines are guides to the eye.
To study the influence of the burst parameters on the specific ablation rate, we explored different burst configurations for cavity milling using a fixed number of passes (15) and the same overlap (48%). Either we set the intra-burst repetition rate with varying numbers of pulses per burst (longer burst duration with higher number of pulses), or we fixed the number of pulses per burst and changed the intra-burst repetition rates. We carried out the experiments on all four selected materials with three combinations of intra-burst repetition rate and number of pulses in order to generate bursts of equal duration. For example, the below three experimental conditions, 0.88 GHz repetition rate with 800 pulses per burst, 1.7 GHz with 1600 pulses per burst, and 3.5 GHz with 3200 pulses per burst, all correspond to a 912-ns burst duration. The results are shown in Fig. 4 for Silicon (a), Copper (b), Aluminum (c), and stainless steel (316L) (d). For a better comparison, we chose a color code in the figures: the same color corresponds to the same burst duration: red, green and blue stand for 912 ns, 228 ns, and 57 ns, respectively. We also encoded the intra-burst repetition rate by applying same symbols: triangles, dots and squares, respectively, stand for 0.88, 1.76 and 3.52 GHz.
Fig. 4. Specific ablation rate as a function of burst fluence obtained with different GHz bursts by machining of (2 × 1) mm2 cavities, 15 passes for Silicon (a), Copper (b), Aluminum (c) and stainless steel (d). The legends are valid for all graphs in Fig. 4 (and in Fig. 5), where red, green and blue correspond to burst durations of 912, 228 and 57 ns, respectively. Triangles stand for 0.88 GHz, dots for 1.76 GHz and squares for 3.52 GHz intra-burst repetition rate. The dashed lines are guides to the eye.
For a fixed intra-burst repetition rate, the specific ablation rate rises with the number of pulses per burst (in the case of 3.52 GHz, this is only true for fluences above 30 J/cm2 for Silicon, Aluminum, stainless steel and above 60 J/cm2 for Copper). On the opposite, when the number of pulses per burst is fixed and the intra-burst repetition rate varies (for instance, 200 pulses per burst at 0.88 GHz and at 3.52 GHz), higher specific ablation rates are obtained with lower intra-burst repetition rates. A comparison of different burst configurations characterized by a fixed burst duration (curves of the same color in Fig. 4) shows that similar specific ablation rates are achieved for 0.88 and 1.76 GHz. With an intra-burst repetition rate of 3.52 GHz, the behavior is slightly different, as the configuration with 3200 pulses per burst shows a different trend than those with 200 and 800 pulses per burst (especially for Copper and Aluminum).
The influence of intra-burst repetition rate, number of pulses per burst and the machining process itself (drilling, cutting and milling) on the specific ablation rate are summarized in Table 1, where are given the maximal specific ablation rates (Eff) reached for all tested laser configurations and for each material together with the corresponding fluences (F).
In previous work [14], we measured a maximal specific ablation rate of 2.5 mm3/min/W for crater formation on Silicon with a 200-pulse burst at 0.88 GHz intra-burst repetition rate. In Fig. 3, for the same experiment, a maximal specific ablation rate of 1.6 mm3/min/W is observed, 37% lower than in [14]. We used the same method to compute the ablation rate for both experiments. The main difference in the experimental conditions was the spot size: 24 µm in [14] and 38.3 µm in the present work. We assume that the different spot sizes explain the reduced ablation rate. Indeed, a strong effect of the laser spot size has been already reported on the ablation rate for spot diameters below 80 µm [16–18], where this phenomenon is explained by a stronger absorption of the incident beam due to stronger plume shielding with bigger spot sizes. Another difference in the experimental conditions between our two experiments was the energy distribution among the pulses of the burst. In this paper, this energy had been homogenized as explained in 2.1 while it was not the case in [14].
3.2. Machining quality
In laser processing with GHz-bursts, the increase of ablation efficiency can lead to a degradation in quality. We first address the case of line scribing (relevant for cutting processes) to quantify the machining quality and study the influence of the burst parameters. In this case, the main effect which leads to quality degradation is the material redeposition outside the trench. We evaluated the quality of engraving by, first, measuring 3D profiles of the sections of redeposited and ablated matter, and then, by assigning a quality factor Q defined as 1 – [(redeposited section) / (ablated section)]. A value close to 1 for the quality factor Q indicates very high quality, where all the removed matter is vaporized and not redeposited around the groove, so that no burr along the scribed line is formed.
Figure 5 presents the results of line engraving on Silicon with three different numbers of pulses per burst (100, 400 and 1600) at a fixed intra-burst repetition rate of 1.76 GHz. The four graphs depict the mean section of ablated material (a), the mean section of the redeposited material (b), the specific ablation rate (c) and the corresponding quality factor Q (d) for different burst fluences. We kept the color code, where red, green and blue stand for 912 ns, 228 ns, and 57 ns burst duration, respectively.
Fig. 5. Results obtained by line scribing of Silicon with 1.76 GHz intra-burst repetition rate bursts: mean section of ablated material (a), mean section of redeposited material (b), specific ablation rate (c) and quality factor Q (d) versus burst fluence. The inset in graph (d) schematically presents how the quality factor Q is computed. The inset in graph (b) depicts line profiles for the three burst configurations at a near-40 J/cm2 burst fluence level. We kept the color legend of Fig. 4, where red, green and blue correspond to burst durations of 912, 228 and 57 ns, respectively. The dashed lines are guides to the eye.
The graph (c) with the specific ablation rate versus burst fluence indicates that, as previously mentioned in the case of milling, that the highest ablation efficiencies are reached with the longest bursts. The Q factor curves in (d) are showing the same tendency for all three different burst durations. The longest bursts present the highest ablated matter sections and lowest redeposited matter corresponding to the highest Q factor. In the range of the tested burst fluences between 7 and 83 J/cm2 for burst configurations with 100 and 400 pulses per burst, and between 23 and 47 J/cm2 for the 1600-pulse burst, the Q factor stays, respectively, nearly constant. Consequently, for line scribing the ablation efficiency and the quality factor can be controlled over a wide range of fluences by adjusting the burst duration.
3.3. Surface homogeneity
In order to study the surface quality obtained for milling, we measured the cavities’ bottom surface with the confocal microscope and computed the roughness as the 2D arithmetical mean deviation Sa. The evolution of this roughness with the burst fluence for Copper, Aluminum, stainless steel and Silicon (with bursts of 1.76 GHz intra-burst repetition rate and 100 pulses per burst) is presented in Fig. 6 (right). An additional observation of these surfaces has been performed with a Phenom ProX scanning electron microscope (SEM). These SEM images are depicted in Fig. 6 (left) for Aluminum samples, for which the roughness is the highest, for five different burst configurations, but at constant burst fluence of around 15 J/cm2.
Fig. 6. Scanning electron microscope images of the cavities’ bottom surface of Aluminum samples for five different burst configurations: 0.88 GHz and 200 pulses per burst, 1.76 GHz and 100, 400 and 1600 pulses per burst and 3.52 GHz and 800 pulses per burst (left), all with fluences close to 15 J/cm2. Graph of the roughness Sa of the cavities’ bottom surface versus the burst fluence for Copper, Aluminum, stainless steel and Silicon (right). The burst parameters are 1.76 GHz intra-burst repetition rate and 100 pulses per burst, corresponding to 57 ns burst duration.
A comparison of the SEM images of the Aluminum samples for 1.76 GHz intra-burst repetition rate shows an increasing cavity bottom surface roughness with an increasing number of pulses per burst, whereas the three bursts of the same burst duration (228 ns) present a similar cavity bottom surface roughness. As shown in Fig. 6 (right), the roughness of the cavity bottom surface is strongly increasing with the burst fluence, especially for Aluminum (from a Sa roughness of 1.32 µm at 7.0 J/cm2 to 5.49 µm at 37.5 J/cm2). Copper is the only tested material presenting a Sa roughness below 1.0 µm up to 37.5 J/cm2.
3.4. Role of overlap
The above presented results of GHz-burst machining for line scribing and cavity milling have been obtained with a fixed burst-to-burst overlap of 48%. This value is different from the one usually used in fs processing with single pulses of around 70% [5]. This deliberately chosen value allows for studying a wider range of burst fluences. Indeed, for high fluences, machining with a pulse overlap higher than 70% often leads to a refill of the groove with melted material, especially for long burst durations, making it impossible to measure the ablated volume. Nevertheless, having a closer look on the influence of the overlap on the ablation rate and machining quality for short burst durations underlines the specificity of GHz-bursts with respect to single-pulse processing.
Figure 7 depicts the graphs of the specific ablation rate (a) and quality factor Q (b) versus burst fluence for different burst-to-burst overlaps (48, 73, 87 and 94%). All these data correspond to a one-pass engraving on Copper, with a 50-pulse burst at 0.88 GHz intra-burst repetition rate. We observe that the overlap variation has a moderate influence on the ablation efficiency (maximum variation of 22% at 23.8 J/cm2) in comparison to the burst parameters, such as the burst duration. The same observation can be done concerning the quality factor evolution with the overlap.
Fig. 7. Specific ablation rate obtained with bursts at a repetition rate of 0.88 GHz and counting 50 pulses for line scribing on a Copper sample (a). The four curves present the ablation rates for four different burst-to-burst overlaps versus burst fluence. Quality factor Q versus burst fluence for the same conditions (b). The dashed lines are guides to the eye.
4. Discussion
The limited influence of the overlap on the ablation efficiency for short burst durations and the refilling of the groove for long burst durations and strong overlap can be interpreted as a result of thermal accumulation which occurs differently for GHz-burst machining than for standard single-pulse machining. In the single-pulse regime, a variation of spatial overlap allows to induce or not thermal accumulation, depending on the pulse repetition rate [5,19,20]. If the overlap is too low, there is no thermal accumulation, an intermediate overlap allows for benefitting of thermal accumulation which enhances the ablation rate, and detrimental effects appear for too high overlap (uncontrolled melting, burrs). In the GHz regime, thermal accumulation already occurs within the burst and rises with the burst duration [21], which mitigates the effects of spatial overlap on thermal accumulation for short burst durations, and therefore on the ablation rate, but makes critical the use of strong overlap for long burst durations. Therefore, a weaker overlap than for single pulse machining should be used for GHz burst machining (close to 50% instead of 70%).
As a direct conclusion of the above presented results, the most relevant parameter of GHz-bursts for the ablation efficiency is the burst duration. Among all tested intra-burst repetition rates (0.88, 1.76 and 3.52 GHz), the 228-ns bursts showed a maximal specific ablation rate up to twice the one of the 57-ns bursts. For the 912-ns bursts, we observed even three to four times higher ablation rates. This is true for craters, lines and cavities and for each of the tested materials. As a consequence, to benefit from the high GHz-burst machining efficiency, it is necessary to privilege bursts of long enough duration, counting up to several hundreds of pulses (depending on the intra-burst repetition rate). It also makes it possible to precisely tune the ablation rate with the number of pulses per burst at a given intra-burst repetition rate. Notice also that we performed most of our experiments in the case of bursts with an individual pulse fluence far below the ablation threshold of the machined material.
Concerning the GHz-burst machining quality, it appears that longer bursts, providing the highest ablation rates, are also showing the highest ablation qualities for line scribing (quality being defined as redeposited matter on ablated matter ratio). Nevertheless, for all tested configurations, a portion of the ablated matter will redeposit along the groove as a burr, which is a signature of thermal effects. This fact points out one specific aspect of GHz-burst ablation: a part of the irradiated matter is melted and displaced instead of being evaporated [15,22]. However, for milling, long-bursts leading to high ablation efficiencies have also presented the most degraded surfaces. Compared to results in literature, the roughness produced on copper is similar to those obtained with single pulse machining [23]. On silicon lower values are reported for single-pulse machining at slightly lower fluences [6]. It is thus not possible to conjugate high efficiencies and low surface roughness in the case of a one-step milling process. But recently published work [24] shows that using bursts counting a large number of low-energy pulses, at a GHz-level repetition rate, allows for obtaining high-quality polishing and with higher efficiency than MHz-burst polishing. As proposed by the authors, a two-step machining process, first ablation with moderate-energy GHz-bursts and then polishing with low-energy GHz-bursts, can allow for overcoming the degraded milling-quality of long GHz-bursts.
The benefits of using GHz-bursts for micromachining are widely discussed, both regarding quality and efficiency compared with machining by single pulses in the femtosecond regime, or by lower intra-burst repetition rate bursts (tens of MHz) or even by single nanosecond pulses. Recent publications allow for an interesting comparison of the ablation rate of these different micromachining approaches. The lack of comparable data on ablation quality in literature do not allow us to conclude on the benefits of one approach with respect to the others. In the case of milling, extensively discussed in this work, Table 2 gathers the maximum specific ablation rates reached in literature for Silicon, Copper and stainless steel, compared for four different laser configurations: single fs pulses, single ns pulses, MHz-bursts and GHz-bursts. Compared with single fs pulses, GHz-bursts are only significantly more efficient for long burst durations. For instance, for copper, 57 ns-bursts (this paper) and 4.6 ns bursts [15] present lower ablation rates than standard fs machining [15,25]. GHz-burst benefits regarding fs machining only appear for 228- and 912-ns GHz-bursts. This observation remains true when comparing GHz-burst and ns-pulse or MHz-burst ablation rates. It appears that there is a burst duration threshold beyond which GHz-burst machining becomes beneficial for increasing the ablation rate.
Table 2. Comparison of maximal specific ablation rates (in mm3/min/W) reached for milling for Silicon, Copper and stainless steel in this study and results from literature. The values are given for single femtosecond-pulses [15,25], MHz-bursts [25], GHz-bursts [15] and single nanosecond pulses [25].
Using a laser source delivering up to 100 W average power allowed us to work with burst fluences up to 80 J/cm2, but we kept the fluence of each sub-pulse of the burst constant and below the ablation threshold, for each material [26–28]. As it has been shown by Povarnitsyn et al. [29,30], and in accordance with all so-far published results on GHz-burst machining to our knowledge, the use of sub-threshold pulse fluences is necessary to reach high ablation rates. Moreover, we have shown that the optimal burst fluences correspond to pulse fluences far below the fluence threshold of single-pulse ablation. In our previous work [14] we have suggested a physical interpretation of GHz ablation in a two-step scheme, which is fully confirmed by the present results. In the first step, matter is slowly heated by thermal accumulation [31], without ablation, thanks to the first pulses of the burst. The thermal rise then reduces the ablation threshold [26]. In a second step, when the ablation threshold is reduced down to the pulse fluence, each subsequent pulse ablates matter very efficiently, since optimal fs ablation is always reached just above the ablation threshold. Our previous work on crater formation in Silicon underlines also a threshold for the burst fluence in GHz ablation. For fluences immediately above this threshold, a deep ablation is observed, corresponding to the thermal length of the material. A very efficient thermal process thus occurs, whose signature is always seen in GHz ablation. If the non-thermal fs ablation phase (the second step in the previous scheme) has not enough time to efficiently occur due to a too small number of pulses in the burst, the thermal nature of ablation and its consequences on the quality are enhanced. With an optimal burst fluence and an optimal number of pulses in the burst, high-volume ablation is achieved with a ratio of ablated volume/redeposited volume compatible with an acceptable surface quality as shown in the line ablation results of the present work.
On the contrary, when the fluence of each pulse of the burst is higher than the ablation threshold, the ablation starts from the first pulse and the first step of accumulative heating cannot take place. The ablation process is thus similar to those occurring in the MHz-burst regime as shown by results obtained in such conditions [15]. If the pulse fluence is lower than the ablation threshold but the number of pulses within the burst is too low, or if the fluence is too low, only a few pulses can contribute to ablation resulting in inefficient ablation or simple bump formation by thermal modification of the surface, as shown in recently published work [14].
As shown in this study using a large enough range of laser parameter variations, the specificity of GHz ablation has to be carefully mastered, especially in terms of burst duration, to obtain an optimum result both in ablation efficiency and quality [32]. A simple physical approach can relieve the determination of optimal laser parameters. The burst fluence threshold has to be estimated for a given material, and the minimal burst length deduced. Then an optimal fluence can be found as it is always the case in fs ablation near the threshold, and an optimal number of pulses can also be found to maximize “non-thermally” ablated volume versus thermally ablated volume.
5. Conclusion
The systematic experimental study presented in this work shows the possible balance between ablation efficiency and processing quality. The burst duration and number of pulses in the burst are the key parameters of fs GHz processing. We can derive not only an optimal ablation efficiency but also a control of the thermal effects added on the known non-thermal femtosecond ablation. The present results underline expanded perspectives for fs GHz lasers to laser processing based on thermal modification, such as welding of all materials (metals, ceramics, dissemble materials) or surface polishing, becoming reachable with the usual precision of fs pulses. A laser developed on a flexible industrial laser basis (up to 100 W in IR, 25 W in UV [33], energy-shaped bursts for a high number of pulses can be used to increase the panel of laser applications open to femtosecond processing, not only due to higher ablation efficiencies.
Funding
Direction Générale de la Compétitivité, de l’Industrie et des Services (172906051); Association Nationale de la Recherche et de la Technologie (2017/0632).
Disclosures
The authors declare no conflicts of interest.
References
1. S. Nolte, F. Schrempel, and F. Dausinger, eds., Ultrashort Pulse Laser Technology: Laser Sources and Applications, Springer Series in Optical Sciences (Springer International Publishing, 2016), 195.
2. N. Matsumoto, H. Itoh, T. Inoue, T. Otsu, and H. Toyoda, “Stable and flexible multiple spot pattern generation using LCOS spatial light modulator,” Opt. Express 22(20), 24722 (2014). [CrossRef]
3. J. Li, Y. Tang, Z. Kuang, J. Schille, U. Loeschner, W. Perrie, D. Liu, G. Dearden, and S. Edwardson, “Multi imaging-based beam shaping for ultrafast laser-material processing using spatial light modulators,” Opt. Lasers Eng. 112, 59–67 (2019). [CrossRef]
4. S. Torres-Peiró, J. González-Ausejo, O. Mendoza-Yero, G. Mínguez-Vega, P. Andrés, and J. Lancis, “Parallel laser micromachining based on diffractive optical elements with dispersion compensated femtosecond pulses,” Opt. Express 21(26), 31830 (2013). [CrossRef]
5. J. Lopez, K. Mishchik, G. Mincuzzi, E. Audouard, E. Mottay, and R. Kling, “Efficient Metal Processing Using High Average Power Ultrafast Laser,” J. Laser MicroNanoengineering 12(3) (2017).
6. B. Neuenschwander, B. Jaeggi, D. J. Foerster, T. Kramer, and S. Remund, “Influence of the burst mode onto the specific removal rate for metals and semiconductors,” J. Laser Appl. 31(2), 022203 (2019). [CrossRef]
7. T. Kramer, Y. Zhang, S. Remund, B. Jaeggi, A. Michalowski, L. Grad, and B. Neuenschwander, “Increasing the Specific Removal Rate for Ultra Short Pulsed Laser-Micromachining by Using Pulse Bursts,” J. Laser Micro/Nanoeng. 12(2), 107–114 (2017). [CrossRef]
8. J. Mur, L. Pirker, N. Osterman, and R. Petkovšek, “Silicon crystallinity control during laser direct microstructuring with bursts of picosecond pulses,” Opt. Express 25(21), 26356 (2017). [CrossRef]
9. A. Žemaitis, P. Gečys, M. Barkauskas, G. Račiukaitis, and M. Gedvilas, “Highly-efficient laser ablation of copper by bursts of ultrashort tuneable (fs-ps) pulses,” Sci. Rep. 9(1), 12280 (2019). [CrossRef]
10. C. Kerse, H. Kalaycıoğlu, P. Elahi, B. Çetin, D. K. Kesim, Ö Akçaalan, S. Yavaş, M. D. Aşık, B. Öktem, H. Hoogland, R. Holzwarth, and FÖ Ilday, “Ablation-cooled material removal with ultrafast bursts of pulses,” Nature 537(7618), 84–88 (2016). [CrossRef]
11. P. Elahi, Ö Akçaalan, C. Ertek, K. Eken, FÖ Ilday, and H. Kalaycoğlu, “High-power Yb-based all-fiber laser delivering 300 fs pulses for high-speed ablation-cooled material removal,” Opt. Lett. 43(3), 535 (2018). [CrossRef]
12. H. Matsumoto, Z. Lin, and J. Kleinert, “Ultrafast laser ablation of copper with ∼GHz bursts,” in Proc. SPIE 10519, Laser Applications in Microelectronic and Optoelectronic Manufacturing (LAMOM) XXIII, 1051902 (2018).
13. G. Bonamis, K. Mishchick, E. Audouard, C. Hönninger, E. Mottay, J. Lopez, and I. Manek-Hönninger, “High efficiency femtosecond laser ablation with gigahertz level bursts,” J. Laser Appl. 31(2), 022205 (2019). [CrossRef]
14. K. Mishchik, G. Bonamis, J. Qiao, J. Lopez, E. Audouard, E. Mottay, C. Hönninger, and I. Manek-Hönninger, “High-efficiency femtosecond ablation of silicon with GHz repetition rate laser source,” Opt. Lett. 44(9), 2193 (2019). [CrossRef]
15. T. Hirsiger, M. Gafner, S. M. Remund, M. W. Chaja, A. Urniezius, S. Butkus, and B. Neuenschwander, “Machining metals and silicon with GHz bursts: Surprising tremendous reduction of the specific removal rate for surface texturing applications,” in Proc. SPIE 11267, Laser Applications in Microelectronic and Optoelectronic Manufacturing (LAMOM) XXV, 112670 T (2020).
16. M. Eyett and D. Bäuerle, “Influence of the beam spot size on ablation rates in pulsed-laser processing,” Appl. Phys. Lett. 51(24), 2054–2055 (1987). [CrossRef]
17. M. E. Shaheen, J. E. Gagnon, and B. J. Fryer, “Excimer laser ablation of aluminum: influence of spot size on ablation rate,” Laser Phys. 26(11), 116102 (2016). [CrossRef]
18. M. Chaja, T. Kramer, and B. Neuenschwander, “Influence of laser spot size and shape on ablation efficiency using ultrashort pulse laser system,” Procedia CIRP 74, 300–304 (2018). [CrossRef]
19. A. Ancona, S. Döring, C. Jauregui, F. Röser, J. Limpert, S. Nolte, and A. Tünnermann, “Femtosecond and picosecond laser drilling of metals at high repetition rates and average powers,” Opt. Lett. 34(21), 3304 (2009). [CrossRef]
20. F. Bauer, A. Michalowski, T. Kiedrowski, and S. Nolte, “Heat accumulation in ultra-short pulsed scanning laser ablation of metals,” Opt. Express 23(2), 1035 (2015). [CrossRef]
21. J. Finger, B. Bornschlegel, M. Reininghaus, A. Dohrn, M. Nießen, A. Gillner, and R. Poprawe, “Heat input and accumulation for ultrashort pulse processing with high average power,” Adv. Opt. Technol. 7(3), 145–155 (2018). [CrossRef]
22. A. D. Zweig, “A thermo-mechanical model for laser ablation,” J. Appl. Phys. 70(3), 1684–1691 (1991). [CrossRef]
23. J. Schille, L. Schneider, P. Lickschat, U. Loeschner, R. Ebert, and H. Exner, “High-pulse repetition frequency ultrashort pulse laser processing of copper,” J. Laser Appl. 27(S2), S28007 (2015). [CrossRef]
24. F. Nyenhuis, A. Michalowski, and J. A. L’huillier, “Surface treatment with GHz-bursts,” Proc. SPIE 11268, 112680B (2020). [CrossRef]
25. M. Domke, V. Matylitsky, and S. Stroj, “Surface ablation efficiency and quality of fs lasers in single-pulse mode, fs lasers in burst mode, and ns lasers,” Appl. Surf. Sci. 505, 144594 (2020). [CrossRef]
26. J. Thorstensen and S. Erik Foss, “Temperature dependent ablation threshold in silicon using ultrashort laser pulses,” J. Appl. Phys. 112(10), 103514 (2012). [CrossRef]
27. J. P. Colombier, P. Combis, F. Bonneau, R. Le Harzic, and E. Audouard, “Hydrodynamic simulations of metal ablation by femtosecond laser irradiation,” Phys. Rev. B 71(16), 165406 (2005). [CrossRef]
28. P. Mannion, J. Magee, E. Coyne, and G. M. O’Connor, “Ablation thresholds in ultrafast laser micromachining of common metals in air,” Proc. SPIE 4876, 470 (2003). [CrossRef]
29. M. E. Povarnitsyn, P. R. Levashov, and D. V. Knyazev, “Simulation of ultrafast bursts of subpicosecond pulses: In pursuit of efficiency,” Appl. Phys. Lett. 112(5), 051603 (2018). [CrossRef]
30. M. E. Povarnitsyn and P. R. Levashov, “Simulation of single-, double- and multi-pulse laser ablation of metals,” J. Phys.: Conf. Ser. 1147(1), 012063 (2019). [CrossRef]
31. S. M. Eaton, H. Zhang, P. R. Herman, F. Yoshino, L. Shah, J. Bovatsek, and A. Y. Arai, “Heat accumulation effects in femtosecond laser-written waveguides with variable repetition rate,” Opt. Express 13(12), 4708 (2005). [CrossRef]
32. G. Bonamis, E. Audouard, C. Hönninger, J. Lopez, K. Mishchik, E. Mottay, and I. Manek-Hönninger, Méthode de Détermination Des Conditions Opérationnelles d’un Procédé d’ablation Laser Femtoseconde à Très Haute Cadence Pour Un Matériau Donné, Patent Pending (2019).
33. G. Bonamis, J. Sanabria, E. Audouard, J. Lopez, I. Manek-Hönninger, E. Mottay, and C. Hönninger, “20-W Ultraviolet Femtosecond GHz Burst Laser,” in 2019 Conference on Lasers and Electro-Optics Europe & European Quantum Electronics Conference (CLEO/Europe-EQEC) (IEEE, 2019).
