1. INTRODUCTION

Laser ablation has been established for several decades due to its versatility and accuracy, although the relatively high-energy consumption makes the application uneconomical in some cases. The enhancement of the efficiency of laser-based processes is thus an important issue. Against this background, coupling atmospheric pressure plasmas (APP) and laser irradiation has the potential to be a promising approach to improve the overall degree of efficiency of ablation processes. Recent studies show that such combination has a number of advantages in comparison to pure laser ablation. For example, the ablation rate can be increased significantly when introducing direct dielectric barrier discharge ($D$-DBD) plasmas to front-side laser drilling or cutting processes. In the case of optical glasses, an increase in ablation depth by a factor of 2.1 was achieved by $D$-DBD APP-assisted laser drilling [1]. Higher ablation rates were also observed for ceramics (aluminum oxide) [2] and metals (aluminum) [3]. Currently, this effect is not yet fully understood, and different possible underlying mechanisms were discussed in the past [1–3]. Most probably, the debris plume that occurs during laser ablation plays a key role since it represents a medium for de-excitation of excited (metastable) plasma species by collisions. The resulting release of energy could contribute to an increased ablation rate in different ways. First, energy could be transferred to the substrate and thus contribute to material removal, and, second, debris particles could be decomposed or even evaporated by the energy provided by the plasma species. In the latter case, the attenuation of the incoming laser beam by scattering, absorption, and reflection at debris particles is reduced, consequently resulting in a higher laser fluence on the work piece surface.

Against the background of these basic considerations, the investigation of the impact of $D$-DBD APP on the particle formation during laser ablation is of great interest. Such investigation was performed in the present work where titanium (Ti) was chosen as sample material. Due to its specific properties, such as low density, high corrosion resistance, and high mechanical strength [4], this element has a number of technical applications. It is used in the automotive and space industries as well as for several medical applications, such as implant technology [5–8]. Ablation of titanium and titanium alloys can be realized by different laser sources such as ${\mathrm{CO}}_2$ lasers [9] or Nd:YAG-lasers [10,3]. In order to minimize the thermal impact on the material and to improve the machining quality during long-pulse ablation, different approaches such as underwater ablation [11] or laser chemical machining [12] were investigated in the past. Another promising method is the use of femtosecond lasers [13–17].

A continuously growing field of laser application on titanium substrates is the production of titanium micro- or nanoparticles. Such particles play an increasing role in the fields of catalysis, biomedical engineering, and solar energy conversion and storage. The laser-generated particles can be collected in vacuum or liquids [18–21]. As a liquid collector medium, deionized water or ethanol is used and, mostly, picosecond or femtosecond laser sources are employed for this task. The sizes of the generated nanoparticles mainly range from 2 up to 200 nm, while the nanoparticle shape is spherical [22–24]. In order to reduce particle agglomeration (consequently resulting in an increase in particle size and the formation of non-spherical, complexly-shaped particles), the impact of electric fields on the particle formation was investigated recently [25].

In contrast to this application, particle formation can represent an unwanted and disturbing effect during laser ablation. In terms of machining quality, the re-deposition of particles on the work piece surface close to the ablation spot leads to the formation of surface-adherent debris. Such debris can have a severe effect on the functionality of micro structures. Moreover, nano- and microparticles represent a potential health risk since particles of this size range are respirable and can enter the stroma [26]. Cytotoxic particles could thus cause diseases such as inflammations [27–30].

The investigation and characterization of laser ablation processes regarding the resulting particle formation thus addresses a number of fields of research and applications, as for example, (i) the study of mechanisms leading to debris formation, (ii) the production of nano- and microparticles, and (iii) the evaluation of hybrid laser-based processes. The latter aspect, i.e., the investigation of $D$-DBD APP-assisted laser ablation, represents the main issue of the present work.

2. MATERIALS AND METHODS

The used experimental setup for the investigation of the impact of assisting plasma on the particle formation during laser ablation of titanium is shown in Fig. 1. This setup was already used in previous work for the investigation of plasma-assisted nanosecond laser ablation of glasses [1], aluminum [3], and ceramics [2]. In the present work, a Yb:YAG picosecond laser (Trumpf, model TruMicro 5025) with a pulse duration of $\tau =8\mathrm{ps}$, an emission wavelength of $\lambda =1030\mathrm{nm}$, and a basic pulse repetition rate of $f_{\mathrm{rep}}=200\mathrm{kHz}$ was used as the laser source. This laser source was operated in burst mode, resulting in an effective burst repetition rate of $f_{\mathrm{burst}}=20\mathrm{Hz}$, where the burst duration was $t_{\mathrm{burst}}=1\mathrm{ms}$. 200 bursts and 40,000 laser pulses, respectively, were applied per ablation spot. The total ablation time per spot was thus 10 s. This strategy was chosen in order to extend the ablation process on time scale since the minimum measurement time of the used particle counter for producing reliable results is 2 s. The raw laser beam was focused by a convex lens with a focal length of $f=160\mathrm{mm}$, resulting in a beam waist diameter of $2w_0=27\mathrm{\mu m}$ on the sample surface. In order to study the dependency of particle formation on the laser energy, five different pulse energies ($E_{\mathrm{pulse}}=10$, 31, 62, 94, and 125 μJ) were applied. Taking the beam waist diameter and the number of laser pulses into account, the laser dose per ablation spot (i.e., the product of laser fluence and number of applied laser pulses) was $D=0.7$, 2.17, 4.33, 6.57, and $8.73\mathrm{kJ}/{\mathrm{mm}}^2$.

 

Fig. 1. Experimental setup for $D$-DBD plasma-assisted picosecond laser ablation of titanium and collection of generated particles.

Download Full Size | PDF

As shown in Fig. 1, the focused laser beam was guided coaxially to a pulsed $D$-DBD plasma. Such guidance was made possible by the use of a hollow-core, high-voltage electrode. The sample was placed on a dielectrically separated ground electrode where the discharge gap between the sample surface and the high-voltage electrode was 14 mm. Dielectric separation was realized by a glass slide with a thickness of 1 mm, leading to the formation of a single and stable plasma filament. This filament features a diameter of approx. 180–200 μm, whereas the plasma footprint diameter on the sample surface is approx. 2.5 mm, as ascertained by high-speed camera measurements and reported in previous work [31]. The process gas applied for plasma ignition was argon 5.0 at a flow rate of 1.54 standard liters per minute (slm). The plasma was driven at a high voltage pulse repetition rate of 7.2 kHz and a high-voltage pulse duration of 400 ns. Consequently, the mean plasma power was 2.7 W and the plasma energy per discharge was 375 μJ. During experimentation, the plasma and laser source were neither triggered nor synchronized.

Experiments were performed on titanium plates (quality grade 2) with a thickness of 1 mm and an arithmetic mean surface roughness of $R_a=1.9\mathrm{\mu m}$. Each data point presented hereafter was averaged from five single measurements at different positions on the sample surface. For the measurement of laser-generated particles, a particle counter (Dekati, model $\mathrm{ELPI}+$), i.e., a low-pressure impactometer, was used. This particle counter was run at low-resolution mode in order to get a fast sampling rate of 10 Hz. Here, the measurement time per value is 100 ms. During this measurement time, 20 data points are generated, where the first five data points are not considered in order to improve the signal-to-noise ratio. The actual value provided by the particle counter is thus given by the average of 15 data points. Working in this mode, the particle counter can classify the collected particles into 14 size classes by separating the particles depending on their aerodynamic diameter. For instance, the minimum detectable particle diameter is 10 nm and the maximum one is 7.26 μm. The measurements of the number of the particles were carried out at a distance of 10, 12, 15, 17, and 20 mm between the ablation spot and the counter nozzle of the particle counter. During measurements, the air dilution was supposed to amount to 1 and the particle density was supposed to be $1\mathrm{g}/{\mathrm{cm}}^3$. Since the used particle counter provides the particle number per ${\mathrm{cm}}^3$, i.e., the particle density $D$ but not the actual number of particles $N$, which are gained during one measurement period, the latter value was calculated based on the air flow rate $Q$ applied for aspiration of particles (taken from the manufacturer’s data sheet) according to $N=Q·D$.

All current particle numbers that were determined during the formation of one ablation spot were integrated to get the total number of particles during the process. Consequently, a time-resolved measurement during the ablation process was not possible. The total volume of the particles within one size range was finally calculated on the basis of the measured mean diameter, assuming a spherical basic shape for each particle, and the total number of particles. During each measurement, the background noise, caused by particles in the ambient air, was subtracted. One has to consider that, due to the different time regimes of the laser (ps) and plasma (ns) pulses and the comparatively long acquisition time (ms) of the particle counter, such measurements do not describe the actual ablation process in detail. However, the particle counter gives integrated values for the two different ablation modes, i.e., with and without plasma. This allows a relative comparison of these modes and the determination of the impact of the assisting plasma as presented in Section 3.

In addition to the measurement of the size, volume, and number of particles generated during pure laser ablation and plasma-assisted ablation, the actually ablated volume on the samples was determined via laser scanning microscopy (LSM) (Zeiss, model LSM 700). For this purpose, the cross-section of each ablation spot and its depth were determined. These values were taken for three equally shifted scan lines shifted by 60°, and the volume was then defined to be the mean value of these three scans, assuming a rotational-symmetric truncated cone. The ablated volume was determined in order to allow a comparison of the measured particle volume and the actually ablated volume on the one hand and to get additional information on the impact of assisting plasmas on the picosecond laser ablation process on the other hand.

3. RESULTS AND DISCUSSION

A. Ablated Volume and Particle Volume

As shown in Fig. 2 and listed in Table 1, the assisting plasma has a significant impact on both the ablated volume and the volume of the generated particles. For better visualization, the data shown in Fig. 2 were normalized, i.e., each data point was divided by the particular maximum value of a series of experiment. It can be stated that the ablated volume increases continuously and reaches a maximum saturation value of $13.24\times {10}^{-5}{\mathrm{mm}}^3$ at a laser dose of $2.17\mathrm{kJ}/{\mathrm{mm}}^2$ (see Table 1) during pure laser ablation. In contrast, the ablated volume increases up to $53.70\times {10}^{-5}{\mathrm{mm}}^3$ at an applied laser dose of $4.33\mathrm{kJ}/{\mathrm{mm}}^2$ but then decreases in the case of plasma-assisted ablation [see Fig. 2(a)]. Such decrease was not observed for the volume of generated particles as shown in Fig. 2(b). Here, the dependencies of the particle volume on the laser dose show the same qualitative relationship for both cases, pure and plasma-assisted ablation.

Table 1. Ablated Volume ${\mathsf{V}}_{\mathsf{a}}$ And Particle Volume ${\mathsf{V}}_{\mathsf{p}}$ for Pure Laser Ablation ($\mathsf{L}$) and Plasma-Assisted Laser Ablation (LP) for the Five Different Applied Laser Doses $\mathsf{D}$^a

View Table

 

Fig. 2. Normalized ablated (a) and particle (b) volume versus applied laser dose for pure laser and plasma-assisted ablation at a fixed distance between the intake port of the particle counter and the ablation spot of 10 mm.

Download Full Size | PDF

The decrease of ablated volume for fluences higher than $4.33\mathrm{kJ}/{\mathrm{mm}}^2$ during plasma-assisted ablation could be explained by different mechanisms.

First, the incoming laser beam could be reflected or absorbed within the laser-induced plasma due to plasma shielding. It can be assumed that the laser-induced plasma features a higher electron density in this case, since the assisting atmospheric pressure plasma contributes to its formation by local-transient modifications of the titanium surface and pre-ionization of the surrounding area close to the ablation spot. Moreover, a higher amount of particles is vaporized or ionized and charged, respectively, when increasing the laser dose. This effect consequently supports plasma shielding.

Second, the laser beam could be attenuated by absorption, reflection, or scattering at particles within the ablation plume. This explanation is supported by the fact that, during plasma-assisted ablation, a higher total amount of particles was formed, as discussed in more detail in Subsection 3.B. In any case, the use of assisting plasma allowed a considerable increase in removed material as listed in Table 1.

The increase in ablated volume becomes obvious by looking at the comparison of LSM pictures in Fig. 3. This comparison reveals that material not only was removed at the focal point after introducing the assisting plasma to the laser ablation process, but also that the area around the actually ablated hole was affected by the incident laser beam, resulting in roughening of a concentric ring-shaped border area [see Fig. 3(b)].

 

Fig. 3. 3D view (top) and cross section (bottom) of holes generated by pure laser ablation (a) and plasma-assisted laser ablation (b) at the maximum applied laser dose of $8.73\mathrm{kJ}/{\mathrm{mm}}^2$.

Download Full Size | PDF

On the one hand, such roughening can be explained by material removal by the impact of single plasma filaments. As shown by the appropriate cross-section in Fig. 3(b), roughness peaks and valleys outside of the actually ablated hole are found in between the lateral positions $-65\mathrm{\mu m}$ and $+70\mathrm{\mu m}$, resulting in a width of the ring-shaped border area of approx. 135 μm. This value is in the order of magnitude of the diameter of the applied plasma beam of approx. 180–200 μm. On the other hand, the plasma has a certain impact on the propagation characteristics of the coaxially guided laser beam. Even though the used plasma could be referred to as “cold” plasma, it features a certain gas temperature (i.e., approx. 14 K above room temperature, as determined in previous work [32]), and can thus act as a thermally induced gradient index lens with certain position instability. Consequently, a plasma-induced wobbling of the laser beam can occur. The radius of the resulting wobble circle is approx. 20 μm for the given operating parameters of the plasma source [32]. Such wobbling could be another reason for the formation of the observed ring-shaped border area of the actual hole and is also responsible for the broadening of the hole diameter and the accompanying increase in ablated volume (compare cross sections in Fig. 3).

It turns out that as a result of introducing assisting plasma to the laser ablation process, the ablated volume increases by a factor of 1.85–3.78 (see Table 1). Furthermore, the particle volume is increased by a factor of 1.36–2.18 (see Table 1). Consequently, the increase in particle volume is generally lower than the increase in actually ablated volume (see right column in Table 1), so the volume and number of particles do not correspond to the volume and amount of removed material. This behavior indicates a reduction or breaking of particles around the ablation spot by the assisting plasma. This assumption is also supported by the observation of the particle size distribution in Fig. 4. Here, the impact of the assisting plasma on the dose-dependent number of particles is displayed for three different particle class sizes, hereafter referred to as nano range (i.e., particle diameters from 10 nm to 754 nm), micro range (i.e., particle diameters from 1.24 μm to 7.26 μm), and full range (i.e., particle diameters from 10 nm to 7.26 μm).

 

Fig. 4. Plasma-induced increase of the number of particles expressed by the ratio $\mathrm{LP}/L$ (detected number of particles for plasma-assisted ablation divided by the detected number of particles as measured for pure laser ablation) versus applied laser dose at a fixed distance between the intake port of the particle counter and the ablation spot of 10 mm for three class sizes (nano, micro, and full range; for definition of particle sizes, see text).

Download Full Size | PDF

It turns out that the observed plasma-induced growth of number of particles in the full range is dominated by the growth of nano-scaled particles, whereas micro-scaled ones do not significantly contribute to this growth. As an example, the absolute number of particles in the micro range, expressed by the ratio of 2.12 shown in Fig. 4, is notably lower than in the nano range ($\text{ratio}=2.91$) at an applied laser dose of $0.7\mathrm{kJ}/{\mathrm{mm}}^2$. However, it should be considered that at this dose, the lowest ablated volume and absolute number of particles, respectively, were generated.

The plasma-induced increase in number of particles remains almost constant when varying the distance between the intake port of the particle counter and the ablation spot, as shown in Fig. 5.

 

Fig. 5. Number of particles versus distance between the intake port of the particle counter and the ablation spot for pure laser and plasma-assisted ablation.

Download Full Size | PDF

Within the considered distance range (10–20 mm), the total number of particles was on average 1.4-times higher in the case of plasma-assisted ablation with respect to pure laser ablation. Merely at the highest distance of 20 mm, a higher factor can be observed, which could be attributed to differences in particle mobility due to different particle shapes.

B. Particle Size Distribution

The distribution of the particle size over the entire detectable size range at the minimum $0.7\mathrm{kJ}/{\mathrm{mm}}^2$ and maximum applied laser fluence ($8.73\mathrm{kJ}/{\mathrm{mm}}^2$), respectively, is shown in Fig. 6.

 

Fig. 6. Number of particles versus particle diameter as detected for the minimum applied laser fluence of $0.7\mathrm{kJ}/{\mathrm{mm}}^2$ (a) and the maximum of $8.73\mathrm{kJ}/{\mathrm{mm}}^2$ and (b) at a fixed distance between the intake port of the particle counter and the ablation spot of 10 mm.

Download Full Size | PDF

Most of the detected particles have a diameter less than 1 μm. Comparing the distributions for low and high laser fluence, it turns out that the particle size distributions are very similar. The particle size obviously depends only to a small extent on the applied laser fluence. In the case of pure laser ablation, a local maximum for the number of particles is found for particles with a mean diameter of 72.5 nm. The number of particles then decreases continuously for higher diameters. A slight decrease is also observed for lower particle diameters.

It can further be stated that the amount of nano-scaled particles is notably increased with respect to pure laser ablation when introducing assisting plasma to the ablation process. As shown in Fig. 7, the highest plasma-induced increase by a factor of 6.63 is found for particles with a mean diameter of 10 nm at the lowest applied laser dose of $0.7\mathrm{kJ}/{\mathrm{mm}}^2$. For higher doses and bigger particles, significantly lower increases by a mean factor of 1.26 were observed.

 

Fig. 7. Plasma-induced increase in number of particles expressed by the ratio $\mathrm{LP}/L$ (detected number of particles for plasma-assisted ablation divided by the detected number of particles as measured for pure laser ablation) versus applied laser dose at a fixed distance between the intake port of the particle counter and the ablation spot of 10 mm for three essential particle diameters.

Download Full Size | PDF

This effect could be explained by a mitigation of agglomeration of small particles to bigger clusters due to the plasma discharge. It was shown by several authors that external electric fields have a strong impact on the size of particles generated by flame synthesis [33] or laser ablation [25]. Even though particle agglomeration was observed by Serkov and co-workers [34], Sapkota et al. have shown that applying external electric fields allows a reduction in particle size [25] and most likely a reduction in particle agglomeration. In the latter work it was reported that introducing such an electric field to tin particle production by laser ablation in water allows a reduction of particle diameter by a factor of 2 for small particles in the single-digit nanometer range (from an average of 8 nm without to 4 nm with an applied external field of 9 V/cm). In the double-digit nanometer range, this approach was even more efficient in terms of particle size reduction (from approx. 40 nm without to about 15 nm with an applied external field of 9 V/cm) [25].

The nature of the particularly applied electric field turns out to have notable impact on the mechanisms during particle formation. The electric field strength of the pulsed plasma discharge used in the present work is in the range of some kilovolts per centimeter [35]. Moreover, the plasma is driven by alternating current at a relatively high pulse repetition rate and high voltage, whereas Serkov applied an electric field based on direct current. This fact could be the main reason for the observed decrease in particle size due to the different mobility of small, medium, and big particles, respectively.

Even though the highest increase in the number of particles as expressed by the ratio shown in Fig. 7 is found for the minimum applied laser dose, the absolute number of generated particles is maximal for high doses. For instance, the number of particles with a mean diameter of 10 nm is $38·{10}^6$ at a dose of $0.7\mathrm{kJ}/{\mathrm{mm}}^2$ and $417·{10}^6$ at a dose of $8.73\mathrm{kJ}/{\mathrm{mm}}^2$ for pure laser ablation. In the case of plasma-assisted ablation, these numbers are $252·{10}^6$ at a dose of $0.7\mathrm{kJ}/{\mathrm{mm}}^2$ and $1030·{10}^6$ at a dose of $8.73\mathrm{kJ}/{\mathrm{mm}}^2$, respectively. Anyhow, it can be stated that the use of assisting plasma results in an increase in generated micro- and nanoparticles.

Collection of the generated particles could finally be realized by charging and separation by an appropriate electro-magnetic setup. When referring to plasma-assisted laser-induced production of particles in air, it must be taken into account that the generated particles are most likely titanium oxide due to reaction with oxygen from ambient air. Such oxidation alters the particle volume and should thus be considered or even reduced or inhibited by appropriate modification of the experimental setup (e.g., by using a shielding gas-filled vacuum chamber). Collection of particles produced by plasma-assisted laser ablation could also be performed in liquids. Actually, plasmas in liquids are already applied to produce nanoparticles [36–38]. Plasma-assisted picosecond laser ablation in liquids could thus open a new field of research for the generation of micro- and nanoparticles or even particles with diameters in the sub-nanometer range, i.e., several hundreds of picometers.

4. CONCLUSIONS

The results presented in this work show that introducing assisting plasma to picosecond laser ablation of titanium leads to significant increase in both ablated volume and generated micro- and nanoparticles. Regarding the generation of particles, the highest ratio and lowest absolute number of particles, respectively, were found at the lowest applied dose of $0.7\mathrm{kJ}/{\mathrm{mm}}^2$. Here, the number of particles with a mean diameter of 10 nm was increased by a factor of 6.63 in the case of plasma-assisted ablation with respect to pure ablation. This effect is most likely attributed to the influence of the electric field provided by the used pulsed plasma, i.e., a direct dielectric discharge.

The results show that the presented plasma-assisted approach could be applied in order to increase the efficiency during laser-induced production of nanoparticles. The measured data also allow the assumption that very small particles with diameters below the detectable range were formed. Here, the number of particles in the nanometer-range is increased when introducing assisting plasma to the laser ablation process. Consequently, respirable and tissue-penetrating particles are formed during picosecond ablation of titanium where the number of particles is even notably increased by the assisting plasma. This issue should be born in mind in order to assure occupational health and safety.