Mechanism and elimination of bending effect in ... - OSA Publishing

5 downloads 7 Views 9MB Size Report
factor, a rough vacuum environment was applied; and the bending effect was also .... Schematic illustration of the experimental setup. Plano-convex lens: f ... 100. 200 300 400 500 600 700 800 900 1000 1100 1200 1300. Top-view. 20μm. (b).

Mechanism and elimination of bending effect in femtosecond laser deep-hole drilling Bo Xia,1 Lan Jiang,1,* Xiaowei Li,1 Xueliang Yan,1 and Yongfeng Lu2 1

Laser Micro/Nano Fabrication Laboratory, School of Mechanical Engineering, Beijing Institute of Technology, 100081, China 2 Department of Electrical Engineering, University of Nebraska-Lincoln, Lincoln, Nebraska 68588-0511, USA * [email protected]

Abstract: In this work, a comprehensive study of the bending effect, which remains one of the most critical challenges during deep-hole drilling, was conducted. The experimental statistics indicate that polarization is not the main factor in bending, but the deviation of the hole tends to be perpendicular to the polarization direction. Also, the dynamic ablated material/plasma was studied. Straight microholes were obtained by extending the interval between laser pulses to avoid dynamic ablated material existing in the millisecond time domain. Therefore, we speculated that the disturbance of the laser beam at the dynamic ablated aerosol, which have not sufficiently dispersed in the millisecond domain, is the main mechanism of bending. However, to more efficiently reduce the disturbance factor, a rough vacuum environment was applied; and the bending effect was also eliminated. The critical pressure for eliminating bending was about 2 × 104 Pa that is about one order of magnitude lower than the atmosphere. The fabricated high-quality microhole arrays without bending show that the proposed drilling method is convenient and efficient with high repeatability and controllability. ©2015 Optical Society of America OCIS codes: (140.7090) Ultrafast lasers; (140.3390) Laser materials processing; (150.5495) Process monitoring and control.

References and links 1.

E. H. Lundgren, A. C. Forsman, M. L. Hoppe, K. A. Moreno, and A. Nikroo, “Fabrication of pressurized 2 mm beryllium targets for ICF experiments,” Fusion Sci. Technol. 51(4), 576–580 (2007). 2. Z. Fei, X. Hu, H. W. Choi, S. Wang, D. Farson, and L. J. Lee, “Micronozzle array enhanced sandwich electroporation of embryonic stem cells,” Anal. Chem. 82(1), 353–358 (2010). 3. H. Varel, D. Ashkenasi, A. Rosenfeld, M. Wähmer, and E. E. B. Campbell, “Micromachining of quartz with ultrashort laser pulses,” Appl. Phys., A Mater. Sci. Process. 65(4–5), 367–373 (1997). 4. L. Jiang, L. Zhao, S. Wang, J. Yang, and H. Xiao, “Femtosecond laser fabricated all-optical fiber sensors with ultrahigh refractive index sensitivity: modeling and experiment,” Opt. Express 19(18), 17591–17598 (2011). 5. H. K. Tönshoff, C. Momma, A. Ostendorf, S. Nolte, and G. Kamlage, “Microdrilling of metals with ultrashort laser pulses,” J. Laser Appl. 12(1), 23–27 (2000). 6. S. Karimelahi, L. Abolghasemi, and P. R. Herman, “Rapid micromachining of high aspect ratio holes in fused silica glass by high repetition rate picosecond laser,” Appl. Phys., A Mater. Sci. Process. 114(1), 91–111 (2014). 7. S. Baudach, J. Bonse, J. Krüger, and W. Kautek, “Ultrashort pulse laser ablation of polycarbonate and polymethylmethacrylate,” Appl. Surf. Sci. 154-155, 555–560 (2000). 8. S. Döring, S. Richter, S. Nolte, and A. Tünnermann, “In situ imaging of hole shape evolution in ultrashort pulse laser drilling,” Opt. Express 18(19), 20395–20400 (2010). 9. T. V. Kononenko, V. I. Konov, S. V. Garnov, S. M. Klimentov, and F. Dausinger, “Dynamics of Deep Short Pulse Laser Drilling: Ablative Stages and Light Propagation,” Laser Phys. 11(3), 343–351 (2001). 10. S. Döring, J. Szilagyi, S. Richter, F. Zimmermann, M. Richardson, A. Tünnermann, and S. Nolte, “Evolution of hole shape and size during short and ultrashort pulse laser deep drilling,” Opt. Express 20(24), 27147–27154 (2012). 11. S. Tao, B. X. Wu, and S. L. Lei, “Study of laser beam propagation in microholes and the effect on femtosecond laser micromachining,” J. Appl. Phys. 109(12), 123506 (2011).

#246964 © 2015 OSA

Received 30 Jul 2015; revised 26 Sep 2015; accepted 29 Sep 2015; published 14 Oct 2015 19 Oct 2015 | Vol. 23, No. 21 | DOI:10.1364/OE.23.027853 | OPTICS EXPRESS 27853

12. L. S. Jiao, E. Y. K. Ng, H. Y. Zheng, and Y. L. Zhang, “Theoretical study of pre-formed hole geometries on femtosecond pulse energy distribution in laser drilling,” Opt. Express 23(4), 4927–4934 (2015). 13. L. Shah, J. Tawney, M. Richardson, and K. Richardson, “Self-focusing during femtosecond micromachining of silicate glasses,” IEEE J. Quantum Electron. 40(1), 57–68 (2004). 14. Y. E. Geints, A. M. Kabanov, A. A. Zemlyanov, E. E. Bykova, O. A. Bukin, and S. S. Golik, “Kerr-driven nonlinear refractive index of air at 800 and 400nm measured through femtosecond laser pulse filamentation,” Appl. Phys. Lett. 99(18), 181114 (2011). 15. L. Shah, J. Tawney, M. Richardson, and K. Richardson, “Femtosecond laser deep hole drilling of silicate glasses in air,” Appl. Surf. Sci. 183(3-4), 151–164 (2001). 16. S. Döring, T. Ullsperger, F. Heisler, S. Richter, A. Tünnermann, and S. Nolte, “Hole formation process in ultrashort pulse laser percussion drilling,” Phys. Procedia 41, 431–440 (2013). 17. S. Nolte, C. Momma, G. Kamlage, A. Ostendorf, C. Fallnich, F. von Alvensleben, and H. Welling, “Polarization effects in ultrashort-pulse laser drilling,” Appl. Phys., A Mater. Sci. Process. 68(5), 563–567 (1999). 18. J. Jiang, M. Chen, Z. X. Bai, C. Yang, and G. Li, “Influence of polarization on the hole formation with picosecond laser,” Opt. Rev. 20(6), 496–499 (2013). 19. D. Esser, S. Rezaei, J. Li, P. R. Herman, and J. Gottmann, “Time dynamics of burst-train filamentation assisted femtosecond laser machining in glasses,” Opt. Express 19(25), 25632–25642 (2011). 20. J. Koch, S. Heiroth, T. Lippert, and D. Günther, “Femtosecond laser ablation: Visualization of the aerosol formation process by light scattering and shadowgraphic imaging,” Spectrochim. Acta B At. Spectrosc. 65(11), 943–949 (2010). 21. B. Xia, L. Jiang, X. W. Li, X. L. Yan, W. W. Zhao, and Y. F. Lu, “High aspect ratio, high-quality microholes in PMMA: a comparison between femtosecond laser drilling in air and in vacuum,” Appl. Phys., A Mater. Sci. Process. 119(1), 61–68 (2015).

1. Introduction Microholes have broad applications in many fields [1,2]. With ultrahigh power intensities and ultrashort irradiation periods, the femtosecond (fs) laser presents unique advantages in the efficient drilling of microholes with wide material adaptation [3–5]. For practical applications, static multi-shot ablation is one of the most common drilling processes [6,7]. As the laser pulses increase, the microholes become deeper, until the increase finally stops [8]. However, the formation of microhole involves many complicated physical processes that have not been well understood, especially in deep-hole drilling. The bending at the tip of microhole is a critical challenge during deep hole drilling, which have been widely investigated. For example, T. V. Kononenko et al. [9] reported the dynamics of a deep microhole drilling process in diamond. A bent tip is apparent in a side-view image. Similarly, S. Döring et al. [10] researched hole formation with short and ultrashort pulse lasers in silicon. Bending was also found inside the tip of the hole. The quality of the hole is seriously affected by bending effect. It has a strong impact on the controllability of the drilling process, such as an uncontrolled exit for a through-hole, especially for a microhole array. The origin of the bending phenomenon must be some mechanism that causes a disturbance and anisotropic interaction. T. V. Kononenko et al. [9] attributed the bending to a linearly polarized laser which caused the anisotropic interaction. Using simulations, S. Tao et al. [11] investigated the distribution of laser intensity inside a microhole proving nonuniform distribution caused by linearly polarized laser. Then L.S. Jiao et al. [12] also proved the nonuniform distribution for a microcrater with a tapered sidewall. Besides the polarization, L. Shah et al. [13] attributed the bending to the filament instability which led to multiple ablated sites and bending. For bending in the bottom of deep holes, S. Döring et al. [10] attributed the imperfection to the laser scattering or deflecting at the ablation residue. The original factors for bending effect need further comprehensive investigation to clarify the fundamentals. In this work, a comprehensive investigation of bending, from the physical mechanism to elimination, was conducted. The anisotropic interaction with the linearly polarized laser, which may impact microholes bending, was investigated. The experimental statistics of the bending direction are presented to reveal the relationship between the bending directions and the polarization. In order to understand the fundamentals of the drilling process in our experiments, the ablated plasma in the nanosecond domain and the vapor cloud expansion in the millisecond domain were investigated using an intensified charge-coupled device (ICCD)

#246964 © 2015 OSA

Received 30 Jul 2015; revised 26 Sep 2015; accepted 29 Sep 2015; published 14 Oct 2015 19 Oct 2015 | Vol. 23, No. 21 | DOI:10.1364/OE.23.027853 | OPTICS EXPRESS 27854

camera. In this work, a method that can prevent the microholes from bending was proposed. Straight microholes were obtained in an air environment when the interval between pulses was longer than 20 ms. It indicates that the disturbance of the laser beam at the ablation matter, which was not sufficiently dispersed in the millisecond domain, was the main mechanism of bending. To more efficiently reduce the disturbance factor, a rough vacuum environment was utilized and the elimination of bending was verified below about 2 × 104 Pa. With this convenient and efficient bending elimination method, high-quality microhole arrays with high repeatability and controllability were obtained. 2. Experimental set-up Fs laser Plano-convex lens Z Ejection

Imaging for ejection/vapor-cloud

Y

ICCD

X Filter 2 Illuminator

Imaging for hole shape/plasma CCD /ICCD Target

Filter 1 Microhole

Fig. 1. Schematic illustration of the experimental setup. Plano-convex lens: f = 100 mm, Filter 1: 300 nm ~700 nm pass-through, Filter 2: 632.8 nm ± 10 nm pass-through.

Figure 1 shows the scheme of the setup used for the experiment. The percussion drilling source was a femtosecond laser system (Spectra-Physics) consisting of a mode-locked Ti:Sapphire oscillator and a regenerative amplifier, producing 50 fs full width at half maximum (FWHM) linearly polarized laser pulses at a central wavelength of 800 nm and a repetition rate which can be adjusted from 1 kHz to 4Hz. The energy of the laser pulses could be tuned with an attenuator assembly of a half-wave plate and a polarizer. The direction of polarization was originally parallel to the y-axis, as shown in Fig. 1. In order to study the impact of polarization on microhole bending, a half-wave plate and a quarter-wave plate were placed behind the polarizer to adjust the direction of polarization and change the polarization to a circularly polarized laser, respectively. The diameter of the laser beam is about 8 mm. The fs laser beam was statically focused onto the top surface of a polished PMMA bulk sample via a plano-convex lens (diameter 25.4 mm) with a focus length of 100 mm. The sample was placed onto a three-dimensional positioning stage (SmarAct - MCS). On the side view of the sample perpendicular to the direction of the laser propagation, an illuminator (white light), a microlens (50x long work distance Objective, Olympus), a filter (300 nm ~700 nm pass-through), and a CCD camera (DMK31BU03) were applied to obtain the images of the shape evolution of the microholes. To prevent the femtosecond laser from scattering into the CCD, a filter (300 nm ~700 nm pass-through) was placed before the CCD, as seen in Fig. 1. Furthermore, the plasma evolution was easily obtained by replacing the CCD with an ICCD camera (Andor istar) with the exposure time set at about 5 ns. In addition, a transverse illumination by a He-Ne laser at a central wavelength of 632.8 nm was placed above the target, as seen in Fig. 1. The imaging system was fixed on a two-dimensional stage which could be adjusted to the region of detection. The light of the He-Ne laser, which was

#246964 © 2015 OSA

Received 30 Jul 2015; revised 26 Sep 2015; accepted 29 Sep 2015; published 14 Oct 2015 19 Oct 2015 | Vol. 23, No. 21 | DOI:10.1364/OE.23.027853 | OPTICS EXPRESS 27855

scattered by ablation ejection, was captured by the ICCD camera. During light scattering, unspecific radiation emitted by the expanding plasma plume or laboratory lightning was rejected by an interference filter (λ = 632.8 nm ± 5 nm) positioned in front of the camera. The exposure time of the ICCD was set at 1 ms. By controlling the delay between the laser shot and the ICCD camera, images of ablated ejection and condensed particles for a wide range of temperatures and densities were obtained on time scales reaching up to the millisecond domain. 3. Results and discussion 3.1 The phenomenon and discussion of bending at the tip of microholes The evolution of the microhole with an increase in the number of pulses (N) from 100 to 3,000 is shown in Fig. 2(a). The energy of the laser pulse was about 30 μJ at a repetition rate of 1 kHz. The microholes became deeper with an increase in the laser pulses, while the diameter of the microholes barely changed. In the beginning (about 0 ~600 pulses), the hole was uniform and straight with a constant ablation rate. Then (about 600 ~2500 pulses), the ablation tip was no longer straight, and the ablation rate became lower. At this time, the formation of minor imperfections on the sidewall and bending at the tip was apparent. Finally (over 1500 pulses), the increase in depth was almost stopped. The quality of the hole became poor if the laser pulses continued to interact with the sample. The entrance of the microhole from the top view (xy-plane) is shown in Fig. 2(b) which was obtained by focusing the microscope on the sample surface. For more details on the bending phenomenon, the specifics of the bending direction could not be obtained only from the longitudinal section in the xz-plane. For the transparent material, the microscope can also be focused into the sample. So, the clear direction of the bending tip can be obtained from a top-view image by focusing the microscope on the tip of the microhole, as seen in Fig. 2(c). Side-view 100

200

300

400

(a)

500

600

700

800

900

Top-view

1000 1100 1200 1300 1400 1500 2000 2500 3000 (b)

The final hole

(b) The Entrance of Microhole

20μm

(c) (c)

Z X

100μm

Y

Bending Direction

θ X

Fig. 2. (a) Images of the longitudinal section of the evolution of a microhole during the drilling process. The number of pulses increased from 100 to 3,000 for a pulse energy of 30 μJ at a repetition rate of 1 kHz. (b) The entrance of the microhole on the sample surface. (c) The clear direction of the bending tip in the xy-plane.

It is worth noting that bending only occurs at the end of the laser drilling process, as shown in Fig. 2. This means that the phenomena may be attributed to factors which disturb the laser propagation and lead to anisotropic interaction at the bottom of the microholes. At the beginning of the drilling process, the laser energy is sufficient to ensure laser propagating without disturbance; and the microhole keeps straight. For a higher depth, especially at the end of the drilling process, the subtle disturbance factor will play a more important role in the laser propagation and anisotropic ablation because of the longer laser propagation distance and the weaker laser energy.

#246964 © 2015 OSA

Received 30 Jul 2015; revised 26 Sep 2015; accepted 29 Sep 2015; published 14 Oct 2015 19 Oct 2015 | Vol. 23, No. 21 | DOI:10.1364/OE.23.027853 | OPTICS EXPRESS 27856

To fundamentally understand the bending, the factors which disturb the laser propagation and lead to an anisotropic interaction at the bottom of the microholes should be researched. From the standpoint of laser propagation, there may be several primary ways to deliver the laser pulse energy to the bottom of the hole, such as propagation by reflecting in the formed hole [11], propagation in a straight path, propagation by a filament channel [14], and so on. In those propagation ways, several possible factors which lead to anisotropic interaction should be discussed. (1) The polarization direction of a laser beam may lead to the anisotropic interaction [9]. The energy distribution of the laser will be impacted by the polarization. The affected laser distribution easily brings up anisotropic ablation, which leads to microhole bending. (2) A narrow plasma filament was produced as a result of self-focusing in the ambient air. For the subsequent laser pulse, the ablated ejection and the vapor cloud may have led to an unstable filament bending slightly deep inside the hole. Then the laser beam energy was deposited along a bent direction [15]. (3) The laser beam is scattered or deflected at the ablation particles that still remain within the hole or are deposited on the sidewalls [16]. The anisotropic interaction at the bottom and the sidewalls may be one of the origins of bending. In the following work, to verify which disturbance factors cause the bending, the polarization was investigated; and the drilling process was also observed. 3.2 The investigation of laser polarization and the impact on microhole bending In general, the microhole created by previous laser pulses may act like a waveguide whose side wall reflects subsequent laser pulses to the bottom of the microhole. For a typical microhole shape, shown in Fig. 2, the taper of the sloped inner sidewall at the tip was about 10 ~20°. This means the incidence angle of the laser is about 70~80°. However, according to the Fresnel reflection coefficient on polarization orientation [17], the reflectivity has a large difference for s-polarized and p-polarized lasers. Hence, the polarization direction of the laser beam may be one of the origins of the anisotropic interaction. In order to investigate the anisotropic interaction, a circularly polarized laser and a linearly polarized laser were compared. Microhole bending occurred in the drilling with both the linearly and circularly polarized laser. The orientation of the bent microhole seems to be random. This means that the polarization is not the main factor leading to the bending of the microhole. To obtain more detail of the bending direction, as shown in Fig. 2(c), the angle between the orientation of the bent microhole and the y-axis is defined as θ, as shown in Fig. 2(c). Statistics of the θ were compiled; and three histograms were made for 100 microholes drilled with a linearly polarized laser parallel to the y-axis [Fig. 3(a)], perpendicular to the yaxis [Fig. 3(b)], and with a circularly polarized laser [Fig. 3(c)], respectively. Figure 3(d)-3(f) show the partial microholes that correspond to Fig. 3(a)-3(c) respectively. The statistical results in Fig. 3(a) indicate that the angle θ tends to be larger than 45°. The statistical results in Fig. 3(b) indicate that the angle θ tends to be smaller than 45°. In contrast, the statistical results of a circularly polarized laser, in Fig. 3(c), has no apparent tendency. That means the orientation of the bending tip tends to be perpendicular to the polarization rather than parallel to the polarization. Though the polarization is not a major factor leading to the bending, it has an impact on the bent direction. As for the impact, the polarization can only provide a preference to enhance the bending. For the phenomenon of ablation enhanced at the direction perpendicular to the polarization, some previous experimental studies have been reported. In [17], S. Nolte et al. attributed non-circular hole shapes in stainless steel foils to polarization-dependent reflections occurred at the walls of the holes, leading to non-uniform intensity distributions. In [9], T. V. Kononenko et al. also used the reflectance anisotropy to explain the bending in CVD diamond. In [18], J. Jiang et al. explained the elliptical exit shapes as a result of the difference in reflectivities between s- and p- polarized incident lasers too. And it is further conrmed by comparative experiments and calculations conducted on copper. In [12], L. S.

#246964 © 2015 OSA

Received 30 Jul 2015; revised 26 Sep 2015; accepted 29 Sep 2015; published 14 Oct 2015 19 Oct 2015 | Vol. 23, No. 21 | DOI:10.1364/OE.23.027853 | OPTICS EXPRESS 27857

Jiao et al. solved Maxwell’s wave equation for fs laser drilling of silicon. The results indicate that the laser intensity at the bottom of the micro-hole has an enhancement perpendicular to the direction of polarization at taper angles of about 15 degrees. According to the previous investigations mentioned above, the polarization dependence of laser reflectivity at the tapered hole wall may be one of reasons for the tendency of bending direction. The intensity anisotropy at the bottom maybe only an enhancement factor of the bending rather than an origin of the hole bending. If some deflected ablation points are formed randomly, the points in the region perpendicular to the polarization can be ablated further than in the other direction. Consequently, the bent tip tended to be perpendicular to the polarization statistically. In contrast, the circularly polarized laser had no anisotropic enhancement; and the orientations of the bent microholes seem random statistically. However, the fundamentals of bending dependence on polarization still remains a challenge. (d)

Frequency

(a)

100μm Y X

Frequency

(b)

(e)

100μm Y X

(c)

Frequency

(f)

100μm Y X

The interval of θ for statistical Fig. 3. Histogram of the angle θ which is between orientation of the bending and y-axis for 100 microholes drilled with a linearly polarized laser parallel to the y-axis (a), perpendicular to the y-axis (b), and a circularly polarized laser (c) respectively at 2,000 pulses. The (d), (e), and (f) are part of the microholes corresponding to (a), (b), and (c), respectively. The energy is about 30 μJ/pulse.

#246964 © 2015 OSA

Received 30 Jul 2015; revised 26 Sep 2015; accepted 29 Sep 2015; published 14 Oct 2015 19 Oct 2015 | Vol. 23, No. 21 | DOI:10.1364/OE.23.027853 | OPTICS EXPRESS 27858

3.3 The verification of plasma and vapor cloud expansion which impact on microholes bending Except for the polarization, there must be some disturbance factors which initially led to random deflected ablation points. Hence, what happened in drilling should be made clear. In this section, the plasma and vapor cloud expansion was observed. For the timescale in the nanosecond domain, the plasma expansion for the experiment is shown in Fig. 4 at a laser pulse energy of 30 μJ. The methods of recording plasma expansion were followed by the work in [19]. The experimental setup is shown in Fig. 1. The image was recorded at the first pulse and the 200th pulse by controlling the delay between the laser shot and the ICCD camera with 5 ns exposure time. In Fig. 4, the plasma can be seen to last to several tens of nanoseconds until the illumination was not enough. For the deeper holes, as shown in Fig. 4(b), the plasma will cool down gradually as it is ejected out of the microhole. Although the sidewall formed is likely to be washed clearly by the followed plasma, the tip of the hole may still contain some ablated particles. It may be one of the disturbance factors. 0~5 ns

5~10 ns

10~15 ns

15~20 ns

20~25 ns

25~30 ns

30~35 ns

Air PMMA

100 μm

1st pulse

(a) 0~5 ns

5~10 ns

10~15 ns

15~20 ns

20~25 ns

25~30 ns

30~35 ns

Air PMMA

200st pulse

100 μm

(b) Fig. 4. The plasma expansion for the first pulse and 200th pulse at a laser pulse energy of 30 μJ.

Besides the plasma, for the longer timescale in the millisecond domain, the vapor cloud expansion and the material ejection in air may not have dispersed sufficiently. The instability phases, such as vapor and ejection, will probably disturb the propagation of subsequent laser pulses, especially for the filament stability which is crucial for femtosecond laser preparation. In order to confirm whether vapor clouds and ablated ejection materials can last up to the millisecond domain in the experiment, it should be directly visualized. However, the backilluminated shadowgraphs did not have enough contrast because the vapor and particles are #246964 © 2015 OSA

Received 30 Jul 2015; revised 26 Sep 2015; accepted 29 Sep 2015; published 14 Oct 2015 19 Oct 2015 | Vol. 23, No. 21 | DOI:10.1364/OE.23.027853 | OPTICS EXPRESS 27859

too thin. So a light scattering image was applied as it enables the monitoring of even strongly dispersed aerosol [20]. The setup is shown in Fig. 1. The scattered light was recorded by ICCD camera with 2 ms exposure time, as shown in Fig. 5. The image was captured at different time delays in the millisecond domain after the first pulse was reached. In the first 2 ms domain, the aerosol, including the vapor cloud and ablated ejection material, was still keeping in a certain degree of condensation, as shown in Fig. 5(a). Then, the aerosol gradually dispersed within the next tens of milliseconds, as shown in Fig. 5(b)-5(j). These results proved that there is something still remaining above the ablation region which lasts up to the milliseconds domain under the experimental conditions. This residue is probably the aerosol, including the vapor cloud and ablated ejection material. With the increase of pulse number, the scattered light can also be captured but becomes weaker and weaker under constant laser illumination intensity. When the pulse number is up to 50, the captured scattered light almost disappears. However, by enhancing the illumination intensity, the captured images (not included in this paper) are also similar to Fig. 5. It indicates that the ablated material above the sample becomes less with the pulse number increasing. However, the ablation rate maintains a constant of about 1 μm/s for hundreds of pulses, as shown in Fig. 2(a). It is suspected that part of the vapor cloud and ablated ejection material remain in the hole. (a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

(i)

(j)

500 μm

Air PMMA

0~2 ms

5~7 ms

10~12 ms

15~17 ms

20~22 ms

25~27 ms

30~32 ms

35~37 ms

40~42 ms

45~47 ms

Fig. 5. The directly visualized vapor cloud and ablated material in the millisecond domain by the light scattering image method. The energy is about 30 μJ/pulse.

Thus, it is the aerosol that last to several tens of milliseconds that may also disturb the subsequent laser pulse propagation and influence the laser energy deposition. If the temporal interval between laser pulses is extensive enough, these instability phases will not influence the subsequent laser pulse. The disturbance factor which causes bending may be eliminated, and extending the interval between laser pulses is easy to realize. Therefore, the impact of the vapor cloud expansion on bending microholes can be verified. As shown in Fig. 6, the microholes were drilled with different repetition rates from 1,000 Hz to 10 Hz. In other words, the temporal interval between laser pulses extended from 1 ms to 100 ms. The applied pulse number for each microhole is 3,000 pulses which is sufficient to end the whole drilling process. For a repetition rate at 1,000 Hz, the bending tip and the formation of minor imperfections on the sidewall are apparent. In contrast, the microholes drilled at 10 Hz are very straight. The critical point for the change was about 100 Hz to 50 Hz, as shown in Fig. 6. Therefore, the factors which led to the bending effect lasted up to about 10 to 20 ms in the experiment. This corresponds to the timescale of the vapor cloud expansion and the material ejection, as shown in Fig. 5(c) and 5(e). So it is suspected that the vapor cloud and the ejection material remaining in the air are the main disturbance factors which led to the bending. The elimination of the bending effect by extending the pulse intervals to tens of milliseconds proves that the dynamic aerosol, including the vapor cloud and ablated ejection material in the millisecond domain, is probably the main disturbance factor which led to the bending. For the repetition rate at 1,000 Hz, the dynamic aerosol remaining in deep channel #246964 © 2015 OSA

Received 30 Jul 2015; revised 26 Sep 2015; accepted 29 Sep 2015; published 14 Oct 2015 19 Oct 2015 | Vol. 23, No. 21 | DOI:10.1364/OE.23.027853 | OPTICS EXPRESS 27860

did not disperse completely when the subsequent laser pulse was reached. On one side, the dynamic aerosol scattered or deflected the incidence laser. Then the laser beam energy was deposited on the sidewall of the channel; and the deflected points will be a new guide for the drilling orientation. On the other hand, the filament in the deep hole tends to be unstable because of the unstable dynamic aerosol. It seriously affects laser beam propagation. The incident laser beam will be deposited along a bent direction. Then the direction of the microhole will deviate from the original straight path. In contrast, for a repetition rate at 10 Hz, the dynamic aerosol were dispersed too thin when the subsequent laser pulse was reached. The unstable factor disappeared. Hence, the microhole was straight. 1000Hz

500Hz

200Hz

100Hz

50Hz

20Hz

10Hz

Fig. 6. The microholes drilled with 3,000 pulses for a pulse energy of 30 μJ with a different repetition rate from 10 Hz to 1,000 Hz. The scale bar is 200 μm.

3.4 The efficient way to eliminate the phenomenon of bending 0~5 ns

5~10 ns

10~15 ns

15~20 ns

20~25 ns

25~30 ns

30~35 ns

Vacuum PMMA

100 μm

1st pulse

(a) 0~5 ns

5~10 ns

10~15 ns

15~20 ns

20~25 ns

25~30 ns

30~35 ns

Vacuum PMMA

200st pulse

100 μm

(b) Fig. 7. The plasma expansion in vacuum environment. The pressure is about 10 Pa.

#246964 © 2015 OSA

Received 30 Jul 2015; revised 26 Sep 2015; accepted 29 Sep 2015; published 14 Oct 2015 19 Oct 2015 | Vol. 23, No. 21 | DOI:10.1364/OE.23.027853 | OPTICS EXPRESS 27861

Avoiding the disturbance factor by extending the pulse interval is a way to eliminate the bending effect. However, for practical applications, reducing the repetition rate seriously affects drilling efficiency. Besides extending the pulse interval, improving the expansion efficiency of the ablated material/plasma by reducing the ambient pressure is a similar method. For a comparison, plasma and vapor cloud expansion in a rough vacuum environment are shown in Fig. 7 and Fig. 8. The condition is the same with Fig. 4 and Fig. 5 expect for the environment. The illumination intensity of the plasma in a vacuum (Fig. 7) is much weaker than the plasma in air (Fig. 4). That means the plasma in a vacuum cools down faster than in air. For a longer time scale, the vapor cloud and ablated ejection material in a rough vacuum disperses too much to hardly capture it in the millisecond domain, as seen in Fig. 8. This work also confirmed that the disturbance factor can be avoided by reducing the ambient pressure. (a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

(i)

(j)

500 μm

10000 Pa PMMA

0~2 ms

5~7 ms

10~12 ms

15~17 ms

20~22 ms

25~27 ms

30~32 ms

35~37 ms

40~42 ms

45~47 ms

Fig. 8. The vapor cloud expansion in a 1 × 104 Pa environment.

By utilizing a vacuum system as seen in our previous work [21], the phenomenon of bending hardly appears in a rough vacuum environment. As shown in Fig. 9, a series of microholes drilled in different ambient pressure for a pulse energy of 50 μJ at a repetition rate of 1 kHz is presented. The number of pulses was set at 5,000 which is sufficient to obtain the final microhole shape. The microholes became straight and deeper as the ambient pressure was lowered. Hence, reducing the ambient pressure is an efficient way to eliminate the bending effect. 30 Pa

1x102 Pa 1x103 Pa 1x104 Pa 2x104 Pa 4x104 Pa 6x104 Pa 8x104 Pa 1x105 Pa

200μm

Fig. 9. The microholes drilled with 5,000 pulses for a pulse energy of 50 μJ at different ambient pressures from 1 x 105 Pa to 30 Pa.

#246964 © 2015 OSA

Received 30 Jul 2015; revised 26 Sep 2015; accepted 29 Sep 2015; published 14 Oct 2015 19 Oct 2015 | Vol. 23, No. 21 | DOI:10.1364/OE.23.027853 | OPTICS EXPRESS 27862

This elimination of the bending phenomenon can probably be attributed to the disappearance of the dynamic aerosol which disturb the laser propagation at the end of the drilling process. It is worth noting that the critical point for eliminating the microhole bending effect is about 2 × 104 Pa, about an order of magnitude lower than the atmosphere. In the experiment, the critical point for bending at a pulse energy of 30 μJ and 70 μJ is almost the same with 50 μJ, which is about 2 × 104 Pa. It is probably attributed to the reduction in pressure where only one order of magnitude is enough to improve the expansion efficiency. As seen in Fig. 8, the vapor cloud has dispersed too much at the ambient pressure, about 1 × 104 Pa. The results in Fig. 8 are similar to those in Fig. 5(e)-5(j). Actually, it was found during the experiment that the vapor cloud almost cannot be captured at a pressure below 1,000 Pa. So the pressure in Fig. 8 was set to about 1 × 104 Pa instead of 10 Pa. However, as shown in Fig. 9, the critical point for bending is different from the critical point for the change of depth, which is about 100 Pa for a pulse energy of 50 μJ. The maximum depth increased gradually as the ambient pressure decreased, even below the critical point for bending, and remained approximately constant at lower pressure. This result indicates that the mechanisms of depth change are not the same with bending. The maximum depth increase in vacuum is not only attributed to the disappearance of a dynamic unstable factor but is also attributed to the weakened air ionization and the reduction of the static ablation debris, which is closely related to the laser beam propagation efficiency. To obtain deeper microholes, both the disturbance and energy loss of the laser beam should be considered. Hence, to increase the depth, it is necessary to decrease the pressure lower than the critical point for bending. In order to demonstrate the significance of eliminating the bending, the exit of microholes arrayed with bending microholes (a) and straight microholes (b) are shown in Fig. 10. Microholes were drilled with a circularly polarized laser with a pulse energy of 50 μJ in air [Fig. 10(a)] and rough vacuum [Fig. 10(b)], respectively, at a repetition rate of 1 kHz. The thickness of a sample was 1 mm. The interval between microholes was about 100 μm. It is obvious that the bending effect had a strong impact on the controllability, repeatability and quality of microholes. A convenient and efficient way to eliminate the bending effect as shown in this work is necessary.

(a)

Bent hole

(b)

Straight hole

Fig. 10. The contrast of the exit of microholes arrayed with bent microholes (a) and straight microholes (b). The microholes drilled with a circularly polarized laser with a pulse energy of 50 μJ in air (a) and vacuum (b), respectively, at a repetition rate of 1 kHz. The thickness of a sample was 1 mm. The interval between microholes was about 100 μm.

4. Conclusions In this work, a comprehensive study of bending phenomenon during femtosecond laser deephole drilling in PMMA was conducted. First of all, the phenomenon of bending at the tip of

#246964 © 2015 OSA

Received 30 Jul 2015; revised 26 Sep 2015; accepted 29 Sep 2015; published 14 Oct 2015 19 Oct 2015 | Vol. 23, No. 21 | DOI:10.1364/OE.23.027853 | OPTICS EXPRESS 27863

microholes was presented. To provide a fundamental understanding of bending, the factors which may disturb the laser propagation and lead to anisotropic interaction were discussed, such as the polarization, the ablated material/plasma, and so on. Secondly, a circularly polarized laser was applied to reduce the anisotropic interaction. As a result, the bending of microholes occurred with both the linearly and circularly polarized laser, indicating that polarization is not a major factor which leads to bending. However, it was found that the deviation of the hole tends to be perpendicular to the polarization. This may be attributed to the polarization-dependent energy distribution that enhances the initially formed defects in the region perpendicular to polarization. And then, to understand the impact of ablated material/plasma on microhole bending, the ablated plasma and the vapor cloud expansion was studied. The images show that the aerosol, including the vapor cloud and ablated ejection, can last into the milliseconds domain. Based on the study, straight microholes can be obtained in an air environment when the interval between pulses is as long as 20 ms for a pulse energy of 30 μJ. Therefore, it is speculated that the disturbance of the laser beam at the dynamic ablated aerosol which have not sufficiently dispersed in the millisecond domain is the main mechanism of bending in the experiment. Lastly, to more efficiently reduce the disturbance factor, a rough vacuum environment was utilized. It was verified that the microholes became straight and deeper as the ambient pressure was lowered. The critical point for eliminating the microhole bending effect was about 2 × 104 Pa, which is just about an order of magnitude lower than the atmosphere. By contrasting the microhole arrayed with bending microholes and straight microholes, it is demonstrated that the proposed drilling method is convenient and efficient with high repeatability and controllability. Acknowledgments This research is supported by the National Natural Science Foundation of China (NSFC) (grants 91323301 and 51305030), the National Basic Research Program of China (973 Program) (grant 2011CB013000) and the Cultivation Fund of the Key Scientific and Technical Innovation Project, Ministry of Education of China (No. 708018).

#246964 © 2015 OSA

Received 30 Jul 2015; revised 26 Sep 2015; accepted 29 Sep 2015; published 14 Oct 2015 19 Oct 2015 | Vol. 23, No. 21 | DOI:10.1364/OE.23.027853 | OPTICS EXPRESS 27864

Suggest Documents