Open Access
Add to BibSonomy
Abstract
Laser-induced periodic surface structure (LIPSS) is an important, high-throughput surface nano-structuring method, which has been used to fabricate various functional surfaces. In this paper, we fabricate double time-delayed orthogonally polarized femtosecond laser beams with a fixed beam power ratio of 1.5:1 that are employed to irradiate the silicon surface and curved periodic ripples with a sub-wavelength period. It is found that the local orientation of the ripples on the silicon surface can be modulated in a range of 0-80° by adjusting the fabrication parameters, such as the laser fluence, the target scanning speed, and the time delay between double laser beams. The transition from the curved ripples to the straight ripples can be achieved by increasing the target scanning speed. Different from previous studies that the curved periodic ripples are fabricated by modulating the laser polarization, the method demonstrated here utilizes the interaction between the linearly polarized subsequent laser beam and the preceding laser beam excited silicon to form curved ripples.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Since the first observation of the laser-induced periodic surface structures (LIPSS) in 1965 [1], LIPSS has received extensive attentions due to its capability of parallel nanopatterning of nearly all kinds of solid materials, including metals, semiconductors and dielectrics. Specifically, 1D and 2D periodic nanostructures have been fabricated by single laser beam or time-delayed double laser beams [2–4], or even bursts of femtosecond pulses with picosecond delay [5] and these nanostructures can be modulated by changing the ambient material used in the fabrication process [6]. After texturing by LIPSS, surface with nanostructures can exhibit a variety of useful functions, such as controllable surface wettability [7], enhanced optical absorption [8] and bactericidity [9], etc.
It is commonly accepted that the low spatial frequency LIPSS (LSFL) on the metal surface are formed due to the interference between the incident laser and the surface plasmon polaritons (SPPs) [10], which is also valid for LSFL on semiconductor materials, such as silicon [11]. When the silicon surface is irradiated by the linear polarized femtosecond laser, J. Bonse et al. find that the orientation of LSFL is perpendicular to the laser’s polarization and its period is slightly smaller than the laser wavelength [11]. They also observe that increasing the number of pulses striking on a single spot can lead to the reduction of the ripples’ period [12]. In addition, when double femtosecond laser beams with zero time delay are employed to irradiate one single spot on silicon, the LSFL orientation is perpendicular to the sum vector of the two beams’ polarizations [13]. The above characteristics of the LSFL morphologies on silicon can be all explained by the SPPs model.
However, recent experiments show that the orientation of LSFL on the silicon surface exhibits complex dependences on the laser parameters and scanning configurations. F. Fraggelakis et al. find that when double cross-polarized femtosecond laser beams with equal energy fluence (total energy fluence is 1.48 J/cm2) irradiate one single spot on silicon, the predominant ripples are perpendicular to the polarization of the secondly arrived laser beam; but when the double femtosecond laser beams with 2:1 energy fluence ratio (the total energy fluence is 1.40 J/cm2) are employed, the predominant ripples are perpendicular to the polarization of the laser beam with higher fluence [14]. Besides, Liu et al. observe that the orientation of LSFL on silicon formed by single or double femtosecond laser beam(s) is not only determined by the laser polarization, but also by the laser scanning direction and speed [15–17]. For example, when double orthogonally polarized and equal-power femtosecond laser beams with a time delay of 1 ps are used, the LSFL’s orientation is always perpendicular to the scanning direction [17]. The complex relations between the orientation of the LSFL on silicon and the fabrication parameters cannot be explained well by the SPPs model. Therefore, the control method of the orientation of LSFL on silicon needs to be further investigated.
According to the previous studies [14–17], it is known that both the power ratio between double laser beams and the target scanning configuration play important roles in the ripple formation. In this paper, double time-delayed orthogonally polarized femtosecond laser beams with 1.5:1 power ratio of the preceding to the subsequent laser beam are employed to fabricate structures on the moving silicon target. It is found that using the fabrication configurations different from those used in the literatures [11–17], totally different surface structures, i.e. curved periodic ripples are fabricated on the silicon surface. It is considered that the transient excited surface induced by the preceding laser beam is critical in forming the curved periodic ripples [18,19]. This paper demonstrates a new fabrication method of curved periodic ripples, not relying on the modulation of the femtosecond laser’s polarization.
2. Experimental setup
The experimental setup is shown in Fig. 1. A Ti: sapphire femtosecond laser amplifier system (Legend elite, Coherence Inc.) is employed to generate 1 kHz, 50 fs, 800 nm laser pulses with a linear polarization. The combination of the half-wave plate (HWP) and the polarization beam splitter (PBS) is used to adjust the laser power. The laser beam reflected from the PBS enters a Michelson interferometer. In order to generate orthogonally polarized double femtosecond laser beams, a quarter-wave plate is inserted in one arm of the Michelson interferometer. The power ratio between the preceding (E1) and subsequent (E2) laser beams is set to be 1.5:1 by a neutral density filter. The mirror (M4) is placed on a translation stage to adjust the time delay between dual laser beams. The dual laser beams are focused by a plano-convex lens (L1) with a focal length of 300 mm and normally irradiate the silicon wafer. The sample surface is placed 2.4 mm away from the geometrical focus of L1. The laser spot radius on the sample surface is calculated to be ∼35 µm. The sample is a 650-µm-thick (100) monocrystalline silicon wafer, which is mounted on a 3D electrically controlled translation stage (WNSC 400, Winner Optical Instruments Inc.). The CCD camera is used to monitor the laser spots of double laser beams to guarantee the complete overlapping of the two laser spots. Scanning electron microscopy (SEM) (VE 9800, Keyence Inc.) is used to investigate the morphology of the silicon surface after irradiated by double femtosecond laser beams.
Fig. 1. Experimental setup of fabricating curved ripples on the silicon wafer using double orthogonally polarized femtosecond laser beams with adjustable time delay. HWP: half-wave plate, PBS: polarization beam splitter, NPBS: 50/50 non-polarization beam splitter, QWP: quarter-wave plate, NDF: neutral density filter, M1-M4: high reflectance mirror, L1 and L2: plano-convex lens.
3. Results and discussions
Two orthogonally polarized femtosecond laser beams are employed to modify the morphology of the silicon wafer. In the experiments, the laser fluence ratio between the preceding beam (E1, vertically polarized) and the subsequent beam (E2, horizontally polarized) keeps constant to be 1.5:1. During the fabrication, the wafer is scanned along the vertical direction, i.e. parallel to the electric field of E1.
Figure 2 shows the SEM images of periodic ripples manufactured on the silicon surface by double 1-ps-time-delayed femtosecond laser beams. The energy fluence of the preceding laser beam is 0.12 J/cm2 close to the LIPSS formation threshold on silicon [20]. In Fig. 2, it is seen that when specific scanning speeds are used, such as 0.30-0.50 mm/s (Figs. 2(f)–2(h)), curved ripples with good uniformity are formed. When the target scanning speed is equal to or smaller than 0.20 mm/s, curved ripples can still be produced, but coexist with traces of excessive ablation. As far as we know, this is the first time that curved ripples are fabricated by double time-delayed linearly polarized femtosecond laser beams.
Fig. 2. SEM images of periodic ripples fabricated on the silicon surface by double time-delayed orthogonally polarized femtosecond laser beams. The vertically polarized laser beam (E1) strikes the silicon wafer earlier than the horizontally polarized laser beam (E2). During the fabrication, the wafer moves along Y direction (see (b)). The energy fluence of the preceding laser (E1) on the silicon surface is 0.12 J/cm2 and the time delay between double laser beams is 1 ps. The wafer’s scanning speeds are respectively 0.16 mm/s (a), 0.17 mm/s (b), 0.18 mm/s (c), and 0.19 mm/s (d), 0.20 mm/s (e), 0.30 mm/s (f), 0.40 mm/s (g), and 0.50 mm/s (h). Scale bar: 10 µm.
In order to obtain the parametric space that can fabricate curved ripples, the silicon surface is irradiated by double femtosecond laser beams with different laser fluences and time delays when different target scanning speeds are used. The dependences of the silicon surface morphology on the fabrication parameters are summarized in Fig. 3. It is seen that at a relatively low scanning speed, the silicon surface is excessively ablated and no periodic structures are formed, and contrastingly at a relatively high scanning speed, the deposited energy on the silicon surface is too small to induce surface modifications. When a moderate scanning speed is employed and the preceding laser beam’s fluence satisfies the condition of 0.11 J/cm2 ≤ F < 0.15 J/cm2, curved ripples can be formed (see green triangles in Fig. 3). Further increasing the scanning speed, the curved ripples become straight and perpendicular to the laser polarization of the preceding laser beam, which is consistent with the experimental results in Ref. [14]. However, in Ref. [14] no curved ripples are fabricated which may be attributed to the fact that the silicon sample does not move relative to the laser beam during the structure formation process. In our experiments, when double laser beams irradiate a fixed spot without scanning the target, the periodic structures are always dominated by the defects on the target surface or the noise on the intensity distribution of the laser beam, which obstructs the formation and recognition of the curved ripples. Therefore, no curved ripples are demonstrated on a fixed spot, which is also consistent with the results in Ref. [14].
Fig. 3. Dependences of the silicon surface morphology on the preceding laser fluence (F), the scanning speed (v), and the time delay (Δt) between dual femtosecond laser beams. Δt is 0.5 ps (a), 1 ps (b), and 2 ps (c), respectively. BT: below modification threshold, CR: curved ripples, SR: straight ripples, EA: excessive ablation without LIPSS.
Our previous experiments [20] show that when the laser fluence ratio between double laser beams is changed to be 1:1, no curved ripples can be formed even a large parameter space is explored. It is also known that curved ripples cannot be formed by single scalar femtosecond laser beam according to the experiments done by ourselves [20] and those by other researchers [16].
In addition, it can be seen from Fig. 3 that for Δt = 2 ps, the parametric region forming the curved ripples shrinks obviously compared with those for Δt = 0.5 ps and 1 ps. This infers that the relaxation of the E1 induced transient silicon surface has adverse effect on the formation of curved ripples. A larger time delay of 4 ps will result in the disappearance of the curved ripples.
To quantitatively analyze the relation between the shape of the curved ripples and the fabrication parameters, ImageJ [21] is used to extract the shape of the ripples. In Fig. 4(a), α is defined as the included angle between the tangential direction of the curved ripples and X axis. The origin of X axis locates in the center of the laser scanning line. Along X axis, we divide the laser scanning line into multiple segments, each of which has a width of ΔX = 2 µm. Figure 4(b) illustrates the variations of α along X axis for curved ripples fabricated at different target scanning speeds while the laser fluence and time delay between dual laser beams are kept constant. It is seen that α with relatively small value often appears at the edge of the scanning line, and large α always appears around the central part of the scanning line. This inspires one to relate α with the laser fluence distribution along X axis since the femtosecond laser beam used in the experiments is a Gaussian laser beam. As the scanning speed increases from 0.16 mm/s to 0.20 mm/s, the maximal α increases to ∼80°. As the scanning speed further increases to 0.50 mm/s, the maximal α decreases to ∼25°. It indicates that the local orientations of the curved ripples are not only related with the laser fluence but also the target scanning speed.
Fig. 4. (a) Definition of the ripples’ local orientation (α). α is defined as the included angle between the tangential direction of the curved ripples and X axis. The center of the scanning line is the origin of X axis. (b) Dependence of α on X for curved ripples fabricated by 1-ps-time-delayed double femtosecond laser beams with the preceding beam’s energy fluence of 0.12 J/cm2. The target scanning speed is changed from 0.16 mm/s to 0.50 mm/s.
Figure 5 shows the 2D fast Fourier transform (2D-FFT) of SEM pictures in Fig. 2. From Fig. 5, it can be determined that the average periods of ripples fabricated at different scanning speeds of 0.16 mm/s - 0.50 mm/s are all 730 nm, and the relative standard deviation (RSD) of the period is less than 0.8%. The spreading angle β (see Fig. 5(a)) of the diffraction orders in 2D-FFT pictures indicates the variation range of the curved ripples’ orientations (α). In Fig. 6, as v increases from 0.16 mm/s to 0.50 mm/s, β decreases accordingly, meaning that the ripples gradually become straight when the laser irradiation dose decreases.
Fig. 5. (a) – (h): Fourier transform of Figs. 2(a) – 2(h). The target scanning speed is indicated at the top of each picture. β is the spreading angle of the diffraction orders of the curved ripples.
Fig. 6. Spreading angle β of the diffraction orders of the curved ripples in 2D-FFT pictures in Fig. 5 as a function of the target scanning speed (square dots). The solid line is used to guide the eye.
Similar surface patterns can be also produced when the time delay between dual laser beams increases to 2 ps. The increase of the time delay does not change the variation tendencies of α with X. However, when the time delay decreases to 0.5 ps, the surface morphologies become slightly different from those at time delays of 1 ps and 2 ps, which is presented in Fig. 7. For Δt = 0.5 ps and v = 0.20 mm/s, the ripples can be divided into three regimes. For X < -15 µm, the ripples are straight and mainly perpendicular to the polarization of the preceding laser beam (E1), so the ripples in this region are predominantly generated by the laser beam (E1). For 0 < X < 23 µm, the ripples are also straight but have a 60° included angle relative to X axis. Only in the range of -15 µm < X < 0 and X > 23 µm, i.e. during the transition between two kinds of straight ripples, curved ripples appear. Similar morphology also appears for v = 0.30 mm/s, but when the scanning speed increases to 0.50 mm/s, curved ripples exist in the major part of the scanning line.
Fig. 7. Relation between the ripples’ orientation (α) and the location along X axis. The periodic ripples are fabricated by 0.5-ps-time-delayed double femtosecond laser beams with the preceding beam’s energy fluence of 0.12 J/cm2. The scanning speed varies from 0.20 mm/s to 0.50 mm/s.
The formation mechanism of the curved periodic ripples cannot be fully understood using the SPPs model without considering the transient non-equilibrium surface state excited by the preceding femtosecond laser beam (E1). The excitation of laser beam (E1) can induce the state and band filling, the renormalization of the band structure, and the free-carrier response, which all have impacts on the dielectric function of the silicon sample [22]. Using the dielectric model presented in Ref. [22] and the calculation method in Ref. [23], the lifetime of SPPs on the air-silicon interface is calculated and presented in Fig. 8.
Fig. 8. Lifetime of SPPs with a circular frequency of 2.356×1015 Hz (which is identical to the frequency of 800 nm laser) in the interface between the air and the excited silicon with different free electron densities.
It should be noted that the effective electron mass me* and the Drude damping time τD are two important parameters in calculating the SPPs’ lifetime. According to the numerical simulations by the first-principle approach [24], for pump femtosecond laser of 5×1012 W/cm2 (in our experiments the preceding beam’s intensity is 2.2×1012–3.0×1012 W/cm2) me* = 0.45me and τD = 25 fs can be used. The lifetime of SPPs calculated using me* = 0.45me and τD = 25 fs is several times longer than that calculated using the parameters in Ref. [22]. Based on the experimental results obtained by the pump-probe technique in Ref. [25], after striking by 800 nm femtosecond laser pulses with a pulse duration of 120 fs and an energy fluence of 0.14 J/cm2, the silicon surface can have a transient electron density of (1-4)×1022 /cm3. In our experiments, when curved ripples are formed, the parameters of the preceding laser beam are similar to those in Ref. [24] and the laser fluence is in the range of 0.11-0.15 J/cm2 as is shown in Fig. 3. Therefore, it is considered that the electron density of the silicon surface after striking by the preceding laser beam is roughly ∼1022 /cm3.
From Fig. 8, it is seen that the lifetime of SPPs increases as the increase of the free carrier density in silicon; for an excited silicon with the electron density of ∼1022 /cm3 in the conduction band, the SPPs lifetime can be up to 0.5 ps. Therefore, for double laser beams with small time delays, the electric fields of SPPs respectively induced by E1 and E2 might superpose each other and the ripples might be generated by the interference between the incident laser E2 and the superposed electric field of two SPPs. This may explain the origin of the ripples in Fig. 7 in the region of X > -15 µm.
However, for larger time delays, such as 1 or 2 ps, based on the calculation result in Fig. 8, the superposition between the two SPPs induced by E1 and E2 is negligible. To understand the formation mechanism of curved ripples fabricated by double laser beams with picosecond time delays, besides the ultrafast dynamics of SPPs on the excited air-silicon surface, the subsequent hydrodynamic evolution of the excited silicon may be also taken into account.
4. Conclusions
This paper demonstrates a new method of fabricating curved periodic ripples by using orthogonally polarized double femtosecond laser beams. This method is different from the conventional curved ripple formation methods which employ the vectorial femtosecond laser or directly modulate the laser polarization direction. The experimental findings in this paper reveal the important role of the preceding pulse excited surface in the surface nanopatterning. The formation of the curved ripples is a powerful challenge to the current LIPSS formation model. It is believed that the observation of curved ripples may motivate the further completion of the formation mechanisms of LIPSS. Seen from another perspective, the curved periodic ripples formed in this paper demonstrate that much more versatile patterns beyond those imagined and predicted by the current theory may be formed by dual-beam or multi-beam LIPSS.
Funding
National Natural Science Foundation of China (12074198, 61875093); Tianjin Research Program of Application Foundation and Advanced Technology of China (19JCYBJC16800, 20JCYBJC01040); Fundamental Research Funds for the Central Universities; State Key Laboratory of High Field Laser Physics; Shanghai Institute of Optics and Fine Mechanics, Chinese Academy of Sciences.
Disclosures
The authors declare no conflicts of interest.
References
1. M. Birnbaum, “Semiconductor surface damage produced by Ruby lasers,” J. Appl. Phys. 36(11), 3688–3689 (1965). [CrossRef]
2. J. Kim, S. Na, S. Cho, W. Chang, and K. Whang, “Surface ripple changes during Cr film ablation with a double ultrashort laser pulse,” Opt. Lasers Eng. 46(4), 306–310 (2008). [CrossRef]
3. J. Cong, J. Yang, B. Zhao, and X. Xu, “Fabricating subwavelength dot-matrix surface structures of Molybdenum by transient correlated actions of two-color femtosecond laser beams,” Opt. Express 23(4), 5357 (2015). [CrossRef]
4. H. Qiao, J. Yang, F. Wang, Y. Yang, and J. Sun, “Femtosecond laser direct writing of large-area two-dimensional metallic photonic crystal structures on tungsten surfaces,” Opt. Express 23(20), 26617 (2015). [CrossRef]
5. G. Giannuzzi, C. Gaudiuso, C. Di Franco, G. Scamarcio, P. M. Lugarà, and A. Ancona, “Large area laser-induced periodic surface structures on steel by bursts of femtosecond pulses with picosecond delays,” Opt. Lasers Eng. 114, 15–21 (2019). [CrossRef]
6. E. V. Barmina, E. Stratakis, M. Barberoglou, V. N. Stolyarov, I. N. Stolyarov, C. Fotakis, and G. A. Shafeev, “Laser-assisted nanostructuring of Tungsten in liquid environment,” Appl. Surf. Sci. 258(15), 5898–5902 (2012). [CrossRef]
7. S. Moradi, S. Kamal, P. Englezos, and S. G. Hatzikiriakos, “Femtosecond laser irradiation of metallic surfaces: Effects of laser parameters on superhydrophobicity,” Nanotechnology 24(41), 415302 (2013). [CrossRef]
8. Z. Ou, M. Huang, and F. Zhao, “The fluence threshold of femtosecond laser blackening of metals: The effect of laser-induced ripples,” Opt. Laser Technol. 79, 79–87 (2016). [CrossRef]
9. A. H. A. Lutey, L. Gemini, L. Romoli, G. Lazzini, F. Fuso, M. Faucon, and R. Kling, “Towards Laser-Textured Antibacterial Surfaces,” Sci. Rep. 8(1), 10112 (2018). [CrossRef]
10. M. Huang, F. Zhao, Y. Cheng, N. Xu, and Z. Xu, “Origin of laser-induced near-subwavelength ripples: Interference between surface plasmons and incident laser,” ACS Nano 3(12), 4062–4070 (2009). [CrossRef]
11. J. Bonse, A. Rosenfeld, and J. Krüger, “On the role of surface plasmon polaritons in the formation of laser-induced periodic surface structures upon irradiation of silicon by femtosecond-laser pulses,” J. Appl. Phys. 106(10), 104910 (2009). [CrossRef]
12. J. Bonse and J. Krüger, “Pulse number dependence of laser-induced periodic surface structures for femtosecond laser irradiation of silicon,” J. Appl. Phys. 108(3), 034903 (2010). [CrossRef]
13. J. Bonse, S. Hohm, S. V. Kirner, A. Rosenfeld, and J. Kruger, “Laser-Induced Periodic Surface Structures-A Scientific Evergreen,” IEEE J. Sel. Top. Quantum Electron. 23(3), 109–123 (2017). [CrossRef]
14. F. Fraggelakis, E. Stratakis, and P. A. Loukakos, “Control of periodic surface structures on silicon by combined temporal and polarization shaping of femtosecond laser pulses,” Appl. Surf. Sci. 444, 154–160 (2018). [CrossRef]
15. P. Liu, L. Jiang, J. Hu, W. Han, and Y. Lu, “Direct writing anisotropy on crystalline silicon surface by linearly polarized femtosecond laser,” Opt. Lett. 38(11), 1969 (2013). [CrossRef]
16. P. Liu, L. Jiang, J. Hu, S. Zhang, and Y. Lu, “Self-organizing microstructures orientation control in femtosecond laser patterning on silicon surface,” Opt. Express 22(14), 16669 (2014). [CrossRef]
17. W. Liu, L. Jiang, W. Han, J. Hu, X. Li, J. Huang, S. Zhan, and Y. Lu, “Manipulation of LIPSS orientation on silicon surfaces using orthogonally polarized femtosecond laser double-pulse trains,” Opt. Express 27(7), 9782 (2019). [CrossRef]
18. W. He and J. Yang, “Formation of slantwise orientated nanoscale ripple structures on a single-crystal 4H-SiC surface by time-delayed double femtosecond laser pulses,” Appl. Phys. A 123(8), 518 (2017). [CrossRef]
19. B. Zhao, J. Yang, J. Cheng, and C. Guo, “Capture of femtosecond plasmon excitation on transient nonequilibrium states of the metal surface,” Phys. Rev. Res. 2(3), 033418 (2020). [CrossRef]
20. N. Zhang and S.-C. Chen, “Formation of nanostructures and optical analogues of massless Dirac particles via femtosecond lasers,” Opt. Express 28(24), 36109 (2020). [CrossRef]
21. J. Schindelin, C. T. Rueden, M. C. Hiner, and K. W. Eliceiri, “The ImageJ ecosystem: An open platform for biomedical image analysis,” Mol. Reprod. Dev. 82(7-8), 518–529 (2015). [CrossRef]
22. K. Sokolowski-Tinten and D. von der Linde, “Generation of dense electron-hole plasmas in silicon,” Phys. Rev. B 61(4), 2643–2650 (2000). [CrossRef]
23. T. J. Y. Derrien, J. Krüger, and J. Bonse, “Properties of surface plasmon polaritons on lossy materials: Lifetimes, periods and excitation conditions,” J. Opt. 18(11), 115007 (2016). [CrossRef]
24. S. A. Sato, K. Yabana, Y. Shinohara, T. Otobe, and G. F. Bertsch, “Numerical pump-probe experiments of laser-excited silicon in nonequilibrium phase,” Phys. Rev. B 89(6), 064304 (2014). [CrossRef]
25. M. Garcia-Lechuga, D. Puerto, Y. Fuentes-Edfuf, J. Solis, and J. Siegel, “Ultrafast Moving-Spot Microscopy: Birth and Growth of Laser-Induced Periodic Surface Structures,” ACS Photonics 3(10), 1961–1967 (2016). [CrossRef]
