Open Access
Add to BibSonomyAbstract
Microstructures with the total length of hundreds of μ m were induced by fixing the focal point of the femtosecond laser at a certain depth in the bulk of SrTiO3 crystal. By different combination of the focusing conditions with the laser parameters, different morphologies have been observed, such as void array, necklace-shaped structures, continuous/ segmental filaments and etc. The possible mechanism of the formation of those diversiform structures is discussed.
©2007 Optical Society of America
1. Introduction
Laser fabrication with powerful femtosecond laser pulses has advantages such as contactless machining, 3D-well-controlled direct writing into dense materials and a smaller heat-affected zone compared with the nanosecond laser fabrication, so the most commonly used Ti: sapphire femtosecond laser has become a promising tool for fabricating various micrometer-scale optical components in the bulk of materials, including waveguides, couplers, gratings and so on. Among these potential applications, photonic device fabrication attracts special attentions for its widespread applications in optical communication and photonic integrated circuits. Photonic crystals are artificially engineered structures that exhibit a periodic variation in one, two, or three dimensions of the dielectric constant [1]. How to fabricate periodic structures is a key issue in photonic device fabrication. Traditionally, either complicated optical interference method involving several beams or the translation method of single-laser-beam direct-writing was used [2, 3]. However, these methods are both time and cost consuming. Recently, Kanehira et al. reported a novel method for fabricating periodic voids in borosilicate glasses [4]. In their report, nanovoid strings were self-formed in the laser propagation direction by focusing a single femtosecond laser beam at a fixed position inside the glass. Subsequently, similar research work was done in other glasses [5]. But the application of this novel method in crystals is rare now.
In this paper, we report the self-organized microstructures in SrTiO3 crystal. The cubic-crystal SrTiO3 belongs to the perovskite-structured transition metal oxide. It has some special characteristics such as high dielectric constant, high dispersion frequency, high linear and non-linear refractive index and etc [6, 7]. Based on these characteristics, strongly nonlinear processes can be expected during femtosecond laser fabrication inside it. In this paper, the different morphologies of the microstructures induced by Ti: sapphire fs laser are reported. Possible formation mechanisms of those structures are also discussed.
2. Experimental setup
The experimental setup is shown in Fig. 1. The SrTiO3 sample was cut in a cubic shape of 10mm×10mm×3mm and four-facet polished so that it could be observed from two orthogonal directions. 800nm mode-locked laser pulses from a regeneratively amplified Ti: sapphire laser, with 120fs pulse length and 1 kHz repetition rate, were focused in the bulk of SrTiO3 crystal by a high NA microscopic objective. The pulse energy could be adjusted by a neutral-density filter and be monitored by an energy meter. The pulse number could be controlled by an electric light shutter. During irradiation, the laser focal point was fixed at the certain depth inside the sample without translation. The optical photographs of the induced microstructures were captured by a CCD camera attached to a transmission optical microscope.
In the experiment, we first focused the fs laser beam on the surface of the SrTiO3 sample, and then raised the three-dimensional PC-controlled stage by a certain fixed distance d. By taking into account the linear refractive index of SrTiO3 (2.34 at 800nm) [8], the actual focal point was located about d×2.30 below the sample surface. In the following section, for simplicity, we use “focal depth d” to describe the above focusing situation. In our experiment, the (0 0 1) crystal plane (l0mm×10mm) was chosen as the entrance surface.
3. The diversiform microstructures induced inside SrTiO3 crystal
3.1 Single-pulse induced filament structures
Femtosecond laser pulse tightly focused in dielectrics can achieve very high intensity of 1013-1014W/cm2, leading to various nonlinear phenomenons, such as multiphoton ionization, tunneling ionization, avalanche ionization and Kerr self-focusing effect. The balance and competition between these processes can lead to different morphologies during the fabrication processes. The most common phenomenon is the filamentation induced in materials. In our experiment, we also found the filamentation tracks during the fabrication. As shown in Fig. 2, even single pulse could induce very long structures in bulk of SrTiO3. When a single pulse of 70 μJ was focused at the focal depth of 200μm through a 100×microscope, a continuous filament with the length of 117μm was formed [Fig. 2 (a)]. However, a single pulse of 35.7μJ focused by a 50×microscope at the same focal depth could lead to a segmental filament of length of 133 μm, which apparently break into six subsections [Fig. 2 (b)]. The length of the subsections and the intervals between them are both becoming longer and longer along the laser propagation direction.
Fig. 2. Optical microscope photographs of the filaments induced by a single pulse. The pulse energy was 70μ J (a) and 35.7μ J (b), respectively. The focusing lenses were 100× and 50×, respectively. The focal depth both were 200μ m.
3.2 Multiple-pulse induced void array structures
Figure 3 shows a side-view optical microscope photograph in the SrTiO3 sample exposed for 1/16 s (about 63 shots) at the focal depth of 400μm. The focusing lens was a 50×microscope objective. The pulse energy was chosen as 12.5μJ [Fig. 3(a)] and 14.25μJ [Fig. 3(b)], respectively. Under the microscope, we measured the lengths of the two void strings as 63.8μm [Fig. 3(a)] and 78.1μm [Fig. 3(b)], respectively. Noticeably, the diameters of the voids in two strings are both about 1.4 μm while the intervals between the neighboring voids become bigger from 1.5μm to 4.2μm along the propagation direction of the laser beam. Besides the void strings with the nonuniform spacings, some longer void strings with more uniform spacings are shown in Fig. 4. The irradiation condition is as follows: 50× microscope objective was used for focusing the laser beam at the focal depth of 400μm. 8 laser pulses with pulse energy of 36.7μJ [Fig. 4(a)] and 38.1μJ [Fig. 4(b)] were used for the irradiation process. As shown in Fig. 4, void array with the length of 133.7μm [Fig. 4(a)] and 145.6μm [Fig. 4(b)] lines up along the propagation direction of the laser beam. The number of voids is 46 in [Fig. 4(a)] and 58 in [Fig. 4(b)] respectively. However, the diameters of voids (about 1.2 μm in both) in Fig. 4 don’t vary much along the laser beam direction like that in Fig. 3. We notice that the void array is longer, more dense and has more uniform spacings between the neighboring voids, compared with those in Fig. 3. Comparing the results in Fig. 3 and Fig. 4, we can conclude that the formed structures of void array are very sensitive to the laser irradiation parameters, such as pulse energy and pulse number.
Fig. 3. Optical microscope photograph of the void array formed in SrTiO3 crystal at the focal depth of 400μ m after 1/16 s irradiation of laser pulses with energy of 12.5μ J (a) and 14.25μ J (b). 50× microscope lens was used.
Fig. 4. The side view of two void strings in SrTiO3 crystal induced by 8 laser pulses with energy of 36.7μ J (a) and 38.1μ J (b). The focal depth is 400μ m and the 50× microscope was used.
To investigate the influence of the focusing condition (numerical aperture and focal depth) on the induced void array, the laser beam with the pulse energy of 41.7μJ was focused by a 100× microscope objective with NA=0.9 at six different focal depths (100μm, 200μm, 300μm, 400μm, 500μm, and 600μm) in the SrTiO3 crystal. The pulse number used was 125. The induced void arrays are shown in Fig. 5(a). It is observed that the induced microstructures basically consist of two parts: The first part is the irregular damage tracks around the focal point on the top, and the following second part is the very regular void array on the bottom. Fig. 5(b) shows the relationship of the void strings’ lengths with the focal depths. We can see that the lengths of the void strings and the voids’ quality do not keep increasing with the increased focal depth. There exists an optimal fabrication depth of 300μm where the well-aligned void array with the longest length of 267.3μm was induced. The reason for the existence of the optimal focal depth may be as follows: for the shallower focal depth, the much bigger damage track on the top indicated that much more plasma electrons were generated through the avalanche ionization initiated by fs-induced multiphoton ionization. Those massive electrons in turn absorbed a large number of pulse energy, and lowered the power of the pulse very much so that pulse could not maintain its power above the self-focusing critical power for a long distance. But for the deeper focal depth, the large group velocity dispersion (about 6865fs2/cm) [8] other than plasma absorption played a dominating role in lowering the peak pulse power by splitting or broadening the pulse, and in addition to this, the spherical aberration, which is caused by the large index mismatch at the interface between the air and the SrTiO3, is more distinct when the focal depth is deeper, so the blurring of the focal spot caused by the serious spherical aberration may also lower the peak power of the pulse. Apparently, compared with the void strings induced by 50× focusing in Fig. 4, the 100× focusing condition leads to more regular, rounder and more-densely-aligned voids. Those voids in diameter of about 1.6μm align regularly and the intervals between the neighboring voids are so small that the voids are almost connected with each other to form necklace-shaped structures [the inset in Fig. 5(a)]. It’s worth noting that, in our experiment, no such void strings or any damage tracks were induced by the lower-NA microscope objectives (eg. 10×NA=0.30 or 20×NA=0.45) even at very high pulse energy of 100μJ. So we can conclude that high numerical aperture is critical to the formation of the void strings in SrTiO3 crystal, this phenomenon is maybe related to the high internal damage threshold in the bulk of SrTiO3.
Fig. 5. The void strings induced at different focal depths. The pulse energy was fixed at 41.7μ J, and the pulse number was 125. The focusing objective lens with 100× and NA=0.9 was used.
4. Dissension
Those different filament structures can be explained by the single-pulse nonlinear propagation in medium. N. Akozbek once analyzed the femtosecond pulse propagation in air by variational analysis [9]. He pointed out that the beam radius on the entrance surface and the input power as well as the focal length of the focusing lens had an important effect on the beam propagation. Different combinations of laser parameters with the focal condition can lead to different propagation phenomenon such as single focusing or multiple refocusing. The femtosecond pulse propagation in solids is very similar to its propagation in gases except for the different dominating ionization mechanism [10]. So it is possible that the single continuous filament induced in Fig. 2(a) possibly corresponds to the single focusing and the following self-guided propagation, while the breaking filament in Fig. 2(b) may result from the multiple refocusing behavior of the pulse nonlinear propagation. Noticeably, the changing tendency of the spacings among the six subsections agrees well with the prediction of the analytical analysis by N. Akozbek, that for a given initial beam radius the distance between focusing and refocusing increase as the power is decreased. There forth, we can conclude that dynamic spatiotemporal evolution of single fs pulse propagation in SrTiO3 crystal may be the most possible formation mechanism of the different morphologies during single-pulse fabrication.
As for the formation mechanism of multiple-fs-pulse induced void string structures, maybe we can obtain some enlightenment from the experimental details. When multiple pulses were employed for fabrication, a filament track was also observed. As shown in Fig. 6, along the laser beam direction, the filament of the length of about 47μm appears before the induced void array. The phenomenon that the filament existed along with the well-aligned void array indicates that the self-organized void array may be formed on the base of the filaments. To certify this conjecture, under the same focusing conditions and the same pulse energy, we compare the single-pulse-induced structure and the two-pulse-induced structure. As is shown in Fig. 7, single pulse induced a uniform filament with 1.5μm wide in the transverse direction and 108μm long in the laser propagation direction, while two pulses fabricated a structure with void array of the diameter of 1.5μm on the top and uniform filament on the bottom. We suspect that the formation mechanism of the void array by multiple pulses is as follows: The first pulse plays a role of inducing a filament. The filament region in the solid indicates a different refractive index from that of the unirradiated region. The index change in the filament generally results from the fs-induced color centers, defects or stress [11]. Many researchers once reported that the measured index change induced by femtosecond filamentation has a parabolic index profile [12, 13].We speculate that similar refractive index change also exists in the first-pulse-induced filament. So, when the second pulse entered into the sample after the first pulse’s irradiation, it seemed to propagate in a “parabolic-index optical fiber” induced by the first pulse. M. Karlsson once reported that light in a parabolic index optical fiber refocused periodically during the propagation [14]. In his paper, the periodic refocusing phenomenon results from the competition of the defocusing effect induced by the diffraction and the self-focusing effect caused by both the kerr self-focusing and the parabolic index profile in the transverse direction. Although in our case, during the propagation of the second pulse, the pulse could also be defocused by the plasma in addition to the diffraction defocusing, we still can expect that the refractive index change induced by the first pulse acted like a serials of positive lens and enhanced the previously dominating kerr focusing effect. The enhanced total focusing effect made the second pulse fail in propagating in the self-guiding mode like that of the first pulse. Instead, the second pulse may experience fast focusing and defocusing over and over again, and then present the periodic refocusing phenomenon. The accumulation of massive electrons in a very small localized region around the refocusing focus led to the onset of microexplosion and finally left a void behind. This may be the underlying mechanism of the void array formed by multiple pulses.
Fig. 6. Optical microscope photograph of the filamentation before the void array appeared. Laser beam with pulse energy of 24μ J was focused at the focal depth of 1 50μ m by a 50× object lens. The irradiation time was 1/63 s and 1/32 s respectively.
Fig. 7. Optical microscope photograph of the microstructures induced by one pulse (above) and two pulse (below). Laser beam with pulse energy of 70μ J was focused at a focal depth of 100μ m beneath the surface by a 100× object lens.
5. Conclusion
In conclusion, we report that different microstructures of the length of hundreds of micrometers can be self-fabricated in SrTiO3 crystal by single laser beam direct writing without translation. The influence of the laser parameter and focusing condition on the induced structures is also investigated; some possible mechanism has been discussed. Among these structures, quasi-periodically and densely-aligned void array can be applied in high-density 3D optical storage, and also have the potential to be used as photonic crystal for optoelectronic devices, while the single-pulse induced uniform filaments can find its application in forming 3D volume gratings, waveguides and microreflectors. SrTiO3 crystal is a key material in the oxide-based electronic device, so the microfabrication of it has much practical application value.
Acknowledgments
This work was supported by the National Basic Research Program of China (No.2006CB806000).
References and links
1. B. Gersten and J. Synowczynski, “Simulation of realizable photonic bandgap structures with high refractive contrast,” Mat. Res. Soc. 692, K.5.6.1 (2002).
2. K. Venkatakrishnan, N. R. Sivakumar, C. W. Hee, B. Tan, W. L. Liang, and G. K. GAN, “Direct fabrication of surface-relief grating by interferometric technique using femtosecond laser,” Appl. Phys. B 77, 959 (2003).
3. K. Itoh, “Laser microengineering of photonic devices in glass,” J. Laser Micro/Nano. 1, 1 (2006). [CrossRef]
4. S. Kanebira, J.H. Si, J.R. Qiu, K. Fujita, and K. Hirao, “Peoriodic nanovoid Structures via femtosecond laser irradiation,” Nano Lett. 5, 1591 (2005). [CrossRef]
5. E. Toratani, M. Kamata, and M. Obara, “Self-fabrication in fused silica by femtosecond laser processing,” Appl.Phys. Lett. 87, 171103 (2005). [CrossRef]
6. H. Katsu, H. Tanaka, and T. Kawai, “Anomalous Photoconductivity in SrTiO3,” Jpn. J. Appl. Phys. 39, 2657 (2000). [CrossRef]
7. R. Adair, L. L Chase, and S. A. Payne, “Nonlinear refractive index of optical crystals,” Phys. Rev. B 39, 3337(1989). [CrossRef]
8. M. Cardona, “Optical properties and band structure of SrTiO3 and BaTiO3,”Jpn. J. Appl. Phys. 39, 2657 (2000).9.N. Akozbek and C. M. Bowden,“Femtosecond pulse propagation in air: Variational analysis,” Phys. Rev. E 61,4540 (2000).
9. S. L. Chin, “Some fundamental concepts of femtosecond laser filamentation,” J. Korean Phys. Soc. 49, 281(2006).
10. Q. Z. Zhao, J. R. Qiu, X. W. Jiang, C. J. Zhao, and C. S. Zhu, “Mechanisms of the refractive index change in femtosecond laser-irradiated Au3+-doped silicate glasses,” J. Appl. Phys. 96, 7122 (2006). [CrossRef]
11. S. -H. Cho, H. Kumagai, and K. Midorikawa, “Fabrication of single-mode waveguide structure in optical multimode fluoride fibers using self-channeled plasma filaments excited by a femtosecond laser,” Appl. Phys. A 359, 77 (2003).
