Open Access
Add to BibSonomy
Abstract
Direct writing of single-mode waveguides into crystalline silicon using ps laser pulses is presented. The embedded structures were fabricated by moving the focal position along the beam axis with the help of a long distance microscope objective. In situ monitoring during inscription was performed to analyze the processing dynamics. The waveguide generation is based on pronounced multi-pulse interaction at moderate pulse energies around 100 nJ. All samples were characterized in terms of mode field distribution and damping losses. Calculations indicate an induced refractive index change in the range of 10−3 to 10−2. Moreover, a Y-splitter was realized to demonstrate the potential of this process.
© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
The approach to bring electronic and photonic circuits together on the same silicon chip is one of the prime desires in communication industry today [1–3]. In particular, the steadily growing demand for processing speed and increased bandwidth requires access to integrated optical pathways packed in high density. One approach is the application of three-dimensional stacked waveguides embedded into the volume. Within this consideration, the writing of waveguides using ultrashort laser pulses is one of the most promising techniques. Today, the inscription of waveguides, which is based on locally induced modifications of the refractive index, is well-developed for different dielectrics and can be extended to functionalize optical pathways by the implementation of gratings or birefringent elements [4–9]. However, in case of silicon, the waveguide writing process cannot be simply transferred. The most limiting factors are the extraordinary high refractive index and the pronounced nonlinearity of silicon [10–15].
In 2005, we demonstrated silicon waveguides close to the surface using femtosecond laser pulses [10]. At this time, the induced modifications were only generated at the SiO2/Si interface and could not be transferred deeper into the volume. In 2016, Chambonneau et al. and Tokel et al. inscribed elongated structures in silicon with ns laser pulses [16]. Nevertheless, it was not completely clarified, whether the achieved structures were able to guide light due to the absence of mode field measurements. In 2017 Tokel et al. and Chanal et al. independently reported on internal structuring using fs laser pulses [16,17]. This approach enabled the first realization of waveguides in the bulk of silicon using fs laser pulses at a wavelength of 1550 nm with the help of a conventional aspherical lens [17,18]. During their investigations, the measured index change was estimated to be in the range of 10−4. The corresponding damping losses were not directly measured, but were declared to be extremely low.
Recently, we published investigations on localized in-volume structuring of silicon using ps pulses in the range from 0.8 to 10 ps using different pulse repetition rates [19]. Within this study, we demonstrated that permanent modifications can be achieved, as far as multi-pulse processing is applied. One key-element is the use of low-energy pulses in order to reduce nonlinear interactions like self-focusing or filamentation.
Based on our previous studies, here we present the inscription of waveguides into the volume of single crystalline silicon using ps laser pulses. We investigated the application of different laser repetition rates ranging from 30 to 400 kHz in order to identify potential pulse accumulation effects. During waveguide writing, in situ monitoring was performed to give insight into the writing dynamics and to evaluate optimum processing windows. The induced structures were characterized using shadowgraphic imaging and near-field detection during light propagation. These measurements were expanded by scattered light detection perpendicular to the waveguide axis, yielding direct information about the damping losses. In addition, the laser induced refractive index distribution was calculated by solving the Helmholtz equation with respect to the measured mode profile. The estimated index change was used afterwards to simulate beam propagation.
2. Experimental setup
The setup for waveguide writing is given in Fig. 1(a). It consists of an Er-doped fiber laser system (“Smart Light 50” from Raydiance. Inc.) and a 3D positioning system. The laser was running at a wavelength of 1552 nm delivering ultrashort laser pulses in the range from 0.8 ps to 10 ps at repetition rates up to 400 kHz. The maximum available pulse energy was 80 µJ (at 30 kHz). For waveguide writing only 0.8 ps pulses were applied. Previous studies have shown that the induced modifications in silicon exhibit a higher degree of localization using the shortest pulse lengths [19].
Fig. 1 (a) Schematic of the waveguide writing setup. The direction of waveguide writing is along the laser beam axis starting at the back-side of silicon. (b) Snapshot during in situ monitoring recorded with (a). The bright spot is the induced plasma at the focal position of the laser during processing. The dark vertical lines are already generated waveguides.
During waveguide writing, the laser beam was focused into 1 mm thick silicon (N/PH, <100>, 1-10 Ωcm) using a long distance microscope objective (20x, NA = 0.40). The writing process was managed with an Aerotech ANT 130 positioning system featuring sub-µm precision. In order to monitor the processing, an in situ observation microscope based on a broadband tungsten lamp in combination with an InGaAs camera was installed. A snapshot performed with the in situ installation can be seen in Fig. 1(b). During writing, the position of the laser focus could be easily identified through the localized bright spot based on plasma generation.
The writing direction was along the beam axis starting at the backside of the silicon and moving the focal position upwards. The approach of longitudinal writing was chosen due to our previous studies, where we have shown, that focusing in silicon involves strong spherical aberrations yielding highly elongated modifications along the beam axis [19]. In particular, for high refractive index materials like silicon, these aberrations strongly depend on the focusing depth. In our experiments we ignored this effect due to a traveling range of only 800 µm.
In order to prevent surface damaging, the processing was stopped approximately 100 µm – 200 µm below the front surface depending on the laser pulse energy. During our experiments, pulse repetition rates of 30 kHz, 100 kHz, 200 kHz and 400 kHz were studied along with pulse energies ranging from 10 nJ to 10 µJ (measured at the sample surface). The processing speed was varied from 1 µm/s – 100 µm/s.
After laser inscription, the induced modifications were characterized using the setup given in Fig. 2. It consists of two NIR pathways, which allow for simultaneous observation perpendicular and parallel to the waveguide axis. In order to investigate the morphologies of the buried structures shadowgraphic imaging was undertaken using an incoherent broadband light source (tungsten lamp). The light source for near-field inspection was a fiber coupled laser diode operating at 1550 nm. In addition to imaging the near-field mode profile of light guided in the induced structures, the scattered light was recorded perpendicular to the waveguide axis. Both measurements served for the characterization of the damping losses during light propagation.
Fig. 2 (a) Setup for waveguide characterization. (b) Broadband illumination of a set of waveguides recorded with this setup. During first inspections, laser generated modifications, which support waveguiding could be easily identified due to a bright region at the modification center. Processing parameters were: pulse repetition rate 400 kHz, writing speed = 20 µm/s.
3. Results and discussion
The generated structures reveal waveguiding behavior independent of the applied pulse repetition rates (30 kHz – 400 kHz), however, key element was the use of high pulse numbers per spot (> 104). This corresponds to writing speeds in the range from 1 - 100 µm/s with respect to the applied repetition rate. In particular, significant processing dependencies regarding the applied laser repetition rates could not be observed. As a consequence, the underlying mechanism for laser writing has to support relaxations times significantly above tens of µs. As a consequence, thermal accumulation can be neglected due to the pronounced heat conductivity of silicon as discussed in [19].
However, silicon exhibits extraordinary long charge carrier lifetimes up to the millisecond range especially at low dopant levels [20,21]. This is basically due to the fact that silicon is an indirect semiconductor. Consequently, accumulation based on avalanche ionization should have a major impact during the processing within the presented parameter range, however, this must be further investigated in future pump probe experiments.
Another important aspect during processing was the application of moderate pulse energies. Best results could be achieved at energies close above the processing threshold inside silicon. This threshold was observed to be in the range between ≈50 −70 nJ. Within this processing window, modifications capable of guiding light could be conveniently identified using shadowgraphic imaging, see Fig. 2(b).
For pulse energies below ≈50 nJ the writing process was observed to be unstable yielding statistical interruptions during inscription accompanied with final break-off. This behavior could be directly observed due to the in situ setup, see Fig. 1(a).
The characterization of the induced waveguides included near-field measurements. A typical near-field intensity profile measured at the waveguide exit buried 200 µm below the surface is given in Fig. 3. The corresponding structure was written using a repetition rate of 400 kHz and 110 nJ pulse energy. The writing speed was 20 µm/s. At these parameter settings the inscription was very stable, offering a high degree of repeatability. The observed near-field exhibits single-mode distribution. The measured full width of half maximum (FWHM, intensity) was in the range 3 – 5 µm for all waveguides. At increased pulse numbers we recognized broader modifications, which went along with larger mode field diameters.
Fig. 3 (a) Typical near-field intensity distribution and (b) cut through the near-field center in x and y orientation (coupling laser wavelength = 1550 nm). Processing parameters were: pulse energy = 110 nJ, repetition rate = 400 kHz, writing speed = 20 µm/s.
In order to support waveguiding, the applied pulse energy was crucial. In contrary, the writing speed (pulse number) could be varied in a comparatively wide range (e.g. 5 to 50 µm/s at 400 kHz). At pulse energies above ≈250 nJ, the induced modifications exhibited poor guiding properties. For these structures, shadowgraphic imaging revealed broader modifications with rough edges. These apparent inhomogeneities were responsible for strongly increased scattering during propagation. Structures written at higher pulse energies can be seen in Fig. 4 (right). For better visibility, the aperture of the imaging system was reduced to increase the depth of field.
Fig. 4 Set of modifications written at higher pulse energies. For better visibility, the aperture of the imaging system was reduced to improve the depth of field. Processing parameters: pulse repetition rate = 400 kHz, writing speed = 20 µm/s.
In order to analyze the waveguide losses, two independent measurements were performed. First, imaging of the scattered light distribution along the waveguide axis was performed to estimate the damping losses, see Fig. 2. A typical measurement, which belongs to the same waveguide displayed in Fig. 3 is plotted in Fig. 5. Typical values for the damping losses were in the range 4 – 6 dB/mm. These losses are based on scattering centers, which can be seen in Fig. 5(a). During waveguide writing, variations of the generated plasma were observable. This also implies the existence of inhomogeneities along the waveguide axis.
Fig. 5 (a) Measurement of the scattered light distribution perpendicular to the waveguide axis (coupling laser wavelength = 1550 nm). (b) Evaluation of damping losses based on (a). The values were filtered using a moving average filter over 24 pixel (33 µm). This measurement corresponds to the waveguide displayed in Fig. 3.
Second, the coupling efficiency was calculated by measuring the intensity distribution at the waveguide entrance and the mode profile at the waveguide exit, see Fig. 6 [22], which reveals values ranging from 35 - 55%.
Fig. 6 (a) Normalized near-field distribution of the laser source at the waveguide entrance using a standard microscope objective (20 x, NA = 0.35) and (b) normalized near-field distribution at the waveguide outlet (coupling laser wavelength = 1550 nm). Solving the overlap integral yielded a coupling efficiency of ≈53%.
In addition, the relative power transmission of the laser light before and after passing through the waveguide were measured with the help of a calibrated InGaAs camera. Since both measurements were recorded inside silicon, the effect of Fresnel reflections was included in both cases. Typical values of the transmitted power were Pin/Pout = 0.25 ± 0.05% (≈7.5 ± 1 dB). If the coupling losses of 50% (3 dB) are subtracted the remaining damping losses account to ≈5 dB/mm taking the waveguide length of 800 µm into consideration. This is in excellent agreement with the results obtained from the scattering measurements.
We evaluated the refractive index profile by solving the Helmholtz equation with respect to the obtained near-field distribution as described in [23]. The corresponding result for the mode profile in Fig. 3 can be seen in Fig. 7. As one can see, the simulations yield a strong increase of the refractive index at the center of the modification (2.5 × 10−3). This level of increase is significantly higher than the results presented in [17].
Fig. 7 Calculated refractive index distribution for a wavelength of 1550 nm, obtained by solving the Helmholtz equation. (a) Cut along x and y direction through the center, (b) surface plot.
A modification of the refractive index within this order of magnitude can be explained by laser induced transition from single crystalline silicon into polycrystalline silicon or even amorphous states as described in [24,25]. However, a detailed analysis of the exact material modification has to be studied in future experiments, which was not in the scope of this work.
In addition, numerical simulations of the beam propagation based on the estimated refractive index profile from Fig. 7 were performed using the software “BeamPROP” from “Synopsys”. The results show very good agreement with the measured near-field profile, see Fig. 8.
Fig. 8 Comparison of simulated and measured mode profile using “BeamPROP” (coupling laser wavelength = 1550 nm). The measured profile corresponds to Fig. 3. The numerical simulation is based on the refractive index profile given in Fig. 7.
In order to further demonstrate the potential of the writing process, a symmetric waveguide splitter was realized, see Fig. 9. The inscription was performed using the following parameter settings: repetition rate = 400 kHz, writing speed = 10 µm/s and pulse energy = 170 nJ. The fabrication of this integrated optical device was based on the interconnection of three separate waveguide sections in a Y-configuration exhibiting a total length of 800 µm. The junction was realized 240 µm after the inlet, followed by two identical branches with an opening angle of approximately 2 degrees. This corresponded to a spatial separation of 20 µm between the outlets. The splitting ratio was measured to be exactly 50:50. For the coupling into the splitter entrance an efficiency of ≈40% was obtained and the damping losses yielded ≈4.5 dB/mm. This means, in total, ≈45% of the launched power was transmitted through the Y-splitter.
Fig. 9 Characterization of the Y-splitter. (a) Normalized near-field distribution at the entrance. (b) Normalized near-field distribution at the outlet. (c) Scattered light distribution along the Y-splitter, plotted logarithmically for better visibility. The damping losses were around 4.5 dB/mm.
4. Conclusions
We demonstrated the inscription of buried waveguides into the bulk of single crystalline silicon using ps laser pulses. The inscription was realized in direction of the beam axis using a long distance microscope objective. The writing process was identified as pronounced multi-pulse interaction based on high pulse numbers (104-105) with energies around 100 nJ. Significant processing dependencies with respect to the investigated laser repetition rates (30 - 400 kHz) could not be observed. The corresponding near-field intensity profiles were in the range from 3 to 5 µm (FWHM), which implicates significant changes of the laser induced refractive index modification. The corresponding refractive index changes are in the order of 10−3 to 10−2. Modifications at this magnitude could be explained with a laser induced transition of crystalline silicon into polycrystalline and amorphous states, however this has to be investigated in future work. The samples were extensively characterized in terms of coupling and damping losses. As a result coupling efficiencies in the range 30 – 50% and damping losses in the range 4 to 6 dB/mm were observed. In order to demonstrate the potential of this technique, an integrated Y-splitter buried in silicon with a splitting ratio of 50:50 was fabricated, featuring damping losses of 4.5 dB/mm.
Funding
German Research Foundation (DFG); German Federal Ministry of Education and Research Project NUCLEUS (BMBF, 03IHS107A).
Acknowledgments
We gratefully acknowledge the superior work of Christiane Otto, with respect to intensive silicon sample preparation involving cutting and polishing. We acknowledge support within the framework of the graduate research school GRK 2101 “guided light, tightly packed” funded by the German Research Foundation (DFG) and the project NUCLEUS funded by the German Federal Ministry of Education and Research (BMBF, 03IHS107A).
References and links
1. A. D. Ashish Dhiman, “Silicon photonics: a review,” IOSR Journal of Applied Physics 3(5), 67–79 (2013). [CrossRef]
2. C. Gunn, “CMOS photonics for high-speed interconnects,” IEEE Micro 26(2), 58–66 (2006). [CrossRef]
3. I. A. Young, E. Mohammed, J. T. S. Liao, A. M. Kern, S. Palermo, B. A. Block, M. R. Reshotko, and P. L. D. Chang, “Optical I/O technology for tera-scale computing,” IEEE J. Solid-St. Circulation 45, 235–248 (2010).
4. K. M. Davis, K. Miura, N. Sugimoto, and K. Hirao, “Writing waveguides in glass with a femtosecond laser,” Opt. Lett. 21(21), 1729–1731 (1996). [CrossRef] [PubMed]
5. R. Osellame, S. Taccheo, M. Marangoni, R. Ramponi, P. Laporta, D. Polli, S. De Silvestri, and G. Cerullo, “Femtosecond writing of active optical waveguides with astigmatically shaped beams,” JOSA B 20(7), 1559–1567 (2003). [CrossRef]
6. K. Itoh, W. Watanabe, S. Nolte, and C. B. Schaffer, “Ultrafast Processes for Bulk Modification of Transparent Materials,” MRS Bull. 31(08), 620–625 (2006). [CrossRef]
7. M. Beresna, M. Gecevicius, and P. G. Kazansky, “Ultrafast laser direct writing and nanostructuring in transparent materials,” Adv. Opt. Photonics 6(3), 293–339 (2014). [CrossRef]
8. A. Zoubir, M. Richardson, L. Canioni, A. Brocas, and L. Sarger, “Optical properties of infrared femtosecond laser-modified fused silica and application to waveguide fabrication,” JOSA B 22(10), 2138–2143 (2005). [CrossRef]
9. F. Chen and J. R. V. de Aldana, “Optical waveguides in crystalline dielectric materials produced by femtosecond‐laser micromachining,” Laser Photonics Rev. 8(2), 251–275 (2014). [CrossRef]
10. A. H. Nejadmalayeri, P. R. Herman, J. Burghoff, M. Will, S. Nolte, and A. Tünnermann, “Inscription of optical waveguides in crystalline silicon by mid-infrared femtosecond laser pulses,” Opt. Lett. 30(9), 964–966 (2005). [CrossRef] [PubMed]
11. P. C. Verburg, G. R. B. E. Römer, and A. J. Huis Veld, “Two-temperature model for pulsed-laser-induced subsurface modifications in Si,” Appl. Phys., A Mater. Sci. Process. 114(18), 21958–21971 (2014).
12. V. V. Kononenko, V. V. Konov, and E. M. Dianov, “Delocalization of femtosecond radiation in silicon,” Opt. Lett. 37(16), 3369–3371 (2012). [CrossRef] [PubMed]
13. A. Mouskeftaras, A. V. Rode, R. Clady, M. Sentis, O. Utéza, and D. Grojo, “Self-limited underdense microplasmas in bulk silicon induced by ultrashort laser pulses,” Appl. Phys. Lett. 105(19), 191103 (2014). [CrossRef]
14. E. V. Zavedeev, V. V. Kononenko, and V. I. Konov, “Delocalization of femtosecond laser radiation in crystalline Si in the mid-IR range,” Laser Phys. 26(1), 016101 (2016). [CrossRef]
15. M. Lipson, “Guiding, modulation, and emitting light on silicon–challenges and opportunities,” J. Lightwave Technol. 23(12), 4222–4238 (2005). [CrossRef]
16. M. Chambonneau, Q. Li, M. Chanal, N. Sanner, and D. Grojo, “Writing waveguides inside monolithic crystalline silicon with nanosecond laser pulses,” Opt. Lett. 41(21), 4875–4878 (2016). [CrossRef] [PubMed]
17. O. Tokel, A. Turnalı, G. Makey, P. Elahi, T. Çolakoğlu, E. Ergeçen, Ö. Yavuz, R. Hübner, M. Zolfaghari Borra, I. Pavlov, A. Bek, R. Turan, D. K. Kesim, S. Tozburun, S. Ilday, and F. Ö. Ilday, “In-chip microstructures and photonic devices fabricated by nonlinear laser lithography deep inside silicon,” Nat. Photonics 11(10), 639–645 (2017). [CrossRef] [PubMed]
18. I. Pavlov, O. Tokel, S. Pavlova, V. Kadan, G. Makey, A. Turnali, Ö. Yavuz, and F. Ö. Ilday, “Femtosecond laser written waveguides deep inside silicon,” Opt. Lett. 42(15), 3028–3031 (2017). [CrossRef] [PubMed]
19. H. Kämmer, G. Matthäus, S. Nolte, M. Chanal, O. Utéza, and D. Grojo, “In-volume structuring of silicon using picosecond laser pulses,” Appl. Phys., A Mater. Sci. Process. 124(4), 302 (2018). [CrossRef]
20. P. T. Landsberg and G. S. Kousik, “The connection between carrier lifetime and doping density in nondegenerate semiconductors,” J. Appl. Phys. 56(6), 1696–1700 (1984). [CrossRef]
21. R. Häcker and A. Hangleiter, “Intrinsic upper limits of the carrier lifetime in silicon,” J. Appl. Phys. 75(11), 7570–7572 (1994). [CrossRef]
22. C. M. Kim and R. V. Ramaswamy, “Overlap integral factors in integrated optic modulators and switches,” J. Lightwave Technol. 7(7), 1063–1070 (1989). [CrossRef]
23. I. Mansour and F. Caccavale, “An improved procedure to calculate the refractive index profile from the measured near-field intensity,” J. Lightwave Technol. 14(3), 423–428 (1996). [CrossRef]
24. E. D. Palik, Handbook of Optical Constants – Silicon & Silicon (Amorphous), (Academic Press, 1985).
25. D. T. Pierce and W. E. Spicer, “Electronic structure of amorphous Si from photoemission and optical studies,” Phys. Rev. B 5(8), 3017–3029 (1972). [CrossRef]
