We report on waveguide writing in fused silica with a novel commercial femtosecond fiber laser system (IMRA America, FCPA µJewel). The influence of a range of laser parameters were investigated in these initial experiments, including repetition rate, focal area, pulse energy, scan speed, and wavelength. Notably, it was not possible to produce low-loss waveguides when writing with the fundamental wavelength of 1045 nm. However, it was possible to fabricate telecom-compatible waveguides at the second harmonic wavelength of 522 nm. High quality waveguides with propagation losses below 1 dB/cm at 1550 nm were produced with 115 nJ/pulse at 1 MHz and 522 nm.
© 2005 Optical Society of America
Internal laser modification of transparent glasses is a promising means for fabricating integrated micro-optical devices and optical circuits in novel 3-D architectures. While ultraviolet (UV) laser induced glass compaction [1,2] enables limited 3-D refractive index modification suitable for the production of buried optical waveguides [3,4], femtosecond lasers provide greater flexibility to construct complex 3-D structures. For the latter, nonlinear absorption confines laser interactions to the focal volume and reduces collateral damage in the surrounding material. Femtosecond lasers have been applied to fabricate buried waveguides [5–12], optical amplifiers [6,12,13], beamsplitters [6,7], directional couplers [8,9], long-period fiber gratings  and birefringent transmission gratings  in a variety of transparent materials. Despite these demonstrations, significant gaps remain in optimizing the laser interactions to increase processing speed and improve control over device characteristics.
In this paper, a novel commercial femtosecond Yb-fiber laser system (IMRA America, FCPA µJewel) was employed to bridge a gap between two classes of laser types that have typically been applied in waveguide writing studies. On one side are amplified Ti:Sapphire laser systems that provide high pulse energies (~1 mJ) at low repetition rates (~1 kHz). Such lasers frequently yield waveguides with asymmetric refractive index profiles that exhibit significant coupling loss and birefringence. On the other side are low energy (~10 nJ) and high repetition rate (~10 MHz) oscillator-only systems [8,10,11,16]. These lasers yield smoother and more symmetric guiding regions due to the heat accumulation effects arising when the interval between laser pulses is less than the time required for the absorbed laser energy to diffuse out of the critical absorption volume [8,10,17]. However, such low energy oscillators demand a high numerical aperture (NA) focusing objective, whose short working distance limits the fabrication of 3-D devices. The combination of high pulse energy and MHz repetition rates was only recently demonstrated with an amplified femtosecond Yb-fiber laser to form waveguides in alkali-free borosilicate glass [18,19].
The laser is a compact commercial femtosecond pulse source providing flexible operating conditions ideally matched to writing optical waveguides in glasses. Variable repetition rate with energies of >2.5 µJ at 100 kHz and >150 nJ at 5 MHz facilitate rapid identification and optimization of laser processing windows for various target glasses. The ~375-fs pulses were applied to fundamental studies of waveguide writing in fused silica. High quality waveguides are shown to be possible without ultrashort pulses (<200 fs), concurring with recent observations with other types of ultrafast lasers [13,18–21]. To expand the range of waveguide writing conditions, the µJewel laser was frequency doubled to 522 nm and compared with the fundamental. A marked improvement in waveguide refractive index contrast is noted at 522 nm in contrast to a previous report of a negative change refractive index when a frequency-doubled Ti:Sapphire amplifier was applied to fused silica . In the same study, a frequency-doubled high repetition rate Ti:Sapphire oscillator was unable to modify the refractive index of fused silica .
2. Experimental method
The arrangement for laser writing of optical waveguides is shown in Fig. 1. The laser is an amplified Yb-fiber laser producing 375-fs pulses at 1045 nm with an M2 of 1.5. The system consists of three diode-pumped Yb-fiber stages: an oscillator and pre-amp operating at ~40 MHz, and a large mode area Yb-fiber power amplifier. The repetition rate is controlled by an acousto-optic downcounter between the pre-amp and power amplifier stages, and can be changed in ten minutes to any value between 100 kHz and 5 MHz. The laser output at the fundamental 1045 nm wavelength is: 200 kHz with 2 µJ/pulse, 1 MHz with 660 nJ/pulse, 2 MHz with 300 nJ/pulse, and 5 MHz with 150 nJ/pulse.
The second harmonic (522 nm) was generated by a 1.85-mm thick BBO nonlinear crystal placed at the focus of a two-lens telescope consisting of 150-mm and 100-mm focal length singlet lenses. A polarizing beam splitter removed the unconverted fundamental radiation. No measurement was made of the pulse duration of the second harmonic beam, but given the phase matching bandwidth and group velocity mismatch for the BBO crystal we calculate that the pulse duration was ≤500 fs. The SHG conversion efficiency of this arrangement was 30%. However, this telescope arrangement was not optimized for the narrow phase matching acceptance angle or the large spatial walk-off of the BBO crystal, resulting in an elliptical 522-nm beam profile with a ~3:1 ratio between the vertical and horizontal beam widths. In these experiments, the long axis of the elliptical beam was perpendicular to waveguide writing direction.
Waveguides were written transversely, with the laser beam directed vertically downward and focused at normal incidence into the fused silica target (Corning 7940). The sample was translated with a motorized stage (Newport TS200) at scan speeds of between 0.05 and 40 mm/s, with the laser polarization parallel to the direction of motion. A variety of aspherical lenses (NA=0.65, 0.55 and 0.40) as well as a microscope objective (Olympus UPLAPO40X, NA=0.85) with variable spherical aberration correction were applied at the fundamental laser wavelength of 1045 nm. Only the microscope objective was applied to the 522-nm study. Waveguides were written at a depth of 150 µm with the aspheric lenses at 1045 nm and a depth of 400 µm when using the microscope objective at both 1045 and 522 nm. Table 1 lists the range of average laser powers and fluences on target for the different conditions tested. Fluence values were based on calculated focal spot sizes, and included beam asymmetry in the second harmonic case.
After waveguide writing, the edges of the fused silica samples were ground ~500 µm and polished to high optical grade. Light from a Nettest Tunics-BT tunable laser was butt-coupled into the input facet of the waveguides with standard SMF-28 fiber. Near-field mode profiles were obtained by imaging the waveguide mode at the output facet of the waveguide sample. A Spiricon SP-1550M phosphor-coated CCD and frame grabber were used with a 100X microscope objective and tube to capture the mode profiles. Mode-field diameter (MFD) measurements were made using the D4σ (second moment) calculation from Spiricon’s LBA-PC software. The total insertion loss was measured by butt-coupling two SMF-28 fibers to the end facets of the waveguide sample and launching light from a 1550-nm laser diode. To minimize measurement error and avoid Fresnel losses, the air gaps between the sample and fibers were filled with index matching oil (n=1.46). The insertion loss was obtained from the ratio of the power signal (Newport 818-IG) transmitted through the waveguide and that transmitted directly from the input and output fibers.
3. Results and discussion
3.1 Processing window for waveguide writing
The 1045-nm light was applied at scan speeds from 0.1 to 40 mm/s, incident pulse energies from 400 to 560 nJ, and aspheric lenses with focal NAs of 0.40, 0.55 and 0.65 NA at repetition rates of 0.2, 1, 2 and 5 MHz. Additional attempts were made using the 0.85 NA microscope objective, with scan speeds from 0.5 to 10 mm/s and pulse energies of 100 to 145 nJ. Material modification and waveguiding was only evident in a narrow processing range that applied the full laser power. Weakly guiding waveguides were observed at 585 mW focused with the 0.4 NA aspheric lens and scan velocities of 0.1 to 0.5 mm/s. The lowest-loss waveguide, written at 0.1 mm/s, had an insertion loss of approximately 15 dB for the 5-cm long sample.
Further improvements were sought by frequency doubling of the laser light to 522 nm. The frequency doubled 522-nm light offers much stronger absorption in wide-bandgap fused silica by reducing the order of multiphoton absorption. The optical arrangement only permitted efficient second harmonic generation at 100 kHz and 1 MHz repetition rates. The majority of waveguides were written using pulse energies (on target) of 115 nJ, 90 nJ and 70 nJ at a repetition rate of 1 MHz and with scan velocities of 0.05 to 5 mm/s. Waveguiding was also observed in features written at 100 kHz with 115 nJ, however writing was only attempted at 0.1 mm/s and poor mode quality and high insertion loss did not reveal an attractive laser processing window.
From basic Gaussian optics, the estimated waist diameter and depth of focus of the elliptical 522-nm beam were 2w 0=0.9 µm and 2z 0=2.4 µm, respectively. For the maximum pulse energy of 115 nJ, the average single pulse fluence was ~20 J/cm2. A useful measure of exposure for optimizing the processing window is net fluence:
where Fp is the average single pulse fluence, R is the repetition rate (1 MHz), 2w 0 is the waist diameter and vs is the translation stage scan velocity. For the waveguides written with maximum single pulse fluence, the net fluence range was 2 to 250 kJ/cm2.
Figure 2 shows an overhead microscope image of two waveguides written in fused silica at 522 nm. A uniform central region with a faint outer modified zone is observable. The modes are guided in the central region, which is approximately 2 µm in diameter. This diameter is slightly larger than the estimated focal diameter which we infer to arise from heat diffusion and heat accumulation effects. Such thermal diffusion effects are much more pronounced in borosilicate glasses and offer greater control of waveguide diameter . To date, there has been no definitive evidence of waveguide diameter control through thermal accumulation effects in pure fused silica [10,11].
3.2 Mode profiles
Examples of guided-light mode profiles of waveguides written with 522-nm light at 115-nJ pulse energy are shown in Fig. 3. The modes are nearly circularly symmetric at 1550 nm, with some asymmetric distortion evident along the writing laser direction, particularly at 0.2 and 0.5 mm/s. The profiles suggest a larger refractive index gradient was formed on the laser irradiation side of the waveguide Mode field diameter increases with scan velocity, varying from 10 to 20 µm (average of width along both axes) for scan speeds between 0.05 and 0.5 mm/s. For scan velocities greater than 0.5 mm/s, modes are not well-formed, which is indicative of weak refractive index modification resulting in larger loss. The trend of MFD increasing with scan velocity leads to greater coupling loss at high scan velocity due to the mode mismatch to SMF-28 fiber (measured MFD of 9.7 µm).
At lower pulse energies (90 nJ, 70 nJ), the modes are generally larger due to decreased index change at lower fluence. As at 115 nJ, increased scan speed resulted in increased insertion loss thus guiding was not detectable where the pulse energy was less than 70 nJ and the scan speed was greater than 0.5 mm/s..
3.3 Refractive index contrast
To estimate the refractive index contrast of the waveguides, Lumerical MODE Solutions was employed. The assumptions made were a circular waveguide cross section with a core diameter of 2 µm (based on optical microscope image) and a step refractive index profile in the radial direction. For the maximum laser exposure (0.05 mm/s, 115 nJ/pulse), the observed 10-µm MFD yielded an estimated refractive index change of Δn=1.0×10-2. The numerical results were verified with a Bessel function analytical solution. Index changes of this magnitude are desirable for efficient coupling to standard single-mode telecommunications fiber where Δn=5×10-3.
3.4 Waveguide losses
To measure the propagation loss, the sample was diced into three different lengths, and each facet was highly polished. The measured insertion loss is plotted against sample length in Fig. 2 and equations for a least square fit are shown for each scan velocity tested. The slope of the linear fit represents the propagation loss in dB/cm and the y-intercept gives the coupling loss from both facets. The coupling loss increases with scan velocity, due to increased mode mismatch between the waveguide and fiber. The minimum coupling loss is about 1.4 dB at 0.1-mm/s scan speed, where the waveguide mode field diameter best matches that of the fiber. The propagation loss is ~0.9 dB/cm for scan velocities between 0.05 and 0.2 mm/s, and decreases to ~0.4 dB/cm (+/-0.2 dB/cm uncertainty) at 0.5 mm/s. Waveguiding was also observed for writing speeds of ≥1 mm/s, but the large insertion losses commensurate with the large mode profiles (>20 µm) precluded a more detailed analysis.
To date, the best reported propagation loss for waveguides written in fused silica is 0.1dB/cm . By comparison, industry-adopted planar lightwave circuit (PLC) fabrication technologies such as plasma-enhanced chemical vapor deposition and flame hydrolysis offer <0.05 dB/cm propagation loss. It is evident that further work is required to improve the propagation and coupling loss of waveguides written with the FCPA µJewel laser. Two significant factors contributing to the 0.9-dB/cm loss are a non-optimized second-harmonic beam and low quality motion stages. The lowest loss waveguides were written at the maximum power generated by the frequency doubling setup. Both the conversion efficiency and output beam symmetry can be significantly improved with a better choice of doubling crystal to enable the writing of higher quality waveguides. Images of scattered light also showed high intensity scattering points that we infer to originate from backlash and non-uniform translation of the lead screw-type motion stages. Higher quality motions stages are expected to reduce such losses.
The results indicate the presence of a weak thermal accumulation effect for waveguide writing in fused silica at 1-MHz repetition rate. Thermal diffusion and heat accumulation effects are expected to play a greater role above 1 MHz in defining low-loss index-modified waveguides in fused silica waveguides. In order to further investigate the influence of thermal accumulation, work is ongoing to improve the frequency doubling efficiency thereby providing higher laser energy and enabling exploration of waveguide writing at repetition rates ≥1 MHz. Likewise, the influence of repetition rates between 100 kHz and 1 MHz on waveguide writing at 522 nm is the subject of continued research.
We have shown that high quality waveguide writing is possible in fused silica using a 375-fs Yb-fiber laser. This commercial laser system is more attractive in an application setting because of superior reliability and robustness as compared with femtosecond Ti:Sapphire amplifier and oscillator-based lasers. In addition, we have demonstrated a new and attractive laser processing window for waveguide formation in fused silica, at the second harmonic wavelength of 522 nm.
In summary, low-loss waveguides have been written in fused silica using a commercial femtosecond fiber laser (IMRA America, FCPA µJewel). The laser was frequency doubled with a BBO crystal because the 1045-nm fundamental wavelength was not amenable to forming low-loss waveguides. To our knowledge, this is the first time a frequency-doubled ultrafast laser has been used to write waveguides in fused silica. This work demonstrates the promise of the ultrafast fiber laser for optical waveguide formation. The laser has the added benefit of turn-key operation, smaller footprint, and greater stability than available from ultrafast Ti:Sapphire oscillators and amplifiers.
We would like to thank Jianzhao Li, Dragan Coric, Rajiv Iyer, Chris Valdivia, Haibin Zhang, Stephen Ho, and Amir Nejadmalayeri from the University of Toronto and Zhenlin Liu, Gyu Cho, Catalin Florea, Fumiyo Yoshino, Tadashi Yamamoto and James Bovatsek from IMRA America for helpful discussions. The University of Toronto research was supported by the Canadian Institute for Photonics Innovation and the Natural Sciences and Engineering Research Council of Canada. Shane Eaton was supported by the Walter C. Sumner memorial fellowship.
References and links
1. N.F. Borrelli, C. Smith, D.C. Allan, and T.P. Seward III, “Densification of fused silica under 193 nm excitation,” J. Opt. Soc. Am. B. 14, 1606–1615 (1997). [CrossRef]
2. N.F. Borrelli, C. Smith, and D.C. Allan, “Excimer-laser-induced densification in binary silicate glasses,” Opt. Lett. 24, 1401–1403 (1999). [CrossRef]
3. D. Coric, P.R. Herman, K.P. Chen, M.X. Wei, P.B. Corkum, R. Bhardwaj, and D.M. Rayner , “Contrasts in writing photonic structures with ultrafast and ultraviolet lasers,” in Optical Devices for Fiber Communication III , M. J. F. Digonnet, ed., Proc. SPIE 4638, 77–84 (2002).
4. C. Valdivia, X. Midori Wei, D. Coric, and P.R. Herman, “F2 laser-induced visible- and infrared-confining buried waveguides in fused silica,” Conference on Lasers and Electro Optics 2003, CWI4.
5. K. Miura, J. Qiu, H. Inouye, T. Mitsuyu, and K. Hirao, “Photowritten optical waveguides in various glasses with ultrashort pulse laser,” Appl. Phys. Lett. 71, 3329–3331 (1997). [CrossRef]
6. M. Will, S. Nolte, and A. Tunnermann , “Single- and multimode waveguides in glasses manufactured with femtosecond laser pulses,” in Commercial and Biomedical Applications of Ultrafast and Free-Electron Lasers , G. S. Edwards, J. Neev, A. Ostendorf, and J. C. Sutherland, eds., Proc. SPIE 4633, 99–106 (2002).
7. D. Homoelle, S. Wielandy, A.L. Gaeta, N.F. Borrelli, and C. Smith, “Infrared photosensitivity in silica glasses exposed to femtosecond laser pulses,” Opt. Lett. 24, 1311–1313 (1999). [CrossRef]
8. K. Minoshima, A.W. Kowalevicz, I. Hartl, E.P. Ippen, and J.G. Fujimoto, “Photonic device fabrication in glass by use of nonlinear materials processing with a femtosecond laser oscillator,” Opt. Lett. 26, 1516–1518 (2001). [CrossRef]
9. A.M. Streltsov and N.F. Borrelli, “Fabrication and analysis of a directional coupler written in glass by nanojoule femtosecond laser pulses,” Opt. Lett. 26, 42–43 (2001). [CrossRef]
10. C.B. Schaffer, J.F. Garcia, and E. Mazur, “Bulk heating of transparent materials using a high-repetition rate femtosecond laser,” Appl. Phys. A. 76, 351–354 (2003). [CrossRef]
11. A.M. Streltsov and N.F. Borrelli, “Study of femtosecond-laser-written waveguides in glasses,” J.Opt. Soc. Am. B. 19, 2496–2504 (2002). [CrossRef]
12. Y. Sikorski, A.A. Said, P. Bado, R. Maynard, C. Florea, and K. Winick, “Optical waveguide amplifier in Nd-doped glass written with near-IR femtosecond laser pulses,” Electron. Lett. 36, 226–227 (2000). [CrossRef]
13. S. Taccheo, G. Della Valle, R. Osellame, G. Cerullo, N. Chiodo, P. Laporta, O. Svelto, A. Killi, U. Morgner, M. Lederer, and D. Kopf, “Er:Yb-doped waveguide laser fabricated by femtosecond laser pulses,” Opt. Lett. 29, 2626–2628 (2004). [CrossRef]
14. Y. Kondo, K Nouchi, T. Mitsuyu, M. Watanabe, P.G. Kazansky, and K. Hirao, “Fabrication of long-period fiber, gratings by focused irradiation of infrared femtosecond laser pulses,” Opt. Lett. 24, 646–648 (1999). [CrossRef]
15. L. Sudrie, M. Franco, B. Prade, and A. Mysyrowicz, “Writing of permanent birefringent microlayers in bulk fused silica with femtosecond laser pulses,” Opt. Comm. 171, 279–284 (1999). [CrossRef]
16. K. Minoshima, A.M. Kowalevicz, E.P. Ippen, and J.G. Fujimoto, “Fabrication of coupled mode photonic devices in glass by nonlinear femtosecond laser materials processing,” Opt. Express 10, 645–652 (2002), http://www.opticsexpress.org/abstract.cfm?URI=OPEX-10-15-645. [PubMed]
17. P.R. Herman, R.S Marjoribanks, and A. Oettl, Burst-ultrafast laser machining method, US Patent (6,552,301 B2)
18. M. Will, J. Burghoff, J. Limpert, T. Schreiber, H. Zellmer, S. Nolte, and A. Tunnermann, “Generation of photoinduced waveguides using a high repetition rate fiber CPA system,” Conference on Lasers and Electro Optics 2003, CWI6.
19. M. Will, J. Burghoff, J. Limpert, T. Schreiber, S. Nolte, and A. Tunnermann , “High speed fabrication of optical waveguides inside glasses using a high rep.-rate fiber CPA system,” in Photon Processing in Micorelectronics and Photonics III, P.R. Herman, J. Fieret, A. Pique, T. Okada, F.G. Bachmann, W. Hoving, K. Washio, X. Xu, J.J. Dubowski, D.B. Geohegan, and F. Trager, eds., Proc. SPIE5339, 168–173 (2004).
20. T. Fukuda, S. Ishikawa, T. Fujii, K. Sakuma, and H. Hosoya , “Improvement on asymmetry of low-loss waveguides written in pure silica glass by femtosecond laser pulses,” in Optical Fibers and Passive Components , S. Shen, S. Jian, K. Okamoto, and K. L. Walker, eds., Proc. SPIE 5279, 21–28 (2004).
21. S. Eaton, F. Yoshino, L. Shah, A. Arai, H. Zhang, S. Ho, and P. Herman, “Thermal heating effects in writing optical waveguides with a 0.1–5 MHz rate ultrafast fiber laser,” in Photonics West 2005, Proc. SPIE5713A-7 (2005).