LETTER pubs.acs.org/NanoLett
Generation of λ/12 Nanowires in Chalcogenide Glasses Elisa Nicoletti,† Douglas Bulla,‡ Barry Luther-Davies,‡ and Min Gu*,† †
Centre for Micro-Photonics and CUDOS, Faculty of Engineering and Industrial Sciences, Swinburne University of Technology, PO Box 218 Hawthorn, 3122 Victoria, Australia ‡ Laser Physics Centre and CUDOS, Research School of Physical Sciences and Engineering, Australian National University, Canberra Australian Capital Territory 0200, Australia
bS Supporting Information ABSTRACT: Nanowires have been widely studied and have gained a lot of interest in the past decade. Because of their high refractive index and high nonlinearity, chalcogenide glasses (ChGs) are a good candidate for the fabrication of photonic nanowires as such nanowaveguides provide the maximal confinement of light, enabling large enhancement of nonlinear interactions and group-velocity dispersion engineering. Here we report on the generation of λ/12 (∼68 nm) nanowires based on the theoretical and experimental study of the influence of the laser repetition rate on the direct laser fabrication in ChGs (λ = 800 nm). Through a numerical model of cumulative heating, the optimum conditions for high-resolution fabrication in As2S3 are found. Nanowires with dimensions down to ∼λ/12 are for the first time successfully fabricated in ChGs. We show that the generated nanowires can be stacked to form a three-dimensional woodpile photonic crystal with a pronounced stop gap. KEYWORDS: Chalcogenide glasses, direct laser writing, nanofabrication, nonlinear nanowires
B
ecause of the advancement in optics, photonics and optoelectronics, there is an increasing demand for faster functional devices in a more compact volume. Recent advances in nanoscale fabrication techniques have opened up new exciting opportunities for the miniaturization of photonic devices. In the past few years, direct laser writing (DLW) based on ultrashort laser pulses has attracted a lot of interest.1 7 DLW by multiphoton polymerization is a highly flexible technique that allows for the fabrication of arbitrary structures with subdiffraction limited resolution. In addition to traditional polymeric photoresists, DLW has been successfully applied to chalcogenide glasses (ChGs), such as arsenic trisulphide (As2S3). ChGs have recently been utilized as the platform for all-optical applications because of their high refractive index (high-n), high nonlinearity, low phonon energy, and high transparency in the near-to-mid-infrared (IR) spectral regime.8 14 Most experiments have been carried out with a low repetition rate (RR) (1 kHz) Ti:sapphire laser.1,2,7 A femtosecond laser beam can induce structural modification in a ChG film in metastable phase through a two-photon induced nonlinear process. The reason for the use of an amplified beam is the requirement for enough energy to activate two-photon polymerization (TPP) in the material. In the case of As2S3, an approximate energy level of 2 nJ is high enough to activate bond restructuring in the glass.1 Recent progress in the development of ultrafast lasers allows pulse energies up to 8 nJ at a RR in the MHz range. At a high RR, local heating effects become significant. Bond restructuring in a glass in metastable conditions can also be induced by the thermal effects. The heat accumulated in the system promotes the formation of a more stable crosslinked network. The use of a high RR laser beam can result in r 2011 American Chemical Society
heat accumulation over a number of pulses and generation of local thermal processes, hence the achievement of structural modifications within the glass.15 Though the high RR fabrication offers fast writing speeds and forms smooth structures, it is hard to reduce the structural size to a nanoscale due to the heat accumulation around the focal volume. Here, we report on the first fabrication of nanowires of a dimension of λ/12 in As2S3 films by DLW through the RR optimization of an ultrafast pulsed laser beam. The laser RR was found to have a significant impact on the structural modifications of the ChGs that in turn affects the nanofabrication quality inside such glasses. To understand the effect of the laser RR on the fabrication in As2S3 and push the threshold limit to a smaller fraction of the laser fabrication wavelength, we need a better understanding of the processes that lead to the structural changes in the material. As such, a variable RR femtosecond laser was applied for the first time to uncover the contributions of thermal diffusion and heat accumulation in nanofabrication in As2S3 films. To verify and study the dynamics of the cumulative heating, a finite-difference thermal diffusion model was applied to simulate the heat dissipation and dynamics at different fabrication conditions16,17 To obtain ideal fabrication conditions, the DLW method should take advantage of the thermal effects while avoiding the overheating of the material. It is necessary, then, to find out at which range of RRs the thermal effects start to occur. As such, we calculated the effective cooling time (τc = d2/D, see Supporting Received: June 27, 2011 Revised: August 15, 2011 Published: August 29, 2011 4218
dx.doi.org/10.1021/nl202173t | Nano Lett. 2011, 11, 4218–4221
Nano Letters
LETTER
Figure 1. Sketch of the two-step fabrication process of 2D nanorods in a ChG film. The as-deposited As2S3 film is locally exposed to a femtosecond laser beam (DLW). The material undergoes structural changes. The unexposed glass is removed by means of wet etching revealing the 2D ChG nanowires.
Information) for the heat to dissipate out of a volume of diameter equal to the focal spot. The τc value for an As2S3 film is equal to 0.65 2.19 μs. The critical RR for the accumulation of heating is the inverse of τc. Thus, the range of RRs where the material starts to experience the thermal effect is between 457 and 1538 kHz. A RR of 1.6 MHz can then ensure the presence of a thermal effect, which avoids the overheating of the glass (see Supporting Information). To investigate if the accumulation of heat occurs during the fabrication and the influence of the thermal effect on the fabrication conditions, we compared the thermal parameters of the material exposed to three different RRs. We calculated the thermal diffusion length (LD), the distance traveled by the sample between two consecutive laser pulses (h), and the number of laser pulses overlapping in the diameter of the focal spot (Nd) and in the region of thermal diffusion (NLD). We studied and compared two extreme conditions of 1 kHz and 82 MHz, as well as the intermediate value of 1.6 MHz (see Supporting Information Table 2). At the 1.6 MHz RR, LD is around 7 times smaller than that at 82 MHz. This means that a thicker portion of the material is entitled to attenuate the laser beam heating. Nd at 1.6 MHz is 1000 times bigger than that at 1 kHz but 100 times smaller than that at 82 MHz. h is much smaller as well, compared to that at the low RR of 1 kHz. Even at a high scanning speed, it is still in the order of 0.1 nm. From the comparison of the thermal parameters at different RRs, we can predict that a high RR allows for the fabrication of smooth and homogeneous nanorods at a higher scanning speed. In particular DLW at 1.6 MHz is expected to offer the optimum fabrication conditions in As2S3 films. In order to confirm the theoretical predictions and investigate the influence of the 3 RRs (1 kHz, 1.6 MHz, and 82 MHz) on the DLW fabrication in ChG films, we fabricated two-dimensional (2D) lines at different laser powers and different scanning speeds. Femtosecond laser pulses operating at a wavelength of 800 nm were focused into an As2S3 film 16 20 μm thick with an oil immersion objective (Olympus, numerical aperture 1.4, 100). For the lower RR (1 kHz), we maintained the same fabrication setup previously used.1 The laser source was an amplified Ti:sapphire laser system (Spitfire, Spectra Physics) with a pulse duration of 110 fs and a RR of 1 kHz. For the higher RRs (1.6 and 82 MHz), an ultrashort pulsed laser (Spectra Physics, Tsunami) with a 10 W pump laser was used. The pulse width and the RR of the laser were 80 fs and 82 MHz, respectively. A Conoptics pulse picker was used to tune the RR of the laser beam as desired.
Figure 2. Experimental plots of the lateral rod dimension versus the laser power at (a) 1 kHz RR and (b) 1.6 MHz RR and different fabrication scanning speeds (vs). (Similar behavior was found for the fabrication at 82 MHz RR.) (A,B) SEM images of rods fabricated at different scanning speeds of 50, 100, 150, and 200 μm/sec; image A for 1 kHz with a laser power of 4 μW and image B for 1.6 MHz with a laser power of 500 μW.
The unexposed glass was removed using a solution of diisopentylamine (Alfa Aesar) and DMSO (Sigma-Aldrich) (2 3 mol %) for 5 10 min (Figure 1). Suspended lines were fabricated at four different scanning speeds (50, 100, 150, and 200 μm/sec) with powers ranging between 1.5 and 12 μW for 1 kHz RR and between 300 and 900 μW for 1.6 MHz RR (see Figure 2A,B). Figure 2 shows the plots of the lateral line width as a function of the laser power at both the 1 kHz and 1.6 MHz RRs. For 1 kHz (Figure 2a) the minimum feature size was found to be around 150 nm, which is comparable to the fabrication in polymer.3 At 1.6 MHz, the threshold thickness limit is lower with a minimum feature of 68 nm (Figure 2b). For the first time, the line widths at a resolution of ∼ λ/12 were fabricated in As2S3 glass. Figure 3 shows the SEM image of three lines fabricated at 1kHz (left), 1.6 MHz (center), and 82 MHz (right). The rods were fabricated at a scanning speed of 150 μm/sec and the threshold power of 2 μW (for 1 kHz), 450 μW (for 1.6 MHz), and 1.1 mW (for 82 MHz). The surface of the rod fabricated at 1 kHz is smooth, although some undulations are observed. The surface quality appears noticeably improved for the rod fabricated at higher RRs. 4219
dx.doi.org/10.1021/nl202173t |Nano Lett. 2011, 11, 4218–4221
Nano Letters
Figure 3. SEM images of rods fabricated (left) at 1 kHz, a laser power of 2 μW and a scanning speed of 150 μm/sec, (center) at 1.6 MHz, a laser power of 450 μW and a scanning speed of 150 μm/sec, and (right) at 82 MHz, a laser power of 1.1 mW and a scanning speed of 150 μm/sec.
LETTER
and a scanning speed of 100 μm/sec. A detailed view of the PhC is presented in Figure 4b, while Figure 4c shows the side view of the woodpile structure. The FTIR spectrum presented in Figure 4d shows a 42% suppression of the transmission at a wavelength of 2.9 μm. As expected, the result is comparable to the one obtained at a higher RR fabrication [1]. The mechanical properties of the written structures are preserved, proving the possible extension of the DLW fabrication at 1.6 MHz to 3D microstructures. In conclusion, we have investigated the effect of the RR on the DLW fabrication in ChG films. It has been found that the heat accumulated over the large number of pulses of a high RR laser beam influences the response of the material at the focal spot. The calculations based on a numerical model of cumulative heating indicates the RR of 1.6 MHz as the ideal case to use for high resolution fabrication as it facilitates the benefits of the heating effect while avoiding the overheating of the glass around the focal spot. Under this condition, it is possible to obtain nanorods with a lateral dimension of 68 ( 14 nm (∼λ/12) with strong mechanical properties that allows us to generate a 3D woodpile PhC with a pronounced stop gap in the near-infrared region. This is a key achievement for the ChG devices as the confinement of the light in a small volume contributes to the enhancement of the nonlinear properties of the material.
’ ASSOCIATED CONTENT
bS
Supporting Information. Supporting Information includes details of the numerical model of cumulative heating together with a table listing the thermal properties of As2S3 and a table containing the calculated thermal parameters for fabrication in As2S3 using a femtosecond laser beam at different repetition rates. This material is available free of charge via the Internet at http://pubs.acs.org. Figure 4. (a) SEM image of an As2S3 woodpile PhC fabricated at 1.6 RR. The lattice parameters are dx = 1 μm and dz = 0.6 μm. (b) Detailed view and (c) side view of the woodpile structure. (d) FTIR transmission spectrum of the ChG PhC showing a 42% suppression of the transmission at a wavelength of 2.9 μm.
’ AUTHOR INFORMATION
Increasing the RR to 82 MHz, rods 80 nm (∼λ/10) thick were obtained. As expected, the threshold fabrication limit is pushed to even smaller features at 1.6 MHz RR. The lateral rod dimension is only 68 ( 14 nm and the nanowires are smooth, homogeneous, and free from structural distortions. The energy per pulse (Ep) used for DLW at 1.6 MHz is 7 8 times lower than that at 1 kHz. However, at a higher RR the TPA process is assisted by heat accumulation over the much larger number of laser pulses per spot and per thermal diffusion length. The combination of TPA processes and thermal effects, followed by selective etching, allows for the fabrication of smoother and smaller rods. This achievement is an important step toward the realization of photonic devices. Such nanowires confine and guide light at the submicrometer-scale, enabling large enhancements of nonlinear interactions. The effective nonlinearity of ChG photonic nanowires is more than 3 orders of magnitude larger than silica glass nanowires,18 which promises the possibility to use more readily achievable power levels for nonlinear optical applications and photonics integration devices. Using the 1.6 MHz RR fabrication condition, we successfully fabricated 3D woodpile PhCs and a stop gap was observed. Figure 4a shows a woodpile structure with lattice parameters dx = 1 μm and dz = 0.6 μm fabricated at a laser power of 550 μW
’ ACKNOWLEDGMENT This work was produced with the assistance of the Australian Research Council under the ARC Centers of Excellence program. The authors acknowledge the ARC Centre of Excellence CUDOS (Centre for Ultrahigh-bandwidth Devices for Optical Systems) for the support. The authors would like to acknowledge Benjamin P. Cumming, from Swinburne University (CMP), for his technical support and Dr. Airan Rodenas, from Universidad Autonoma de Madrid, for the valuable discussions.
Corresponding Author
*E-mail:
[email protected].
’ REFERENCES (1) Nicoletti, E.; Zhou, G.; Jia, B.; Ventura, M. J.; Bulla, D.; LutherDavies, B.; Gu, M. Opt. Lett. 2008, 33 (20), 2311–2313. (2) Wong, S.; Deubel, M.; Perez-Willard, F.; John, S.; Ozin, G. A.; Wegener, M.; Von Freymann, G. Adv. Mater. 2006, 18 (3), 265–269. (3) Serbin, J.; Gu, M. Adv. Mater. 2006, 18 (2), 221–224. (4) Straub, M.; Ventura, M.; Gu, M. Phys. Rev. Lett. 2003, 91, 043901. (5) Straub, M.; Gu, M. Opt. Lett. 2002, 27 (20), 1824–1826. (6) Sun, H. B.; Kawakami, T.; Xu, Y.; Ye, J. Y.; Matuso, S.; Misawa, H.; Miwa, M.; Kaneko, R. Opt. Lett. 2000, 25 (15), 1110–1112. (7) Nicoletti, E.; Bulla, D.; Luther-Davies, B.; Gu, M. Opt. Lett. 2011, 36 (12), 2248–2250. 4220
dx.doi.org/10.1021/nl202173t |Nano Lett. 2011, 11, 4218–4221
Nano Letters
LETTER
(8) Ta’eed, V. G.; Baker, N. J.; Fu, L.; Finsterbusch, K.; Lamont, M. R. E.; Moss, D. J.; Nguyen, H. C.; Eggleton, B. J.; Choi, D. Y.; Madden, S.; Luther-Davies, B. Opt. Express 2007, 15 (15), 9205–9221. (9) Ta’eed, V. G.; Fu, L.; Pelusi, M.; Rochette, M.; Littler, I. C. M.; Moss, D. J.; Eggleton, B. J. Opt. Express 2006, 14 (22), 10371–10376. (10) Ta’eed, V. G.; Shokooh-Saremi, M.; Fu, L.; Moss, D. J.; Rochette, M.; Littler, I. C. M.; Eggleton, B. J.; Ruan, Y.; Luther-Davies, B. Opt. Lett. 2005, 30 (21), 2900–2902. (11) Asobe, M.; Ohara, T.; Yokohama, I.; Kaino, T. Electron. Lett. 1996, 32 (15), 1396–1397. € Wise, F. W.; Sanghera, J. S.; (12) Harbold, J. M.; Ilday, F. O.; Nguyen, V. Q.; Shaw, L. B.; Aggarwal, I. D. Opt. Lett. 2002, 27 (2), 119–121. (13) Madden, S. J.; Choi, D. Y.; Bulla, D. A.; Rode, A. V.; LutherDavies, B.; Ta’eed, V. G.; Pelusi, M. D.; Eggleton, B. J. Opt. Express 2007, 15 (22), 14414–14421. (14) Galili, M.; Xu, J.; Mulvad, H. C. H.; Oxenløwe, L. K.; Clausen, A. T.; Jeppesen, P.; Luther-Davies, B.; Madden, S.; Rode, A.; Choi, D. Y.; Pelusi, M.; Luan, F.; Eggleton, B. J. Opt. Express 2009, 17 (4), 2182–2187. (15) Lesik, M. A.; Averina, A. V.; Shimko, A. A.; Man’shina, A. A. J. Opt. Technol. 2009, 76 (1), 48–50. (16) Nejadmalayeri, A. H.; Herman, P. R. Opt. Express 2007, 15 (17), 10842–10854. (17) Jaluria, Y.; Atluri, S. N. Comput. Mech. 1994, 14 (5), 385–386. (18) Freeman, D.; Madden, S.; Luther-Davies, B. Opt. Express 2005, 13 (8), 3079–3086.
4221
dx.doi.org/10.1021/nl202173t |Nano Lett. 2011, 11, 4218–4221
