We describe the fabrication of freestanding optical fibers in a glass chip by femtosecond laser three-dimensional (3D) micromachining. The process has mainly four steps: (1) femtosecond laser scanning of the areas surrounding the fibers; (2) postannealing of the sample for the modification of the exposed areas; (3) chemical etching of the sample for the selective removal of the modified areas; and (4) second postannealing for smoothening the surfaces of the fibers. The measured optical loss of the fibers is approximately 0.7dB/cm. Integrating the freestanding fibers with 3D micromirrors is also demonstrated, enabling functions such as light folding or splitting in the glass chips. Furthermore, we demonstrate that the freestanding fibers can be incorporated into a microfluidic circuit for on-chip biophotonic applications.
©2005 Optical Society of America
CorrectionsYa Cheng, Koji Sugioka, and Katsumi Midorikawa, "Freestanding optical fibers fabricated in a glass chip by femtosecond laser micromachining for lab-on-a-chip application: erratum," Opt. Express 14, 11910-11910 (2006)
In recent years, we have witnessed the rapid prototyping of three-dimensional (3D) hollow structures buried in glasses using femtosecond (fs) laser micromachining [1–7]. Currently, both microfluidic structures and microoptical structures can be easily fabricated in a commercially available photosensitive glass (Foturan) [3–6], and these two structures have been successfully integrated into a single glass chip to construct functional devices, such as 3D microfluidic dye lasers embedded in a glass which is a very attractive tool for optical analysis in lab-on-a-chip devices . Naturally, the next step is the fabrication of an “all-in-one” biosensor or a micro-total analysis system (μ-TAS) device that incorporates a microfluidic circuitry with integrated microoptical networks. In these integrated lab-on-a-chip devices, first, a microfluidic laser beam is coupled into a waveguide, and then, the waveguide brings light to a liquid sample in a microfluidic chamber to realize optical absorption spectroscopy or fluorescence detection. Lastly, a second waveguide, which is connected to an optical spectrometer or a photodetector, collects the light transmitted from or emitted by the liquid sample for information acquisition.
Although optical waveguides buried in a Foturan glass have been achieved by refractive index modification using fs laser microprocessing [8–9], two main issues hamper the use of these waveguides for the above-mentioned biosensor application. First, the refractive index modification of the Foturan glass is realized by the selective precipitation of silver nanoparticles. The silver nanoparticles not only change the refractive index, but also cause a significant optical loss due to light absorption, particularly in the visible range . Second, the fabrication techniques used to produce the buried waveguides and the internal hollow structures are incompatible; thus the waveguides must be written after the formation of the hollow structures. This two-step approach would require the highly accurate repositioning of the etched sample on the stage for the waveguide writing process, thereby increasing the number of manufacturing steps and the cost. In this paper, we show that these two issues can be resolved by fabricating freestanding optical fibers on a glass chip. We demonstrate that these fibers can be integrated with micromirrors, which can be fabricated using the same process , to realize functions such as light folding or splitting on microchips. Finally, we show that the freestanding fibers can also be incorporated into microfluidic circuit to enable analyses of liquid samples by optical means, such as photoabsorption spectroscopy or fluorescence detection.
The glass used in this work is a commercially available Foturan glass from Schott Glass Corporation, which is composed of lithium aluminosilicate doped with trace amounts of silver, cerium, and antimony [3, 10]. The experiments were carried out at a commercial fs laser workstation, as described elsewhere . The repetition rate, wavelength and pulse width of the fs laser were 1kHz, 775nm and 145±5fs, respectively. To fabricate the fiber, first we scanned the areas surrounding the fibers in the glass using the fs laser beam tightly focused by an objective lens with a numerical aperture (NA) of 0.46. The focal spot size is estimated to be approximately 1μm (FWHM). Based on the previous investigation on the critical irradiation dose needed for the fs laser modification of Foturan glass , a pulse energy of 525nJ/pulse (measured after the objective lens) and a scanning speed of 200μm/s were used throughout this experiment. The exposed areas were created by horizontally scanning the laser beam to form many parallel lines with a spacing of 2.5μm for the horizontal areas and a spacing of 10μm for the vertical areas. We chose larger spacing in the vertical direction because the focused laser beam spot is elongated in axial direction . The total scanning time for fabricating an 8mm-long freestanding fiber is approximately 55min. Using a fs laser working at higher repetition rate can significantly shorten the scanning time, because a higher scanning speed can be employed.
After the laser scanning process, we baked the exposed sample at 500°C for 1h and then at 605°C for another 1h in a programmable furnace for the modification of the exposed areas. After this step, crystalline phase of lithium metasilicate, whose etching rate in a diluted solution of 5% hydrofluoric (HF) acid could be 50 times higher than that of the glass matrix, were formed in the exposed areas. Since the size of the lithium metasilicate crystallites ranges from 3μm to 5μm, the fabrication resolution of this processing is limited to a few microns, despite that the fs laser beam could be focused into a sub-micron spot. Next, we etched the sample in a solution of 5% HF acid diluted in water in an ultrasonic bath for one hour to remove modified areas. Lastly, we baked the etched sample again at 570°C for 5h for further smoothening the surfaces of the fibers. More details including the mechanism of this process can be found elsewhere [10–12].
First, we fabricated a structure composed of a freestanding fiber integrated with a 45° micromirror at the entrance of the fiber in the glass, as illustrated in Fig. 1(a). The optical path of the coupling scheme is indicated in Fig. 1(a) by arrows. The 45° micromirror allows us to couple the light into the fiber from the side of the sample. Figure 1(b) shows the optical micrograph of the micromirror and entrance of the fiber. The inset of Fig. 1(b) (upper right corner) shows the cross-sectional shape of the fabricated fiber, which has dimensions of approximately 100μm×80μm (width×height). Coupling light into the fiber was obtained by focusing a He-Ne laser beam at the micromirror using the objective lens with NA of 0.46, as shown in Fig. 1(c). The guided light was clearly observed at exit of the fiber. The total length of the fiber in Fig. 1(c) was 8mm, which is sufficient for many microchip applications. The fabrication of a longer fiber is not a problem, but it requires a longer fabrication time.
We then fabricated five freestanding fibers of different lengths of 2mm, 3mm, 4mm, 6mm, and 8mm to evaluate the optical loss of these fibers. In this case, only the freestanding fibers were fabricated on the glass chip, without the integration of the micromirror. The He-Ne laser beam was directly focused onto the entrance of the fiber, and the power of the exiting light spot was measured using an optical power meter. For each measurement, we carefully aligned the fiber until maximum output power was achieved. Moreover, we carefully ensured that the entire focal spot of the He-Ne laser beam was positioned within the area of the entrance of the fiber. We found that when a good alignment was obtained, we could see a far-field pattern containing many interference fringes on a receiving screen placed near the exit of the fiber, due to the fact that the freestanding fiber is a multimode fiber. Figure 2 shows the transmittance of the fiber as a function of fiber length. The measured optical loss of the fiber is the sum of the coupling loss and the propagation loss. Here, we assume that the coupling loss is the same for all fibers, and the propagation loss is proportional to the fiber length. Therefore, by performing the linear least-square fitting of the data in Fig. 2, the calculated propagation loss and the coupling loss are approximately 0.7dB/cm and 1.8dB, respectively. Since the Foturan glass does not exhibit absorption at the wavelength of the He-Ne laser, we conclude that the propagation loss is caused by the surface scattering of light on the fiber sidewall.
As shown in Fig. 1, the freestanding fibers can be easily integrated with the micromirrors to redirect the light on the glass chip. To further demonstrate this capability, we fabricated a structure containing three fibers and two 45° micromirrors to create a folded light path, as shown in Fig. 3(a). Figure 3(b) shows an optical micrograph of the folding point on the glass chip. Two orthogonal fibers and one micromirror were combined at this point. Figure 3(c) shows a digital-camera-captured side view image of the coupling of a He-Ne laser beam into the structure. It is clear that the light focused by the objective lens was first coupled into this first fiber with a length of 2mm, and then, it was guided to the first micromirror. The first micromirror then reflected the light downward, so that light could be coupled into the second fiber with a length of 2mm. Lastly, the second micromirror redirected the light into the third fiber with a length of 5mm. The guided light spot is indicated by an arrow in Fig. 3(c); also indicated by arrows in Fig. 3(c) are scattering light spots at the micromirrors and an image of the scattering light spot on the reflective rear surface of the glass sample. Figure 3(d) shows a front view of the exit surface of the structure, showing that the guided light spot was shifted downward by 2mm from the coupling position.
Furthermore, we fabricated a more complex optical circuit containing five fibers, two micromirrors, and a microbeam splitter, as shown in Fig. 4(a). The microbeam splitter, which is composed of two micromirrors at right angles, was positioned at the meeting point of the optical axes of the three fibers, as shown in Fig. 4(b). When a guided light came from the first fiber on the left-hand side, half of the light was reflected upward into the upper fiber, whereas the other half was reflected downward into the lower fiber. Figure 4(c) shows the front view image of the surface of exit. The upper and lower light spots in Fig. 4(c) are the two light beams exiting from the two output arms of the fabricated optical circuit, and the middle light spot is the scattering light at the microbeam splitter. This beam splitter is particularly useful for accurate chemical or biological analysis, because one arm can be used for measuring unknown liquid samples, whereas another arm can be used as a reference to monitor the fluctuation of the light source.
The freestanding fibers can easily be incorporated into a microfluidic circuit for on-chip biophotonic applications. For this purpose, we fabricated a structure that comprises two series of freestanding fibers intercepted by a microwell fabricated on the glass chip, as illustrated in Fig. 5(a). Figure 5(b) is the optical micrograph of a part of the fabricated microstructure. To demonstrate that the exiting light from the first fiber can still be effectively coupled into the second fiber, we focused the He-Ne laser beam into the entrance facet of the first fiber by the 20× objective lens. As shown in Fig. 5(c), both scattering light at the microwell and the guided light at the end of the second fiber can clearly be seen. We then measured the coupling loss between the two fibers intercepted by the microwell which was approximately 1dB. To evaluate the coupling optical loss, we fabricated another single freestanding fiber (a pure fiber of 4.5mm-length which is the same length from the entrance of the first fiber to the end of the second fiber) on the same sample using the same processing parameters. The difference between the optical losses of the two structures should be attributed to the additionally scattered light at the microwell between the two fibers. In our opinion, two factors responsible for the low optical loss between the two fibers are: (1) the relatively large diameter of the fiber allows propagation of light beam of large mode area in the fiber, thereby reducing the divergence angle of the exiting light; and (2) the internal wall of the microwell between the two fibers is smooth and fabricated nearly perpendicular to the fiber axis, causing little distortion of the beam during exiting from the first fiber and then entering into the second fiber.
4. Discussions and conclusion
The freestanding fibers demonstrated here are all multimode fibers with diameters approximately 100μm. Thinner fibers can also be fabricated, but the length must be reduced. Actually, we have already fabricated a fiber with a diameter of 20μm and a length of 500μm. In our experiment, we found that a long thin fiber was first slightly bent after the wet etching step, perhaps because the thin fiber was not sufficiently strong to resist the force generated by the surface tension of the HF solution in the ultrasonic bath. In the subsequent postannealing step, the bent part of the fiber then collapsed, and finally fused to the wall surrounding the fiber. This issue can be resolved by fabricating several supporting structures beneath the freestanding fiber to enhance its strength. The price paid on these additional supporting structures is increased optical loss. For instance, each of the freestanding fibers described in this paper has two supporting structures of 100μm-length at the entrance and the exit of the fiber. The evaluated optical loss of each supporting structure is less than 0.12dB. To fold the light beam in the chip, the use of the 45° micromirrors is not the only way. A bend incorporated in the freestanding fiber would also enable the same function with less bending loss. However, since we are now using an XYZ stage which can only produce linear structures, the bend structure cannot be fabricated using the current 3D fs laser microfabrication system.
To summarize, we demonstrated the direct fabrication of freestanding optical fibers in glass chips by fs laser 3D micromachining. Although currently a variety of technologies can routinely produce fibers with performance significantly higher than the freestanding fibers demonstrated here, few of them have the ability to directly assemble the fabricated fibers into a chip to form a compact 3D optical circuit. Incorporating the optical circuit with microfluidic structures poses further challenge. The technique described here provides a unique way to resolve these issues, enabling mass production of cheap biosensors based on optical means, like photoabsorption spectroscopy or fluorescence detection.
We thank K. Furusawa of RIKEN for helpful discussion, and K. Shihoyama of HOYA CANDEO OPTRONICS CORPORATION for providing the femtosecond laser workstation.
References and links
1. Y. Kondo, J. Qiu, T. Mitsuyu, K. Hirao, and T. Yoko, “Three-Dimensional Microdrilling of Glass by Multiphoton Process and Chemical Etching,” Jpn. J. Appl. Phys. 38, L1146–L1148 (1999). [CrossRef]
2. A. Marcinkevicius, S. Juodkazis, M. Watanabe, M. Miwa, S. Matsuo, H. Misawa, and J. Nishii, “Femtosecond laser-assisted three-dimensional microfabrication in silica,” Opt. Lett. 26, 277–279 (2001). [CrossRef]
3. M. Masuda, K. Sugioka, Y. Cheng, N. Aoki, M. Kawachi, K. Shihoyama, K. Toyoda, H. Helvajian, and K. Midorikawa, “3-D microstructuring inside photosensitive glass by femtosecond laser excitation,” Appl. Phys. A76, 857–860 (2003).
4. Y. Cheng, K. Sugioka, K. Midorikawa, M. Masuda, K. Toyoda, M. Kawachi, and K. Shihoyama, “Control of the cross-sectional shape of a hollow microchannel embedded in photostructurable glass by use of a femtosecond laser,” Opt. Lett. 28, 55–57 (2003). [CrossRef] [PubMed]
5. Y. Cheng, K. Sugioka, K. Midorikawa, M. Masuda, K. Toyoda, M. Kawachi, and K. Shihoyama, “Three-dimensional microoptical components embedded in photosensitive glass by a femtosecond laser,” Opt. Lett. 28, 1144–1146 (2003). [CrossRef] [PubMed]
6. K. Sugioka, M. Masuda, T. Hongo, Y. Cheng, K. Shihoyama, and K. Midorikawa, “Three-dimensional microfluidic structure embedded in photostructurable glass by femtosecond laser for lab-on-chip application,” Appl. Phys. A79, 815–817 (2004).
8. Y. Cheng, K. Sugioka, M. Masuda, K. Shihoyama, K. Toyoda, and K. Midorikawa, Opt. Express 11, 1809 (2003); http://www.opticsexpress.org/abstract.cfm?URI=OPEX-11-15-1809. [CrossRef] [PubMed]
9. S. Ho, P. R. Herman, Y. Cheng, K. Sugioka, and K. Midorikawa, “Direct ultrafast laser writing of buried waveguides in Foturan glass,” in Conference on Lasers and Electro-Optics (CLEO), Vol. 97 of OSA Trends in Optics and Photonics Series (Optical Society of America, Washington, D.C., 2004), paper CThD6.
10. H. Helvajian, P. D. Fuqua, W. W. Hansen, and S. Janson, “Laser microprocessing for nanosatellite microthruster applications,” RIKEN Review 32, 57–63 (2001), http://www.riken.go.jp/lab-www/library/publication/review/pdf/No_32/32_057.pdf.
11. M. Masuda, K. Sugioka, Y. Cheng, N. Aoki, M. Kawachi, K. Shihoyama, K. Toyoda, and K. Midorikawa, “3D microfabrication in photosensitive glass by femtosecond laser,” SPIE Proc. 4830, 576–580 (2002). [CrossRef]
12. T. Hongo, K. Sugioka, H. Niino, Y. Cheng, M. Masuda, J. Miyamoto, H. Takai, and K. Midorikawa, “Three-dimensional micromachining of glass using femtosecond laser for lab- a-chip device manufacture,” J. Appl. Phys. 97, 063517 (2005) [CrossRef]