1. Introduction

Polymers are very attractive materials for use in biophotonic microchips and micro total analysis systems, due to their high flexibility, high chemical stability, and light weight. The commercially available polymers, poly(methyl methacrylate) (PMMA) and poly(dimethylsiloxane) (PDMS), are two of the most commonly used polymers in various applications, including microfluidic devices. These two polymers exhibit excellent transparency from the near-UV region to the near-IR region. Optical waveguides have been fabricated in these polymers by femtosecond laser irradiation [1–3]. However, these optical waveguides transmit light only with wavelengths greater than about 532 nm. This is also true of waveguides fabricated in UV-transparent fused silica, which is another popular material for microbiochips [4]. To the best of our knowledge, there have been no reports of UV waveguides fabricated in materials by femtosecond laser irradiation. This is despite the high demand by biologists for microchips that transmit UV light for analyzing UV-induced biological phenomena, such as UV fluorescence of DNA in electrophoresis systems [5] and UV damage of DNA chains in living cells [6]. One promising material for these purposes is the amorphous fluoropolymer, CYTOP, developed by Asahi Glass Co. Ltd [7,8]. CYTOP has a transmittance of over 90% for wavelengths in the range 200 nm to 2 μm. CYTOP also has a high chemical resistance to acids, alkalis, and organic solvents. The greatest advantage of CYTOP over PMMA, PDMS, and fused silica is that its refractive index is very similar to that of water. This is indispensable for microscopic observations of biocells since cells are always immersed in water during observations so that the refractive index match between the substrate and water minimizes image distortion and reflections at the interface [9].

Our group has used F₂ laser ablation to fabricate electrophoresis microchips from CYTOP, which can be used for analyzing separated DNAs with different base pairs using 254 or 312 nm wavelength irradiation [10]. In addition, we have modified the CYTOP surface by F₂ laser irradiation, making it hydrophobic so that culture adherent cells can be selectively adhered to the CYTOP microchips. As a result of CYTOP and water having similar refractive indices, the fabricated microchips could be used to obtain clear 3D microscopic images of HeLa cells in a culture medium [9]. Our next step will be to modify these microchips to enable them to analyze UV-induced biological phenomena. For this purpose, in this paper, we demonstrate fabrication of UV waveguides in CYTOP by femtosecond laser direct writing for the first time and we investigate the modification mechanism and optical properties of the written waveguides.

2. Experiments

To fabricate the optical waveguides in this study, the refractive index of CYTOP was modified using a commercial femtosecond laser system: a 800-mW regenerative amplified laser with an emission wavelength of 775 nm, a pulse duration of 150 fs, and a repetition rate of 1 kHz. The detailed configuration has been described elsewhere [11]. The pulse energy was adjusted using a polarizer and a neutral density filter. The 6-mm beam width of the output laser was reduced to 3 mm by passing it through an aperture to improve the beam quality. The beam was focused using a × 20 microscope objective with a numerical aperture (NA) of 0.46. A 0.5-mm-wide slit was placed before the objective lens to improve the cross-sectional shape of the optical waveguide [12]. Optical waveguides were fabricated by translating the CYTOP samples perpendicular to the laser beam axis using a computer-controlled xyz-stage with a resolution of 1 µm. The laser scanning speed was varied in the range from 0.1 to 3.0 mm/s and the laser energy was varied between 3 and 11 µJ/pulse before the slit. After the slit, this laser energy range decreased to 0.65 to 2.5 µJ/pulse because only 22% of the light passed through the slit. The whole fabrication process was recorded using a CCD camera and displayed on a PC monitor. To make 500-µm-thick CYTOP films, commercial CYTOP solution was placed in a programmable furnace and the temperature of the furnace was increased to 230 °C in 30 hrs. The film was then cut and mechanochemically polished along the cut edges to achieve an optical finish. The mode profile, the induced refractive index change, and the propagation loss were measured to optically characterize the waveguides. The mechanism for the refractive index increase was investigated by X-ray photoelectron spectroscopy (XPS) measurements.

To optimize the conditions for fabricating waveguides inside the CYTOP samples, we first conducted systematic experiments for various writing conditions. Figure 1 shows microscope images of the cross-sectional profiles of waveguides fabricated under different writing conditions. Figure 1 shows that optical waveguides with no laser-induced damage were written at pulse energies before the objective lens in the range 0.8 to 2.0 µJ/pulse at a laser scanning speed of 1.5 mm/s and at a pulse energy of 0.9 µJ/pulse for laser scanning speeds between 1.5 and 3.0 mm/s. The cross-sectional profiles of the written waveguides are longitudinally elliptical in shape; this could be rectified by employing a narrower slit. A refractive index increase could not be induced below a pulse energy of 0.8 µJ/pulse or above a scanning speed of 3.0 mm/s, while above a pulse energy of 2.0 µJ/pulse or below a scanning speed of 1.5 mm/s, laser-induced damage may be generated inside the CYTOP sample (see Fig. 1). Based on these results, a laser energy of 0.9 µJ/pulse and a scanning speed of 1.5 mm/s were selected for the subsequent experiments.

 

Fig. 1 Cross-sectional optical microscope images of waveguides written in CYTOP by femtosecond laser irradiation at (a) various laser pulse energies with a laser scanning speed of 1500 µm/s and (b) various laser scanning speeds with a laser pulse energy of 0.9 µJ/pulse.

Download Full Size | PDF

To measure the profiles of guided light beams, a 632.8-nm-wavelength He-Ne laser beam was coupled to one facet of the written waveguide by a × 50 objective lens (NA: 0.55). The near-field profile and the intensity distribution profile were obtained by imaging the output beam at another facet of the waveguide (see Fig. 2 ).

 

Fig. 2 (a) Near-field pattern of He-Ne laser beam transmitted by a waveguide and (b) its intensity distribution.

Download Full Size | PDF

The intensity profile in Fig. 2 demonstrates that the waveguide essentially functions as a single-mode waveguide at a wavelength of 632.8 nm. We confirmed that it also functions as a single-mode waveguide even at 355, and 266 nm wavelengths. The full-width at half-maximum of the intensity profile is approximately 3.4 µm in the y-direction and 6.5 µm in the z-direction. The refractive index change (Δn) induced by laser irradiation was estimated by measuring the NA of the 5-mm-long waveguide using the method of Homoelle et al [13]. The 632.8-nm-wavelength He-Ne laser beam was coupled to the waveguide and the NA of the cone of the guided light that exited from the waveguide was measured in the far field. The NA of the waveguide was measured to be 0.07. Assuming a step index change, Δn was estimated to be 1.7 × 10⁻³.

The propagation loss at various wavelengths was evaluated by fabricating waveguides with five different lengths. Much care was taken to assure that identical coupling conditions were achieved for all five waveguides. 632.8, 355, and 266 nm wavelength CW lasers were coupled to the waveguide using an objective lens ( × 50, NA: 0.55 for 632.8 nm; × 15, NA: 0.32 for 355 nm and 266 nm) to determine the optical losses. Several measurements were taken at each waveguide and the average of these measurements was adopted. Figure 3 shows the throughput data for the optical loss as a function of the waveguide length. The data for each wavelength were fitted with a linear equation. The gradient of the linear fit represents the propagation loss and the y-axis intercept gives the coupling loss.

 

Fig. 3 Linear fit of optical loss data for different length waveguides at wavelengths of (a) 632.8, (b) 355, and (c) 266 nm.

Download Full Size | PDF

From Fig. 3, the coupling losses are 0.3 dB for 632.8 nm, 1.1 dB for 355 nm, and 1.7 dB for 266 nm. The coupling losses are caused by either a mismatch between the size of the beam and that of the waveguide or misalignment between the waveguide and the focused CW laser beam. The propagation losses are estimated to be 0.49, 0.77, and 0.91 dB/cm for wavelengths of 632.8, 355, and 266 nm. For biochip applications, the propagation losses obtained for the two UV wavelengths are acceptable due to the short waveguide length in biochips.

3. Discussion

When focused femtosecond laser pulses with low energies are introduced to PMMA or PTFEP, partial bond breaking in the polymer is induced; this causes the volume of the polymer to contract, which increases its refractive index [14]. A similar mechanism is thought to be responsible for the observed increase in the refractive index of CYTOP. XPS measurements were conducted to characterize changes in the chemical bonding states in irradiated regions of CYTOP. The XPS analysis sample was prepared by line-by-line femtosecond laser irradiation on the CYTOP surface at a pitch of 10 µm for the same experimental conditions as those used for fabricating the optical waveguides; unirradiated CYTOP was used for the control. Figure 4(a) shows the chemical bonding structure of unirradiated CYTOP and Figs. 4(b) and (c) show XPS spectra of C1s for the CYTOP surface before and after femtosecond laser irradiation, respectively. Figures 4(b) and (c) reveal that the relative area ratios of the peaks (CF₂O:CFO and CF₂:CF) are 24:59:17 before femtosecond laser irradiation and are 19:63:18 after laser irradiation. This reveals that laser irradiation increases the CFO and CF₂ bond concentrations, while the CF₂O bond concentration is reduced by laser irradiation. This suggests that femtosecond laser irradiation induces partial bond breaking of the CO and CF bonds in the CF₂O bond (labeled (3) in Fig. 4(a)). Although the detailed changes in the chemical bonding structure of femtosecond laser irradiated CYTOP still remain unclear, we suspect that the partial bond breaking of CF and CO bonds by the femtosecond laser, which would cause the volume to contract, may contribute to the increase in the refractive index of CYTOP.

 

Fig. 4 (a) Chemical bonding structure of unirradiated CYTOP. C1s XPS spectra of (b) unirradiated and (c) irradiated CYTOP.

Download Full Size | PDF

4. Conclusion

In conclusion, we have fabricated UV waveguides in CYTOP by femtosecond laser direct writing. The refractive index increase in the core was estimated to be 1.7 × 10⁻³, which might be caused by partial bond breaking in the CF₂O bonds in CYTOP due to the femtosecond laser irradiation. The waveguides functioned as single-mode waveguides at 632.8, 355, and 266 nm wavelengths. Propagation losses of 0.49, 0.77, and 0.91 dB/cm were obtained at wavelengths of 632.8, 355, and 266 nm, respectively. The fabricated waveguides presented here will be very useful for UV analysis of samples and UV-induced stimulation of biocells in microbiochips.