1. Introduction

The advancement of microfabrication techniques for lab-on-a-chip systems has enabled integration of optical and fluidic components. By this approach, the field of optofluidics has emerged, including both devices where optical functionality is added to fluidic systems and devices where fluidics is adding functionality to optical systems [1]. It is of particular interest to integrate tunable light sources to lab-on-a-chip microsystems [2, 3, 4], aiming for novel sensor concepts [5].

One approach for compact integrable light sources is microfluidic dye lasers in which a liquid laser dye is pumped through a microfluidic channel with an embedded optical resonator [6, 7, 8 , 9, 10, 11, 12, 13, 14]. For integration, a laser resonator with lateral emission is preferred. Of different laterally emitting device layouts, distributed feedback (DFB) laser resonators have proven particularly suitable for obtaining single mode and low threshold lasing [7, 11, 13].

In order to achieve a low threshold for lasing the optical losses should be minimized. Two main challenges to be addressed in this context are waveguiding in the liquid dye (liquid-core waveguiding) and out-of-plane scattering from high order Bragg reflection. Liquid-core waveguiding was achieved using a low refractive index polymer (polydimethylsiloxane, PDMS) in combination with a high refractive index liquid [9, 11]. Wavelength tuning by mechanical deformation of a PDMS-based 15th order DFB laser was reported in Refs. [1, 15]. A tuning range of 29 nm was achieved with a single laser dye, and a total tuning range between 565 and 638 nm was demonstrated by employing two different laser dyes.

The requirement of a low refractive index polymer and a high refractive index liquid may be relaxed by reducing the dimensions of the resonator segments as the subwavelength regime is entered [16]. Following this approach, a periodic array of nanoscale fluidic channels forms a third order DFB resonator, as in our previous demonstration [13]. The optofluidic third order DFB dye laser devices are defined by combined electron beam and UV lithography (CEUL) [17] in a thin film of SU-8 resist on a SiO₂ substrate and the channels are sealed by polymer-mediated wafer bonding of a glass substrate. The hydrophilicity of the SiO₂ [18] at the bottom of the nanochannels facilitates filling of the nanochannels with a dye solution by capillary action. This approach simplifies fluidic handling as the dimensions of the resonator segments enter the nanoscale.

Despite the lack of waveguiding in the dye-filled nanofluidic channels, the third order Bragg grating yields low out-of-plane scattering losses and a large free spectral range (FSR). In Section 2, the FSR at the third order Bragg reflection wavelength is estimated to be 290 nm for a grating of period Λ = 599 nm filled with an ethylene glycol solution. This very large FSR ensures that, for visible light, only the third order resonance falls inside the wide dye gain spectrum and facilitates tuning of the laser over the full gain spectrum of the chosen laser dye.

In this work, we investigate the optofluidic tuning range of the third order DFB dye laser. By combining different grating periods and dye solution refractive indices, we demonstrate a tuning range of 45 nm using a single laser dye.

2. Tunable optofluidic third order DFB dye lasers

The design and fabrication of the third order optofluidic DFB dye lasers are described in Ref. [13]. For clarity, the key parameters are presented here. The laser resonator structure, see Fig. 1, is based on a planar polymer waveguide structure supporting a single propagating TE-TM mode. The basic waveguide structure consists of a SiO₂ buffer substrate (refractive index n = 1.46), a 300 nm thick core of the negative-tone resist SU-8 (n = 1.59), and a top cladding of polymethylmethacrylate (PMMA) (n = 1.49). An array of nanofluidic channels, defined lithographically in the SU-8 film, constitutes the third order Bragg grating DFB laser resonator. The resonator has a phase shift of $\frac{\pi }{2}$ in the middle, see Fig. 1(b), to obtain a single resonance for each Bragg reflection. The laser resonator is embedded in a 300 nm high 500 μm wide fluidic channel which is followed by a 100 µm wide 16 cm long meandering channel to facilitate fluid flow through the resonator by capillary action. The laser emits light laterally in the chip plane. The light is coupled directly into the SU-8 layer which guides the light to the edge of the chip for measurement. The laser structure provides tuning of the wavelength by changing the grating period and through microfluidic functionality by changing the refractive index of the dye solution.

 

Fig. 1. Sketch of a fabricated optofluidic third order DFB dye laser. (a) Top-view illustrating the 500×500 μm² DFB laser resonator with central phase shift situated in a microfluidic channel. (b) Atomic force micrograph of the central region of the third order DFB laser resonator. In the middle an extra channel width is included to yield the $\frac{\pi }{2}$ phase shift. The width of the white scale bar is 600 nm. (c) Side-view showing the refractive index distribution of the structure.

Download Full Size | PDF

In order to estimate the resonances and the FSR of the passive DFB laser resonator, we consider the periodic arrangement of polymer and liquid segments in the layered structure shown in Fig. 1(c). The SU-8 (n ₁ = 1.59) in the polymer segments acts as the core of a planar dielectric waveguide supporting a single well-confined propagating TE-TM mode. Since the refractive index of the liquid in the nanofluidic channels n ₂ is lower than that of the buffer and cladding layers, a well-confined TE-TM-like mode does not exist in the liquid segments. This lack of waveguiding in the liquid segments gives rise to losses and complicates the modelling of light propagation in the structure. The coupling losses for light traversing a dye-filled nanofluidic channel are estimated to a few percent using finite-difference beam propagation method calculations [13]. In general, a full solution to Maxwell’s equations in the geometry is required.

The resonances of the grating satisfy the Bragg condition

where m = 1,2,3,… is the reflection order, λ_m is the free space wavelength of the m’th reflection order, and Λ_op is the optical path length of one grating period. The spatial dimensions of the resonator segments are carefully chosen so that the DFB laser oscillates at the third order Bragg reflection wavelength [13]. From the experimental data, summarized in Table 1, λ_op can be estimated. Given Λ_op, the FSR is calculated as FSR = λ_m/(m- 1).

As an example, a grating of period Λ = 599 nm filled with an ethylene glycol solution (n ₂ = 1.43) is considered. From (1) using m = 3 and λ3 = 580.60 nm, we obtain Λ_op = 871 nm and a FSR of 290 nm at the third order Bragg reflection wavelength. This very large FSR ensures that, due to the third order Bragg grating, only the third order resonance lies within the gain spectrum of the chosen dye (for visible wavelengths), thus enabling a wide tuning range.

3. Results and discussion

To operate the devices, a droplet of liquid dye solution is applied to the fluid inlet hole and the fluidic network is filled by capillary action with no need for external syringe pumps. The speed of the advancing fluid front in the meandering channel decreases with time. The dye solution in the resonator is replaced within 30 s – 5 minutes during the first 30 minutes of experiments (ethylene glycol solution) [13]. Here, the laser dye rhodamine 6G (R6G) dissolved in (i) ethylene glycol (n = 1.43), (ii) a 2:1 mixture of ethylene glycol and benzyl alcohol (n = 1.467), and (iii) a 1:1 mixture of ethylene glycol and benzyl alcohol (n = 1.485), all with a concentration of 2 × 10^-2 mol/L is used. The refractive indices of the mixtures are determined by experimentally confirmed linear extrapolation of the refractive indices of the pure solvents.

Table 1. Summary of the data presented in Fig. 2. The laser wavelength λ is tuned by changing the grating period Λ and the refractive index n of the R6G solution.

View Table

 

Fig. 2. Laser spectra from different laser devices showing a tunability of the lasers of 45 nm by changing the grating period and the refractive index of the R6G solution. Lasers with three different periods were fabricated. The data is summarized in Table 1.

Download Full Size | PDF

The laser is optically pumped at 532 nm perpendicularly to the chip plane by a frequency doubled Nd:YAG laser (5 ns pulse duration, 10 Hz repetition rate) and the output dye laser light is collected by an optical fiber at the edge of the chip and analyzed using a fixed grating spectrometer (resolution 0.15 nm).

Lasers with three different grating periods Λ = 579, 599, 620 nm were fabricated by varying the width of the nanochannels and keeping the width of the SU-8 walls fixed, cf. Fig. 1. The periods were derived from atomic force micrographs of the SU-8 structures prior to bonding, see Fig. 1(b). Figure 2 shows laser spectra for laser devices with different grating periods using different R6G solutions. The lasers oscillate in a single mode (except the laser of Fig. 2(b)) and are polarized perpendicular to the chip plane (TM). The data are summarized in Table 1 and demonstrate wavelength tunability of 45 nm by changing the grating period (coarse-tuning) and dye solution refractive index (fine-tuning). The minor side-modes in the spectrum in Fig. 2(b) may be due to grating defects due to fabrication errors. The spectral shape and linewidth of the lasers remain undetermined due to limited spectrometer resolution.

 

Fig. 3. Typical pump pulse fluence/output laser power graph for the laser device with spectrum shown in Fig. 2(e). The graph follows the standard pump/output relation of two linear segments around a laser threshold of ~ 7 μJ/mm².

Download Full Size | PDF

 

Fig. 4. Left: Laser wavelength as a function of grating period for different R6G solutions. The fits indicate a linear dependence for each solution. Right: Laser wavelength as a function of dye solution refractive index for different R6G solutions with fixed grating period Λ = 599 nm. The fit indicates a linear dependence.

Download Full Size | PDF

Figure 3 shows the output laser power as a function of the average pump pulse fluence for the laser of Fig. 2(e). The graph follows the standard pump/output relation of two linear segments around a threshold fluence of approximately 7 μJ/mm² and is typical for the lasers. The error bars of the output laser power are calculated as the standard deviation of a number of output laser power measurements using the same average pump pulse energy fluence and reflect the pump pulse energy fluctuations from our Nd:YAG system.

Figure 4 shows the laser wavelengths plotted as a function of grating period and dye solution refractive index (for Λ = 599 nm). The graphs provide linear tuning curves and indicate a linear dependence of the wavelength on grating period and dye solution refractive index (refractive index tuning). This may be interpreted as a linear regime where the Bragg condition (1) applies although the resonator structure of Fig. 1 is not a simple layered dielectric stack but rather a two-dimensional structure of sub-wavelength dimensions which, in general, requires a full solution of Maxwell’s equations in the geometry [16].

4. Conclusion

We demonstrate tunable third order optofluidic DFB dye lasers. The lasers rely on light-confinement in a nanostructured polymer film and third order Bragg reflection. The lasers are fabricated by CEUL in SU-8 and polymer-mediated wafer bonding, ideal for fast and flexible prototyping of device designs. Due to the large FSR of the third order Bragg grating, the laser resonator provides a wide tuning range of 45 nm using a single laser dye by changing the grating period and dye solution refractive index (refractive index tuning). Despite the lack of waveguiding in the dye-filled nanofluidic segments of the laser resonator, the lasers exhibit low threshold fluences down to approximately 7 μJ/mm² due to the sub-wavelength dimensions of the resonator segments and low out-of-plane scattering. To provide real-time wavelength tunability, the laser could be combined with a microfluidic mixer [19]. The laser is suitable for integration on lab-on-a-chip micro-systems, e.g. for novel sensor concepts, where coherent, tunable light in the visible range is desired.