Antiresonant reflecting optical waveguides for biosensing applications

Over the last few years, much effort has been devoted to the development of so-called lab-on-chip (LOC) systems that are based on optical detection methods.¹ Applications of these systems include environmental monitoring, health care, biosensing, and industrial analysis. To carry out complete assays with LOC devices, it is necessary to integrate a wide variety of microfluidic and optical function components onto the same microchip. Moreover, although very low detection limits can be realized with optical devices, incorporating these components into an LOC system is not an easy task.

Conventional approaches to chip integration are based on solid-core waveguides.² The light–sample interaction in such waveguides, however, is weak. This is because the interaction is related to only a small fraction of the mode field (i.e., the evanescent tail). These problems, as well as the miniaturization trend, have motivated the pursuit of innovative design concepts and technologies to provide more effective solutions for on-chip integration of microfluidic systems and optical devices.³ Indeed, interest in liquid-core waveguides has recently increased.⁴

In our work, we have demonstrated liquid-core waveguides as novel optofluidic devices that can be used in highly functional microsensors for LOC devices. The strength of these waveguides is that they intrinsically provide (through the wave-guiding structure itself) a hollow cavity that can serve as a microchannel for liquid flow. Light can directly couple with the substance of interest in the small and well-defined volume of this channel and thus provide a strong enhancement in the sensitivity of optical sensors.⁵ In addition, our integrated liquid-core waveguides, known as antiresonant reflecting optical waveguides (ARROWs), can be fabricated with silicon photonic technology.

We achieve light confinement in the low-refractive-index core of an ARROW by depositing a high-reflectivity multilayer stack on the sidewalls of the core. We produce this stack with repeating bilayers of alternating high- and low-index layers, where each layer is antiresonant at a specific wavelength (see Figure 1). The confinement mechanism itself arises from optical interference effects in the claddings. These effects then result in low-loss propagation of a broad spectral band centered around the operating wavelength. By properly selecting the cladding materials and the waveguide geometry, we can design our ARROWs to operate over the entire visible to far-IR spectral window. This tailoring of the ARROW design for the fabrication of complex integrated photonic devices means that our waveguides are promising candidates for many applications (ranging from the detection of chemical/biological agents to clinical diagnostics and environmental monitoring).

Figure 1. Schematic cross section of a liquid hybrid antiresonant reflecting optical waveguide (h-ARROW), showing a simulated power distribution of the fundamental mode.

Our concept of antiresonant wave-guiding for the realization of liquid-core waveguides is successful for several reasons. For example, ARROWs (unlike photonic-crystal-based waveguides) do not require a periodicity in the alternating layer structure, which means that the fabrication tolerance can be relaxed. In addition, ARROWs can be fabricated with the use of planar silicon-compatible technology, which is particularly useful for on-chip integration. We also note two other important properties of ARROWs. First, their modal operation can be tailored by exploiting the weaker reflectivity that is experienced by the higher-order modes of reflection at the cladding interfaces. Second, the modes experience attenuation losses that depend on the state of polarization. By combining these two intrinsic properties, along with an accurate design strategy, we can obtain low-loss, single-mode, linearly polarized ARROWs. The devices can then be used in the design of high-performance optical devices, such as high-visibility integrated interferometers or a high-quality-factor optical ring resonator for liquid sensing.⁵

We fabricate our ARROWs through a bulk micromachining process. In addition, we have made many technological developments so that we can use ARROWs as the building blocks of photonic structures that integrate several optical elements (e.g., intersecting solid-core waveguides, low-loss straight and curved waveguide sections, and power splitters). For example, we have used ARROWs with simple straight waveguides to act as a long path absorbance cell for colorimetric detection of bovine serum albumin protein in a water solution (with a loss-on-drying concentration of 1.5 × 10⁻³mg/mL).⁶

In addition to the use of ARROWs for guiding liquid flow in optical sensing analyses, recent progress has also been made in integrating the devices with other microfluidic functionalities for sample processing.^{7, 8} For example, we have developed the hybrid-ARROW (h-ARROW) by using hybrid silicon/polymer technology. The h-ARROW device comprises an optofluidic channel from a full-silicon ARROW that is sealed with a thin polydimethylsiloxane polymer layer (see Figure 1). This configuration allows us to merge the strong optical performance of silicon technologies with the advantages of polymers. That is, the microfluidic layer integrates several important functional elements, e.g., a fluid inlet and outlet, and channels for sample delivery.

So far, we have exploited this hybrid approach for the development of an optofluidic platform prototype for fluorescence measurements.⁹ In this prototype, we integrate the liquid-core h-ARROW with self-aligned solid-core waveguides and a microfluidic mixer, in a multilayer approach. This results in a compact 3D device assembly (see Figure 2). The h-ARROW could also be used as a tool to realize more sophisticated sensing devices, i.e., that can overcome the decreased sensitivity of miniaturized devices (caused by the reduction in optical path length). For example, we have developed hybrid liquid-core ring resonators, with quality factors of up to 1.44 × 10⁴ (see Figure 3).¹⁰ These devices benefit from the strong light–matter coupling that occurs in the liquid core, as well as from the recursive interaction between the light and the sample. These attributes make the liquid-core ring resonators suitable candidates for high-sensitivity label-free biosensors. We thus evaluated the sensitivity of the liquid ring to changes in refractive index. We measured the shift of the resonance wavelengths, and found—see Figure 3(c)—a bulk sensitivity of 700nm per refractive index unit (RIU) and a detection limit of about 1.6 × 10⁻⁶ RIU.

Figure 2. (a) Schematic drawing of the h-ARROW optofluidic platform and (b) a photograph of the fabricated device.

Figure 3. (a) Schematic view of the hybrid liquid-core ring resonator and (b) a scanning electron microscope image of its bottom silicon part. (c) The change in resonant wavelength of the ring resonator as a function of change in the refractive index of the core (Δn_core). The inset shows a plot of optical resonances and the fitted Lorentzian curve (red line). a.u.: Arbitrary units. RIU: Refractive index unit.

In summary, we have developed a new type of optofluidic device with liquid cores that are known as ARROWs. These waveguides offer a range of advantages, and we have therefore conducted a large amount of work to demonstrate them as suitable building blocks for high-performance sensing devices that can be used in complex optical analyses. In our research group, we are now focusing on the use of ARROWs to develop a new type of highly sensitive optofluidic biochemical platform that is based on ring resonators. This platform should be suitable for point-of-care applications such as sepsis biomarker detection.

The authors are grateful for support from the Italian Ministry of Education, University and Research, through the Futuro in Ricerca program (grant RBFR122KL1).