*2.1. Planar Waveguide Biosensor Design and Testing*

The detection principle of the planar waveguide (PW) biosensor acting as a polarisation interferometer (PI) is similar to Mach–Zehnder (MZ) interferometers, but instead of two optical arms in the MZ biosensor, the p- and s-polarisations of light were used as parallel parameters in this set-up. The PW being the main element of the developed biosensor was devised on an Si wafer and consisted of a 200 nm thick silicon nitride (Si3N4) ore layer (having a RI *n* = 2.01) placed between two much thicker (3 µm) cladding silicon oxide (SiO2) layers of lower RI (*n* = 1.46). Such design allowed the propagation of a single mode electromagnetic (EM) wave through the waveguide by multiple internal reflection; the large difference in RIs between the Si3N<sup>4</sup> core and the SiO<sup>2</sup> cladding resulted in light propagation at a steep angle of 47◦ , creating a large number of internal reflections of light (up to 3000 reflections/mm) along the PW.

As shown in Figure 1, the polarised 630 nm light from a laser diode (1) was coupled into waveguide (4) via a slant edge, which was polished at a 47◦ angle to provide a 90◦ incidence angle and therefore maximal efficiency of coupling. The light was converted to circular polarisation using a λ/4 plate (2) and focused on the slant edge using a lens (3). The outcoming light is going through a polariser (7), which converts the changes in the EM wave polarisation into modulation of its intensity, and collected by a charge-coupled device (CCD) array photodetector (8). The waveguide (4) with the dimensions of 25 × 8 mm is held between two pieces of black nylon with the upper piece forming an 8 × 2 × 6 mm (≈0.1 mL) cell (6) sealed against the top side of the waveguide and equipped with inlet and outlet tubes enabling injecting different chemicals into the cell. In the earlier versions of the set-up, the top layer of SiO<sup>2</sup> is etched away by injecting 1:10 diluted hydrofluoric acid into the cell to form the sensing window. Later on, in the advanced experimental set-up, both the waveguiding core and sensing window were formed by photolithography.

**Figure 1.** (**a**) The planar waveguide (PW) biosensor experimental set-up: laser diode (1), λ/4 plate (2), collimating lens (3), PW (4) on Si wafer support (5), reaction cell (6) with inlet and outlet tubes, polariser (7), and charge-coupled device (CCD) array (8); (**b**) Cross-section of waveguide section showing schematically the multiple reflections of light, the sensing window, the reaction cell, and antibodies immobilised on the PW surface binding zearalenone molecules.

The resulted set-up operates as a planar polarisation interferometer (PPI); the pcomponent of polarised light (lying in the plane of incidence) is affected by changes in the RI of the medium, while the s-polarised component (orthogonal to the plane of incidence) is almost invariant to the RI variation in the medium and subsequently used as a reference. Any changes in the medium RI in the sensing window including the variations of RI caused by molecular adsorption result in a multi-periodic sensor response cause by a variable phase shift between p- and s- polarisations of light, which could be converted by a polariser to a multiperiodic signal. In a way, the principle of PI is a logical expansion of the TIRE method [3], which is based on the detection of a phase shift between p- and s-components of polarised light, utilising a large number of reflections in the optical waveguides.

The experimental set-up for PI went through several stages of optimisation. Previously, the light from a fan-beam laser diode was coupled into the PW and was propagated over the entire width (≈8 mm) of the waveguide. As the result of a modal dispersion of light across the waveguide, and therefore not equal conditions of light propagation (see Figure 2a), averaging of the light intensity over the entire width of the waveguide has led to losing the contrast of the interference pattern. To avoid that, the number of pixels for light averaging had to be limited. However, improved results were obtained with photolithography to form a narrow strip (≈2 mm) of silicon nitride (Figure 2b). Another advantage of photolithography was the formation of a well-defined sensing window. The photographs of Figure 3 show the PW biosensor set-up (Figure 3a), the top views of the waveguide at different preparative stages in Figure 3b, e.g., a 24 × 6 mm chip with an SiO<sup>2</sup> and Si3N<sup>4</sup> layer deposited (1), after etching of Si3N<sup>4</sup> to form a narrow (2 mm) waveguide core (2), and the final structure with the sensing window (2 × 8 mm) etched in the top SiO<sup>2</sup> layer (3), and the waveguide inserted in the cell (Figure 3c). A Thorlabs LC100 camera was

interfaced to a PC; the output signal acquisition was carried out using SPLICCO software (A Thorlabs GmbH, Bergkirchen, Germany).

**Figure 2.** Propagation of light through the planar waveguide (PW) in a wide core (**a**) and narrow core (**b**) set-up.

**Figure 3.** The upgraded planar polarisation interferometry device. Planar waveguide-based polarisation interferometry experimental set-up upgraded from the prototype (**a**); photolithography steps the waveguide preparation (**b**); the reaction cell with the waveguide inserted (**c**).
