JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 40, NO. 1, JANUARY 1, 2022 237 Novel Sensing Algorithm for Linear Read-Out of Bimodal Waveguide Interferometric Biosensors Bernat Bassols-Cornudella , Patricia Ramirez-Priego , Maria Soler , M.-Carmen Estévez, Heriberto J. Díaz Luis-Ravelo, Maria Cardenosa-Rubio , and Laura M. Lechuga , Member, IEEE, Fellow, OSA Abstract—Biosensors employing photonics integrated circuits, on interferometric architectures have shown a remarkably good and specifically those that rely on interferometric evanescent wave sensitivity; 10−7 –10−8 refractive index units (RIU) for bulk working principles, have outstanding performances due to the ex- detection and in the order of few pg/mm2 for surface sensing, treme sensitivity exhibited in one-step and direct assay, without the need of amplification. Within the interferometric configurations, while presenting clear advantages for label-free analyses [1], the Bimodal Waveguide (BiMW) interferometric sensor stands [3], [4]. out due to its demonstrated sensitivity for real-life applications Classical interferometric architectures such as the Mach- and the simplicity of its design. To overcome the ambiguities that Zehnder (MZI) [7]–[10], or Young interferometers [11]–[13] arise from the periodic nature of interferometric read-outs, a new are based on two-path schemes. In these devices, an initial all-optical modulation and the subsequent trigonometry-based al- gorithm have been proposed and applied to the BiMW biosensor. light beam is split into a reference branch where the evanescent This new algorithm has been successfully employed for the se- probing field remains intact and a sensing branch responsible for lective identification and quantification of the external Spike (S) guiding the light beam through a sensing area. The properties protein of the Severe Acute Respiratory Syndrome Coronavirus 2 of the light beam traveling through the sensing branch will be (SARS-CoV-2). Our biosensing results from this simple, quick, and altered depending on the refractive index (RI) of the analysed user-friendly method demonstrate high sensitivity and specificity and pave the way towards a point-of-care device for general use. solution. As a result, interference takes place after recombin- ing both light beams, producing quantifiable variations on the Index Terms—Biosensors, integrated optics, silicon photonics, resulting light intensity. waveguide interferometry, coronavirus. Alternatively, single-path configurations have been proposed to simplify fabrication processes, as well as to reduce costs and I. INTRODUCTION footprint [14], [15]. In these devices, two light modes of different HE need for rapid, highly sensitive, inexpensive, and T portable platforms for the real-time detection of low con- centrations of analytes has increased in recent years and has order or polarization propagate through the same waveguide, where their respective evanescent fields are affected by the RI of the analysed solution simultaneously. The displacement between been revealed of paramount importance during the COVID-19 these modes results in an interferometric output, measurable pandemic. Portable biosensors able to be used at the point-of- either as a variation of the far-field pattern created at the sensor care (POC) are the most effective means of early diagnostics output and quantifiable by their relative phase variation or as an regarding clinical and pharmaceutical analysis, medical diag- intensity deviation after recombining both waves into a single nostics [1], [2], environmental monitoring [3], and many other mode [16], [17]. This idea was first put into practice with industrial applications. In this rapidly increasing field, photonic optical fibres [18]–[22] and later applied to silicon photonics, biosensors based on evanescent wave detection have shown to be resulting in the emergence of the Bimodal Waveguide (BiMW) one of the best candidates for such POC devices given their ex- interferometric biosensor [17], [23]. tremely high sensitivity [4]–[6]. More specifically, those based Interferometric biosensors can achieve very accurate and sen- sitive results at the expense of hindering the read-out response. Manuscript received June 1, 2021; revised August 28, 2021; accepted Septem- The periodic (sinusoidal) nature of the output intensity with ber 30, 2021. Date of publication October 6, 2021; date of current version January 3, 2022. This work was supported by the H2020 Research and Innovation respect to phase variations raises a series of analytical difficulties Programme of the European Commission (CONVAT project, H2020-SC1-PHE- when trying to evaluate the phase displacement, i.e., the evalu- CORONAVIRUS-2020) under Grant 101003544. The ICN2 was supported by ated parameter Δϕ, such as sensitivity fading, fringe order, and the CERCA programme of the Generalitat de Catalunya and by the Severo Ochoa Centres of Excellence Program, funded by the AEI under Grand SEV- direction ambiguities, as shown in Fig. 1. 2017-0706. (Corresponding author: Laura M. Lechuga.) In order to overcome these drawbacks and obtain an The authors are with the Nanobiosensors and Bioanalytical Applica- unambiguous read-out, a wide range of phase compensation tions Group (NanoB2A), Catalan Institute of Nanoscience and Nanotech- nology (ICN2), CSIC, CIBER-BBN and BIST, 08193 Barcelona, Spain (e- and modulation techniques have been suggested. They are mail: bernat.bassols-cornudella20@imperial.ac.uk; patricia.ramirez@icn2.cat; based on a variety of physical principles induced by electro maria.soler@icn2.cat; mcarmen.estevez@icn2.cat; heriberto.diazluis@gmail. and magneto-optical components [24]–[27], thermo-optical com; maricruz.cardenosa@icn2.cat; laura.lechuga@icn2.cat). Color versions of one or more figures in this article are available at properties [28], or rely on the use of liquid crystals and https://doi.org/10.1109/JLT.2021.3118103. photosensitive layers [29], [30]. Further research has also been Digital Object Identifier 10.1109/JLT.2021.3118103 carried out for the implementation of spectral interrogation This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License. For more information, see https://creativecommons.org/licenses/by-nc-nd/4.0/ 238 JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 40, NO. 1, JANUARY 1, 2022 allows the propagation of the fundamental and first TE modes, previously shown to be the most sensitive ones [38]. More specifically, as shown in Fig. 2, when light travels through the waveguide, it first encounters a single-mode section (150 nm thickness) in which the light solely propagates in its fundamen- tal mode. This region is followed by a modal splitter, i.e., a rigorously engineered step junction that geometrically tolerates the appearance of the first mode. After this point, both modes propagate through the bimodal section (340 nm thickness) until the waveguide exit. In this bimodal section, a rectangular part of the upper cladding is removed to create the sensing area (centred with dimensions = 50 μm width × 15 mm length). In general, as light travels through a waveguide, its physical properties are determined by the core and the top and bottom Fig. 1. Usual interferometric output ambiguities that hinder the quantification claddings refractive indexes. Therefore, any change in the mate- of the signal: sensitivity fading, directional ambiguities and fringe order ambi- rials will produce a variation of the light propagation properties. guities. The correct value to be measured is highlighted in red and the quadrature To be more precise, as both modes propagate along the bimodal point, where sensitivity is maximized, is also shown. section (see Fig. 2), their evanescent field, i.e., the electric field exponential tails that slightly penetrate into the claddings (see techniques that have been successfully applied on slightly equation (3.53) in [40]), are altered by the corresponding RI modified MZI versions [31]–[36]. However, the practical of the material. When the two modes reach the sensing area, execution of these methods requires the addition of expensive they are affected by the solution or the molecules attached to its and space-demanding equipment such as high precision surface. Any change to this external environment will imply a spectrometers, narrowband tunable lasers, or an extensive modi- change on the top cladding RI, inducing a quantifiable variation fication of optimized fabrication protocols. Our group previously to the properties of the probing light beam. proposed an all-optical modulation technique together with an Once the propagating modes have crossed the device, they algorithm based on Fourier Series deconvolution for achieving a exit the waveguide such that, if projected on a screen, light linear read-out of the MZI and the BiMW biosensor [37]–[39]. A expands vertically, forming a bright, straight line containing the simple implementation was feasible due to the use of commercial interferometric output as shown in Fig. 2. laser diodes under high-frequency intensity modulations, without the need to incorporate any extra apparatus. B. Experimental Setup and Output Now we propose a novel and improved BiMW read-out algorithm based on simple trigonometric properties of the The visible light travelling in the BiMW interferometric sen- periodic read-out signal, which has been implemented using sor was generated by a commercial laser diode (HL6545MG, custom-made software. The proposed method uses an all-optical λ = 600 nm, Thorlabs, Newton, New Jersey, USA) plugged modulation technique as previously applied in [37], although in a miniature temperature-controlled Cooled Mount (LDM21, slightly modified for employing the new algorithm, resulting in Thorlabs, Newton, New Jersey, USA) and kept at a constant a faster, more sensitive, and user-friendly modulation technique temperature of 25.00 ± 0.01 ◦ C. An aspheric lens (C240TMD-B, with no additional costly equipment. We show the excellent Thorlabs, Newton, New Jersey, USA) directed the emitted light performance of this new read-out algorithm with a biosensing towards the preferred direction as a collimated light beam. proof-of-concept by detecting the external Spike (S) protein of The light beam was then polarized in the TE direction by a the Severe Acute Respiratory Syndrome Coronavirus 2 (SARS- free-space isolator (IO-3D-660-VLP, Thorlabs, Newton, New CoV-2). This new algorithm has been optimized for its use Jersey, USA) and focused on the waveguide entrance by a with the BiMW interferometric biosensor, although it could be 40× optical objective (Olympus PLN 40× Objective, Edmund extended to any other interferometric-based platform that allows Optics, Barrington, New Jersey, USA). A total of 20 opera- a sinusoidal intensity-induced phase modulation such as the one tive sensing bimodal waveguides are included in each sensor presented below. chip (3 cm long × 1 cm width), fabricated at wafer-scale in the cleanroom facilities of the Institute of Microelectronics II. THE BIMW INTERFEROMETRIC BIOSENSOR of Barcelona (IMB-CNM-CSIC), as previously described [23]. The sensor chips were placed on a custom-made, temperature- A. Sensor Working Principle controlled holder adjustable for the coupling of the light in the The BiMW biosensor working principle relies on the change waveguide. A fluidic system ensures the continuous flow of of the properties of a monochromatic transverse electric (TE) a buffer solution over the sensor chip and the liquid sample polarized light beam [23]. The waveguide is a three-dimensional injection to be analysed on the sensing area. The fluidic system (2.5–3 μm width × 2–3 nm height × 3 cm length) rib waveguide includes a polydimethylsiloxane (PDMS) microfluidic cell with made of a silicon dioxide (SiO2, RI = 1.46 ± 0.01 RIU) cladding five channels (channel dimensions = 1.25 mm wide × 500 μm and a silicon nitride (Si3N4, RI = 2.00 ± 0.01 RIU) core that height) that is hermetically sealed to the sensor chip; a syringe BASSOLS-CORNUDELLA et al.: NOVEL SENSING ALGORITHM FOR LINEAR READ-OUT 239 Fig. 2. BiMW sensor sketch showing the single mode propagating section, the bimodal section, the step junction, and the sensing area. Travelling light modes are simulated in orange for both sections, single and bimodal. As the laser beam, in red, exits the waveguide, it expands vertically when projected on a screen. Horizontally separated intensities Iup and Idown are recorded in real-time. When a sample is injected in the sensor, and under no laser intensity modulations, the light distribution output will shift up and down as shown in the magnified window. pump (New Era, US) for the continuous flow of the running buffer solution (e.g., milli-Q water, phosphate-buffered saline, etc.); and a 6-port valve (CHEMINERT, VICI, USA) for the injection of the samples in the sensor. The light output was recorded by a horizontally split two-sectional photodetector (S4349, Hamamatsu, Japan). The received data was collected by an acquisition card (NI USB-6361, National Instruments, USA) and treated by a custom-made Python software. Due to the interferometric nature of the device, RI changes on the chip sensing area caused the vertical output line to shift up and down. This oscillatory movement was recorded and converted into a quantifiable signal able to measure any RI changes in the sensing area. Having a two-sectional photodetector provided real-time values for the intensities Iup and Idown shown in Fig. 2, such that without any external tunning of the laser intensity, i.e. Fig. 3. Recorded SR signal during the detection of HCl 0.1 M injection under under a non-modulating regime, we could define the sensitivity milli-Q water constant flow, with a RI change Δn = (8.21 ± 0.01) × 10−4 RIU. parameter, SR [23], [37]–[39], [41], as follows, Iup − Idown SR = = A + B cos ϕ, (1) Iup + Idown C. Phase Retrieval Techniques. A New Algorithm where ϕ is the light phase. Changes like the flow of a new A phase retrieval process must be carried out in order to obtain substance on the sensing area or the binding of biomolecules the temporal evolution of the phase ϕ. This is only possible in resulted in variations on the light propagation characteristics relative terms by computing the phase difference Δϕ between and, therefore, on its phase ϕ and the signal SR . Hence, Eq. (1) that of the initial solution and the introduced one. Then, Δϕ is shows that as time goes by and no changes on the sensing area compared to a set of calibration values from which the solution occur, the sensor output will remain a continuous straight line, RI is finally acquired. However, the non-bijective periodic nature and ϕ will not be altered. In other words, if a solution of the same of the SR signal and its resulting phase fringe and direction RI as the buffer flows through the sensing area (e.g., milli-Q ambiguities as well as a potential sensitivity fading, hinder water while also running milli-Q water), no changes in the SR this process and make it much more complex than desired. will be observed. However, the flow of a different refractive Previous methods involved counting fringes and periods pro- index solution (e.g., hydrochloric acid (HCl) 0.1 M, Δn = (8.21 vided the computed signal is a 2π-periodic function under static ± 0.01) ×10−4 RIU) generates the oscillations on the output as intensity or applying an experimentally restrictive sinusoidal ϕ varies, changing the SR sinusoidally. This behaviour is shown intensity modulation followed by an algorithm based on Fast in Fig. 3, where the SR signal sinusoidally oscillates if and only Fourier transforms and the Bessel functions of the first kind if a change in ϕ occurs. of order 2 and 3 [37]–[39]. To solve these drawbacks, a new 240 JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 40, NO. 1, JANUARY 1, 2022 all-optical modulation algorithm based on simple trigonometric be known before any measurement as it is possible if the RI relationships was designed and implemented to easily retrieve increases with respect to the running buffer and negative if it de- the pursued phase difference, Δϕ. creases. Thus, the term (−1)k0 +1 shall not be taken as an incon- As presented in [39] and [42], we can induce a new term in venience. Furthermore, this algorithm is applicable to any type of the SR expression by sinusoidally modulating the intensity, such sinusoidal signal in which a time-dependent phase variation be- that Eq. (1) is modified to obtain Eq. (2): having like the modulating term presented above can be induced. To modulate the laser diode emission and apply the algorithm SR = A + B cos(Δϕ + ϕ0 + μ sin ωt), (2) presented, a custom-made Python application was developed. where t is the time, ω is the modulating frequency, and μ is the Specifically, the laser intensity was sinusoidally modulated from modulation depth, altogether constituting the modulating term, 100 to 200 mA at a frequency of 10 Hz. Such intensity changes μ sin ωt. Note that Δϕ = ϕ − ϕ0 , being ϕ0 the phase corre- induced a linear displacement of the central wavelength emitted sponding to the previously mentioned running buffer. Notice by the laser diode, resulting in the addition of the modulat- that neither A nor B constants are the same as those previously ing term μ sin ωt. It must be noted that for each probed laser defined in Eq. (1). intensity, the emitted wavelength distribution is very narrowly The first step in the algorithm consists in removing the offset A supported near its mean, covering less than 0.5 nm around the from the signal. This can be done assuming μ ≥ π, which allows centre, while being displaced almost 10 nm for the considered the cosine to fill all its image space, from −1 to 1 regardless of intensity modulation. At the same time, these characteristics are ϕ0 and Δϕ, ensuring the SR to be well defined and surjectively not dependant on the modulating frequency and do not induce a fill the interval [A − B, A + B]. Thus, the new centered signal perceivable noise on the system. presented in Eq. (3) will be considered from now on, specifically: The only restriction the proposed algorithm imposes is the need to have a value of μ ≥ π. This has been proven to be true  max SR + min SR SR = SR − as long as the intensity modulating amplitude is higher than 2 80 mA. To ensure a correct behaviour, it was decided to modulate = B cos(Δϕ + ϕ0 + μ sin ωt), (3) from 100 to 200 mA while remaining in the recommended intensity range of the laser diode and not forcing it to extreme for which, due to an abuse of notation, we shall drop the prime values. Moreover, the measuring times t0 and t1 used in the superindex. Let us consider the beginning of a measurement, presented phase retrieval method are not unique in most cases. in which the running buffer is kept constant such that Δϕ = 0. Depending on the SR obtained, more than one value for t0 or In this case, the recorded SR signal from Eq. (3) reaches its t1 can be found, thus allowing us to apply the algorithm to all maximum B at a time t1 , if ϕ0 + μ sin ωt1 = 2πk1 , for k1 ∈ possible pair combinations and consider the one showing a lower Z, and becomes 0 at a time t0 if ϕ0 + μ sin ωt0 = π/2 + πk0 , noise, overall increasing the biosensor sensitivity with respect for k0 ∈ Z. Therefore, constantly modulating under the same to previously employed phase retrieval methods. sinusoidal signal we can determine times t0 and t1 such that if If compared to the previous technique proposed in [39], there we compute the SR at these times, we obtain Eq. (4) and Eq. (5): are three key differences to be outlined. First, our new algorithm SR,t0 = B cos(Δϕ + ϕ0 + μ sin ωt0 ) allows for the use of a larger number of sensor chips since the   previous method requires a strictly fixed value of μ = 1.2π. π = B cos Δϕ + + πk0 Although usually achievable, slight fabrication tolerances may 2 make this unfeasible for certain sensor chips. Second, the mod- = −B sin(Δϕ + πk0 ) ulating frequency of 10 Hz ensures at least the same number of phase retrievals per second and can be increased accordingly = (−1)k0 +1 B sin Δϕ, (4) with compounds inducing strong RI which are used during the SR,t1 = B cos(Δϕ + ϕ0 + μ sin ωt1 ) experiments. Finally, the algorithm takes place in the time do- main instead of the frequency domain, which eases its technical = B cos(Δϕ + 2πk1 ) understanding by multidisciplinary users. Hence, the proposed = B cos Δϕ. (5) algorithm is faster, simpler, and more user friendly, altogether opening the way to be implemented in POC biosensor platforms. The phase difference information can now be directly retrieved dividing these expressions to get Eq. (6) and finally Eq. (7): D. Bulk Sensing Evaluation SR,t0 (−1)k0 +1 B sin Δϕ = In order to evaluate the bulk refractive index sensitivity of the SR,t1 B cos Δϕ BiMW interferometric biosensor under the discussed algorithm, = (−1)k0 +1 tan Δϕ, (6) five different HCl solutions (0.025, 0.05, 0.1, 0.25, and 0.5 M), SR,t0 whose respective refractive indexes were measured with a Δϕ = (−1)k0 +1 arctan , (7) digital refractometer (Rudolph Research Analytical, USA), were SR,t1 sequentially injected through the microfluidic system into the since the arctangent function can be easily unwrapped compu- sensing area, while milli-Q water was used as running buffer at tationally [39]. Note that the sign of the phase shift will always a constant flow rate of 20 μl/min. BASSOLS-CORNUDELLA et al.: NOVEL SENSING ALGORITHM FOR LINEAR READ-OUT 241 to follow real-time large and small refractive index variations happening on the sensing area with more precision, as well as lower noise (σ = 0.00184 rd) as compared with previous methods (see [43]). E. Biosensing Evaluation To fully demonstrate the new all-optical modulation capabil- ities for biosensing purposes, the rapid and direct detection of a specific structural protein of the novel Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) was assessed. Among all the SARS-CoV-2 viral antigens, the external spike (S) protein is a trimetric glycoprotein that plays an essential role in the infection, mediating the binding of the viral particles to host surface cellular receptors. The S protein comprises two func- tional subunits: S1 subunit that contains the receptor-binding do- main, which directly binds to the host cell receptor angiotensin- converting enzyme 2, and S2 subunit, which is responsible for the membrane fusion. As a proof-of-concept, S1 subunit was selected as the target analyte to be recognized by its specific antibody previously immobilized on the sensor chip surface. The BiMW sensor chip was first functionalized with tri- ethoxysilane polyethylene glycol carboxylic acid (silane-PEG- COOH 600 Da; Nanocs, New York, USA), following our pro- tocol previously detailed in [44]. The silanized sensor chip was directly placed into the experimental setup for further in-situ covalent binding of the antibodies. Carboxylic groups pro- vided by the silane-PEG-COOH were activated employing the well-known N-(3-dimethylaminopropyl)-N  -ethylcarbodiimide hydrochloride (EDC)/N-hydroxysulfosuccinimide (sulfo-NHS) Fig. 4. (A) Real-time computed phase difference during HCl calibration procedure with five different concentrations. The RI difference with respect to chemistry, highly reactive to primary amine groups acces- milli-Q, Δn is given in RIU with an error of 0.01 × 10−4 . (B) Calibration curve sible in the lysine residues of the antibodies. A solution with its corresponding linear fit, r2 = 0.9998, LOD = (3.21 ± 0.01) × 10−7 with 0.2 M EDC/0.05 M sulfo-NHS in MES buffer (2-(N- RIU. Plotted points correspond to the average of six measurements with a standard deviation of approximately 0.01 that cannot be appreciated in the plot. morpholino)ethanesulfonic acid 0.1 M, pH 5.5) was flowed over the sensor surface. Next, a solution with 50 μ g/mL of the specific chimeric monoclonal antibody against Spike protein (anti-S The new phase retrieval method was applied and the respec- chAb, 40150-D001, Sino Biological Inc., China) in acetate tive phase differences of each injection, Δϕ, were computed. buffer (10 mM, pH 5.5) was injected. After antibody binding, a Figs. 4(A) and (B) show the variation of the phase depending stable covalent amide bond was formed. The remaining activated on the HCl concentration and the calibration curve obtained by carboxyl groups were blocked with ethanolamine (1 M, pH plotting the phase difference for each HCl solution as a function 8.5) for 2 min. All steps of the immobilization of the anti-S of the refractive index change. The limit of detection (LOD) chAb were monitored in real-time with the new algorithm (see of the sensor is defined as the minimum variation of RI that Fig. 5 A). Milli-Q water was used as the running buffer during produces a signal at least three times higher than the detectable the immobilization step and was then switched to PBST buffer noise of the system. It is computed as LOD = 3σ/Sbulk , where (phosphate buffer saline (PBS 10 mM) with 0.05% Tween 20, σ is the standard deviation of the retrieved phase, which results pH 7.4) for the detection of S1 subunit protein. from the propagated error of the mean of the standard deviation Different S1 concentrations ranging from 50 to 1000 ng/mL after five experiments that run for 200 seconds with milli-Q were flowed over the sensor surface. As shown in Fig. 5(B), the water, and Sbulk is the slope of the linear regression in Fig. 4(B). biosensor response in real-time exhibits a correlation between Evaluations of the same sensor chip and different sensor chips the analysed concentration and the response. Furthermore, the showed consistent results of a LOD between 2 and 5 × 10−7 injection of a control protein (1 μg/mL C-reactive protein, RIU. These values demonstrate the validity of the newly pro- BBI Solutions, U.K.) confirms that the antibody exclusively posed algorithm as the results are identical to those obtained recognized the specific SARS-CoV-2 S1 subunit protein with employing previous techniques [37]. In fact, we did not expect no cross-reactivity. The obtained calibration curve shows a the algorithm to provide different phase variation values from the linear concentration-dependent response (r2 = 0.98) resulting ones obtained using other methods since the evaluated parameter in a LOD of 6.45 ng/mL. These results demonstrate that we was the same, Δϕ. On the other hand, this method was able could develop an effective biosensor with the BiMW sensor and 242 JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 40, NO. 1, JANUARY 1, 2022 Grand SEV- 2017-0706. We are very grateful to EPI Industries (Barcelona, Spain) for its kind donation supporting our research on COVID-19. REFERENCES [1] C. S. Huertas, D. Fariña, and L. M. Lechuga, “Direct and label-free quantification of Micro-RNA-181a at attomolar level in complex media using a nanophotonic biosensor,” ACS Sensors, vol. 1, no. 6, pp. 748–756, 2016. [2] M. Soler, C. S. Huertas, and L. M. Lechuga, “Label-free plasmonic biosensors for point-of-care diagnostics: A review,” Expert Rev. Mol. Diagn., vol. 19, no. 1, pp. 71–81, 2019. [3] B. Chocarro-Ruiz et al., “Interferometric nanoimmunosensor for label- free and real-time monitoring of Irgarol 1051 in seawater,” Biosensors Bioelectron., vol. 117, pp. 47–52, 2018. [4] P. Kozma, F. Kehl, E. Ehrentreich-Förster, C. Stamm, and F. F. Bier, “Inte- grated planar optical waveguide interferometer biosensors: A comparative review,” Biosensors Bioelectron., vol. 58, pp. 287–307, 2014. [5] D. P. Campbell, “Interferometric biosensors,” in Princ. of Bacterial Detec- tion: Biosensors, Recognit. Receptors and Microsystems. New York, NY, USA: Springer, 2008, pp. 169–211. [6] A. Brandenburg, R. Krauter, C. Künzel, M. Stefan, and H. Schulte, “In- terferometric sensor for detection of surface-bound bioreactions,” Appl. Opt., vol. 39, no. 34, pp. 6396–6405, 2000. [7] R. G. Heideman, R. P. Kooyman, and J. Greve, “Performance of a highly sensitive optical waveguide Mach-Zehnder interferometer im- munosensor,” Sensors Actuators: B. Chem., vol. 10, no. 3, pp. 209–217, 1993. [8] F. Prieto et al., “An integrated optical interferometric nanodevice based on silicon technology for biosensor applications,” Nanotechnol., vol. 14, pp. 907–912, 2003. [9] B. H. Schneider, J. G. Edwards, and N. F. Hartman, “Hartman interferome- Fig. 5. Real-time sensorgrams showing (A) the immobilization of 50 µg/mL ter: Versatile integrated optic sensor for label- free, real-time quantification anti-S chAb, and (B) the following detection of different S1 subunit concen- of nucleic acids, proteins, and pathogens,” Clin. Chem., vol. 43, no. 9, trations (50 - 1000 ng/mL) in comparison to the control protein (1000 ng/mL pp. 1757–1763, 1997. CRP). [10] B. H. Schneider, E. L. Dickinson, M. D. Vach, J. V. Hoijer, and L. V. Howard, “Highly sensitive optical chip immunoassays in human serum,” Biosensors Bioelectron., vol. 15, no. 1-2, pp. 13–22, 2000. the new algorithm to identify and quantify SARS-CoV-2 spike [11] A. Brandenburg and R. Henninger, “Integrated optical young interferom- protein in less than 10 min with high sensitivity and specificity. eter,” Appl. Opt., vol. 33, no. 25, pp. 5941–5947, 1994. [12] A. Ymeti, J. S. Kanger, R. Wijn, P. V. Lambeck, and J. Greve, “Develop- ment of a multichannel integrated interferometer immunosensor,” Sensors III. CONCLUSION Actuators, B: Chem., vol. 83, no. 1-3, pp. 1–7, 2002. [13] K. Schmitt, B. Schirmer, C. Hoffmann, A. Brandenburg, and P. Meyrueis, A new algorithm has been implemented with a custom-made “Interferometric biosensor based on planar optical waveguide sensor chips for label-free detection of surface bound bioreactions,” Biosensors Bio- software application in order to retrieve the phase difference electron., vol. 22, no. 11, pp. 2591–2597, 2007. from the real-time recorded data. This method based on simple [14] I. Barth, D. Conteduca, C. Reardon, S. Johnson, and T. F. Krauss, trigonometry was proven to provide the same results as previ- “Common-path interferometric label-free protein sensing with resonant dielectric nanostructures,” Light: Sci. Appl., vol. 9, no. 96, 2020. ously employed techniques while also reducing the sensor noise [15] R. Magnusson, K. J. Lee, and D. Wawro, “Guided-mode resonance biosen- as well as the time and expertise needed to employ the biosensor, sors employing phase detection,” in Diffractive Opt. Micro-Opt., 2004, making the BiMW platform more user-friendly. In addition, the Paper DTuC 2. [16] K. A. Murphy, M. S. Miller, A. M. Vengsarkar, and R. O. Claus, “Elliptical- algorithm can be easily applied to any interferometric device core two-mode optical-fiber sensor implementation methods,” J. Lightw. whose output signal behaves similarly to the SR described, Technol., vol. 8, no. 11, pp. 1688–1696, 1990. without increasing the intricacy of design nor fabrication pro- [17] C. D. Horna, K. E. Zinoviev, and L. M. Lechuga, “Interferometer and sensor based on bimodal optical waveguides, and detection method,” U.S. tocols. The aforementioned advances have pushed the BiMW Patent no. 8279445-B2, 2007. interferometric biosensor much closer to a user-friendly com- [18] M. Spajer, “Linear phase detection for a bimodal fiber sensor,” Opt. Lett., petitive tool, increasing its potential to be assembled into a truly vol. 13, no. 3, p. 239, 1988. [19] A. M. Vengsarkar, W. C. Michie, B. Culshaw, L. Jankovic, and R. O. POC device. Claus, “Fiber-optic dual-technique sensor for simultaneous measurement of strain and temperature,” J. Lightw. Technol., vol. 12, no. 1, pp. 170–177, 1994. ACKNOWLEDGEMENT [20] J. Canning and A. L. G. Carter, “Modal interferometer for in situ measure- ments of induced core index change in optical fibers,” Opt. Lett., vol. 22, This work was supported by the H2020 Research and In- no. 8, p. 561, 1997. novation Programme of the European Commission (CONVAT [21] R. X. Gao, Q. Wang, F. Zhao, B. Meng, and S. L. Qu, “Optimal design and project, H2020-SC1-PHE- CORONAVIRUS-2020) under Grant fabrication of SMS fiber temperature sensor for liquid,” Opt. Commun., vol. 283, no. 16, pp. 3149–3152, 2010. 101003544. The ICN2 was supported by the CERCA pro- [22] R. Poozesh et al., “Experimental study and simulation of modular in- gramme of the Generalitat de Catalunya and by the Severo terference in two-mode fiber,” Opt. Spectrosc. (English Transl. Optika i Ochoa Centres of Excellence Program, funded by the AEI under Spektroskopiya), vol. 111, no. 2, pp. 284–286, 2011. BASSOLS-CORNUDELLA et al.: NOVEL SENSING ALGORITHM FOR LINEAR READ-OUT 243 [23] A. B. González-Guerrero, C. Domínguez, K. E. Zinoviev, and L. M. Bernat Bassols-Cornudella was born in Barcelona, Spain, in 1997. He received Lechuga, “Integrated bimodal waveguide interferometric biosensor for the B.Sc. degree in mathematics and the B.Sc. degree in physics from the label-free analysis,” J. Lightw. Technol., vol. 29, no. 13, pp. 1926–1930, Autonomous University of Barcelona, Bellaterra, Spain, in 2020, and the M.Sc. 2011. degree in applied mathematics from Imperial College London, London, U.K., [24] U. Minoni, E. Sardini, E. Gelmini, F. Docchio, and D. Marioli, “A high- in 2021, where he began his studies of random processes. He is currently frequency sinusoidal phase-modulation interferometer using an electro- working toward the Ph.D. degree in mathematics of random systems in a optic modulator: Development and evaluation,” Rev. Sci. Instruments, joint CDT Program between Imperial College London and the University of vol. 62, no. 11, pp. 2579–2583, 1991. Oxford, U.K. In 2019, he joined Nanobiosensors and Bioanalytical Applications [25] B. Maisenhölder, H. P. Zappe, R. E. Kunz, P. Riel, M. Moser, and J. Group (NanoB2A), where he specialised in integrated photonics and highly Edlinger, “A GaAs/AlGaAs-based refractometer platform for integrated sensitive optical biosensors, focusing on their mathematical modeling and signal optical sensing applications,” Sensors Actuators, B: Chem., vol. 39, no. 1-3, processing algorithms. His main research focuses on the birth of chaos in random pp. 324–329, 1997. dynamical systems and their application to biological and ecological processes. [26] P. V. Lambeck, “Remote opto-chemical sensing with extreme sensitivity: Design, fabrication and performance of a pigtailed integrated optical phase-modulated mach-zehnder interferometer system,” Sensors Actua- tors, B: Chem., vol. 61, no. 1, pp. 100–127, 1999. [27] B. Sepúlveda, G. Armelles, and L. Lechuga, “Magneto-optical phase modulation in integrated Mach-Zehnder interferometric sensors,” Sensors Patricia Ramirez-Priego was born in Vilafranca del Penedès, Barcelona, Spain, Actuators A: Phys., vol. 134, no. 2, pp. 339–347, 2007. in 1993. She received the B.Sc. degree in biotechnology and the M.Sc. degree in [28] V. M. N. Passaro, F. Magno, and A. V. Tsarev, “Investigation of thermo- omics data analysis from the University of Vic - Central University of Catalonia, optic effect and multi-reflector tunable filter/multiplexer in SOI waveg- Spain, in 2015 and 2016, respectively, and the Ph.D. degree in biotechnology uides,” Opt. Exp., vol. 13, no. 9, p. 3429, 2005. from the Autonomous University of Barcelona, Bellaterra, Spain, in 2021, fo- [29] A. Hamori, K. Cottier, and R. Horvath, “Grating coupled interferometry cused on the development of a low-cost point-of-care test for active Tuberculosis for optical sensing,” Appl. Phys. B, vol. 97, no. 1, pp. 5–8, 2009. detection in human urine samples using a nanophotonic platform. Since 2015, [30] A. Dér, S. Valkai, A. Mathesz, I. Andó, E. K. Wolff, and P. Ormos, “Protein- she has been with the Nanobiosensors and Bioanalytical Applications Group based all-optical sensor device,” Sensors Actuators B: Chem., vol. 151, (NanoB2A), Catalan Institute of Nanoscience and Nanotechnology (ICN2), no. 1, pp. 26–29, 2010. where she is currently a Postdoctoral Researcher. Her main research interests [31] T. Pinheiro-Ortega, E. Silvestre, P. Andrés, B. Maes, and P. Bienst- include the development of silicon photonic biosensors, mainly focusing on their man, “Design of bimodal PCFs for interferometric gas sensors with biofunctionalization and their final validation with the evaluation of real samples high sensitivity,” IEEE Sensors J., vol. 10, no. 7, pp. 1180–1184, to demonstrate their potential in multiple applications ranging from biomedical Jul. 2010. applications to environmental monitoring. [32] R. Halir, L. Vivien, X. Le Roux, D.-X. Xu, and P. Cheben, “Direct and sensitive phase readout for integrated waveguide sensors,” IEEE Photon. J., vol. 5, no. 4, pp. 6800 906–6800906, Aug. 2013. [33] M. Kitsara, K. Misiakos, I. Raptis, and E. Makarona, “Integrated optical Maria Soler was born in 1987 in Valencia, Spain. She received the Degree frequency-resolved Mach-Zehnder interferometers for label-free affinity in chemistry from the University of Valencia, Valencia, Spain, in 2010 and sensing,” Opt. Exp., vol. 18, no. 8, pp. 8193–8206, 2010. the Ph.D. degree in biochemistry, molecular biology, and biomedicine from [34] M. L. Notte and V. M. Passaro, “Ultra high sensitivity chemical photonic the Autonomous University of Barcelona, Bellaterra, Spain, in 2015, with sensing by Mach-Zehnder interferometer enhanced vernier-effect,” Sen- international honor mentions. In 2011, she joined the Laboratory of Prof. Laura sors Actuators, B: Chem., vol. 176, pp. 994–1007, 2013. M. Lechuga with the Catalan Institute of Nanoscience and Nanotechnology [35] T. Hutter, S. R. Elliott, and S. Ruschin, “Dynamic range enhancement (ICN2), Barcelona, Spain, for her master and doctorate studies. After a 3-year and phase-ambiguity elimination in wavelength-interrogated interfero- postdoctoral stage with the Ecole Polytechnique Federale de Lausanne, EPFL, metric sensor,” Sensors Actuators, B: Chem., vol. 178, pp. 593–597, Switzerland, since 2018, she is a Senior Researcher of the Nanobiosensors 2013. and Bioanalytical Applications Group (NanoB2A), ICN2. To date, she has [36] K. Misiakos et al., “Broad-band Mach-Zehnder interferometers as high authored or coauthored about 20 peer-reviewed articles in top-ranked journals, performance refractive index sensors: Theory and monolithic implemen- with more than 800 citations. Her research interests include the design and tation,” Opt. Exp., vol. 22, no. 8, pp. 8856–8870, 2014. development of novel nanophotonic biosensors for biomedical applications, [37] S. Dante, D. Duval, D. Fariña, A. B. González-Guerrero, and L. M. with a special focus in sensor biofunctionalization methods and technology Lechuga, “Linear readout of integrated interferometric biosensors using validation for point-of-care clinical diagnostics and novel immunoengineering a periodic wavelength modulation,” Laser Photon. Rev., vol. 9, no. 2, therapeutic studies. She has also been invited as keynote and plenary speaker in pp. 248–255, 2015. international conferences and actively contributes to scientific dissemination in [38] D. Duval et al., “Nanophotonic lab-on-a-chip platforms including novel outreach events and publications. bimodal interferometers, microfluidics and grating couplers,” Lab a Chip, vol. 12, no. 11, pp. 1987–1994, 2012. [39] S. Dante, D. Duval, B. Sepúlveda, A. B. González-Guerrero, J. R. Sendra, and L. M. Lechuga, “All-optical phase modulation for integrated interferometric biosensors,” Opt. Exp., vol. 20, no. 7, pp. 7195–7205, M.-Carmen Estévez was born in Barcelona, Spain. She received the Bachelor 2012. of Chemistry degree from the University of Barcelona, Spain, in 1998 and the [40] G. Lifante, Integrated Photonics: Fundamentals. Chichester, U.K.: Wiley, M.Sc. and Ph.D. degrees in chemistry in 1999 and 2005, respectively, focused on Ltd, 2003. the development of immunochemical techniques for environmental monitoring. [41] D. G. García, D. Jiménez, J. R. Sendra, and L. M. Lechuga, “Lab-on-a- She got a Postdoctoral Fellowship from the Government of Catalonia (Beatriu chip integration of the bimodal waveguide nanointerferometric biosensor,” Pinós Grant) to perform a stay at the University of Florida (Dept. Chemistry, UF) Ph.D. dissertation, Universitat Autònoma de Barcelona: Barcelona, Spain, between 2006 and 2008. She focused on nanobiotechnology applied to in-vitro 2018. diagnostics, through the production and application of fluorescent nanoparticles [42] O. Sasaki and H. Okazaki, “Sinusoidal phase modulating interferometry for bioanalysis. In 2009, she joined the Nanobiosensors and Bioanalytical for surface profile measurement,” Appl. Opt., vol. 25, no. 18, pp. 3137– Applications Group, Institut Català de Nanociència i Nanotecnologia (ICN2) 3140, 1986. and CSIC, Barcelona, Spain, as a Senior Researcher, where she focused on [43] P. Ramirez-Priego, M.-C. Estévez, H. J. Díaz-Luisravelo, J. J. Manclús, surface biofunctionalization and clinical diagnostics applications of label-free Á. Montoya, and L. M. Lechuga, “Real-time monitoring of fenitrothion in optical biosensors. She has coauthored more than 40 papers in leading journals water samples using a silicon nanophotonic biosensor,” Analytica Chimica in nanotechnology, analytical chemistry, biosensors and optics. She is an Active Acta, vol. 1152, 2021, Art. no. 338276. Reviewer in ACS, Elsevier, RSC, and Wiley journals. Her main research interests [44] J. Maldonado, M. C. Estévez, A. Fernández-Gavela, J. J. González-López, include plasmonic and silicon photonics sensors, their integration in compact A. B. González-Guerrero, and L. M. Lechuga, “Label-free detection prototypes and the development of biosensor based applications to demonstrate of nosocomial bacteria using a nanophotonic interferometric biosensor,” their potential for real clinical, industrial, and/or environmental applications, Analyst, vol. 145, pp. 497–506, 2020. aiming at transferring them to decentralized settings. 244 JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 40, NO. 1, JANUARY 1, 2022 Heriberto J. Díaz Luis-Ravelo was born in Santa Cruz de Tenerife, Spain, in Laura M. Lechuga is currently a Full Professor with Spanish National Re- 1988. He received the B.S. degree in telecommunications engineer, specialized in search Council (CSIC) and the Head of the Nanobiosensors and Bioanalytical electronic systems, from Las Palmas de Gran Canaria University, in 2015 and the Applications Group, Catalan Institute of Nanoscience and Nanotechnology M.S. degree in telecommunications engineer from Las Palmas de Gran Canaria (ICN2), Barcelona, Spain and with the Networking Biomedical Research Center University, in 2017. He joined the Nanobiosensors and Bioanalytical Applica- (CIBER-BBN). The principal focus of her research is the development of novel tions Group (NanoB2A), Catalan Institute of Nanoscience and Nanotechnology nanobiosensor devices based on nanoplasmonics and silicon-based photonics (ICN2), Barcelona, Spain, as a Research Engineer from 2017 to 2021. He worked principles for point-of-care diagnostics. She has been at the forefront of the Pho- in nanoplasmonic and bimodal waveguide interferometric biosensors, in which tonic Biosensor area during more than eighteen years, making key contributions he worked on the development to control and monitor biosensors working on and opening new horizons in this field. She has authored or coauthored more than both hardware and software. Since 2021, he has been a Lead Software Engineer 260 articles, book chapters and proceedings, has eight families of patents, and with QUSIDE Technology, Barcelona, Spain. He is working on the development has presented her work worldwide in more than 365 invited talks. Her research to control and monitor the quantum random number generator (QRNG). activities have encompassed from fundamental research to the technological operation of complete sensing platforms, including the technological transfer into products of social applicability. Her Group is considered as a world reference in this field. She has co-founded two spin-offs companies. The quality of her research has been recognised by prestigious prizes and awards, as the Physics, Mari Cardenosa-Rubio is currently a PROBIST Postdoctoral Fellow under the Innovation and Technology Prize from the Spanish Royal Physics Society Marie Skłodowska-Curie Agreement No. 754510 with the Catalan Institute of (RSEF) and BBVA Foundation (2016), the Ada Byron Prize (2020), the King Nanoscience and Nanotechnology, Barcelona, Spain. She did her undergraduate Jaume I Award in New Technologies and the Spanish National Research Prize studies in biotechnology with the University of Salamanca, Salamanca, Spain. in Tech Transfer in 2020 (the two last ones are the most prestigious research After that, she moved to Belgium and France to pursue the Erasmus Mundus prizes in Spain), and the Burdinola Research Award in 2021. In 2020, during Master in nanosciences and nanotechnology. Then, she got a fellowship from La the COVID-19 pandemic, she has been included in the Expert Scientific Panel Caixa to pursue post-graduate studies in North America. This fellowship allowed advising the Ministry of Science and Innovation and the Spanish Government her to carry out her Ph.D. in chemistry with the University of Michigan, Ann in the management of the pandemic. She is a Fellow of OSA. Arbor, MI, USA. During this stage, she worked in silicon photonic microring resonators for the detection of noncoding RNAs and polyphosphates in complex matrices. In the NanoB2A, her current work is based on the development of protocols for the multiplex detection of biomarkers in clinical settings using the silicon bimodal waveguides.