Improving Limit-of-Detection in High-Sensitivity Plasmo-Photonic Mach-Zehnder Interferometer Refractive Index Sensors

IEEE PHOTONICS JOURNAL, VOL. 17, NO. 5, OCTOBER 2025 4800107 Improving Limit-of-Detection in High-Sensitivity Plasmo-Photonic Mach-Zehnder Interferometer Refractive Index Sensors Stelios Simos , Lamprini Damakoudi, Dimosthenis Spasopoulos, Konstantinos Fotiadis , Evaggelia Chatzianagnostou , Jose Carreira, Gabriele Navickaite, Michael Geiselmann, Juan Arocas, Jean-Claude Weeber, Dimitris. V. Bellas, Eleftheria Lampadariou, Elefterios Lidorikis, Konstantinos Vyrsokinos , and Nikos Pleros Abstract—This work presents a dual plasmo-photonic branch Mach-Zehnder interferometer refractive index sensor integrated on a Si3N4platform, with both the sensor and reference arms hosting identical aluminum plasmonic metal stripes. Experimental evaluation of the configuration reveals a bulk sensitivity up to 8801 nm/RIU and an increased environmental noise resilience compared to state-of-the-art plasmo-photonic MZI sensors with the plasmonic section residing only on the sensor arm. The enhanced noise resilience along with the high sensitivity results in an experimentally obtained low detection limit value of 4.4 ×10−6RIU, im-proved by an order of magnitude compared to the detection limit of 5 ×10−5RIU that was measured in a reference MZI sensor with a plasmonic stripe only at its sensor arm. Index Terms—Aluminum, MZI, Photonic integrated circuits, plasmonic sensors, refractive index, SiN. I. INTRODUCTION PHOTONIC and plasmonic sensors are advanced optical sensing technologies that leverage light-matter interaction for highly sensitive detection of biological, chemical, and environmental applications [1]. Photonic sensors mainly rely on Received 24 July 2025; accepted 4 August 2025. Date of publication 11 August 2025; date of current version 22 August 2025. This work was supported by the European project AMBROSIA under Grant 101093166. (Corresponding author: Stelios Simos.) Stelios Simos, Konstantinos Fotiadis, Evaggelia Chatzianagnostou, and Nikos Pleros are with the Department of Informatics, Aristotle University of Thessaloniki,54124Thessaloniki,Greece,andalsowiththeCenter for Interdisciplinary Research and Innovation (CIRI-AUTH), Balkan Center, Buildings A & B, 57001 Thessaloniki, Greece (e-mail: [email protected]). Lamprini Damakoudi and Konstantinos Vyrsokinos are with the Center for Interdisciplinary Research and Innovation (CIRI-AUTH), Balkan Center, Buildings A & B, 57001 Thessaloniki, Greece, and also with the Department of Physics, Aristotle University of Thessaloniki, 54124 Thessaloniki, Greece. Dimosthenis Spasopoulos is with the Department of Informatics, Aristotle University of Thessaloniki, 54124 Thessaloniki, Greece, also with the Center for Interdisciplinary Research and Innovation (CIRI-AUTH), Balkan Center, Buildings A & B, 57001 Thessaloniki, Greece, and also with the EOS Technologies, OX4 1 JB Oxford, U.K.. Jose Carreira, Gabriele Navickaite, and Michael Geiselmann are with the LIGENTEC SA, EPFL Innovation Park, CH-1024 Ecublens, Switzerland. Juan Arocas and Jean-Claude Weeber are with the Photonic Department, ICB UMR 6303, Université de Bourgogne, 21078 Dijon, France. Dimitris. V. Bellas, Eleftheria Lampadariou, and Elefterios Lidorikis are with the Department of Materials Science and Engineering, University of Ioannina, 45110 Ioannina, Greece. Digital Object Identifier 10.1109/JPHOT.2025.3597420 waveguides [2],[3], ring resonators [4],[5],[6], or MachZehnder interferometers (MZI) [7],[9] to facilitate light-matter interaction and sensing, offering certain advantages such as high sensitivity [10], label-free detection [11] and mass-production credentials through the use of CMOS technology. However, photonic sensor configurations mainly utilize evanescent wave sensing and require rather lengthy structures in order to achieve high sensitivity values. Plasmonic sensors, on the other hand, can have even the whole optical mode interacting with matter by exploiting surface plasmon resonances (SPRs) in metallic nanostructures or surface plasmon polaritons (SPPs) in metaldielectric interfaces. This allows for ultra-sensitive detection at the nanoscale [12],[13] with real-time monitoring capabilities [14], suffering, however, from higher optical losses and bulky I/O structures. The loss and I/O interface drawbacks can be overcome by converging plasmonics with photonics into plasmo-photonic sensor configurations. Merging the best of both worlds can combine the low-loss passive circuitry from the photonic domain, like interferometry, optical I/O interfaces, filtering, phase shifting and power balancing, with the ultra-high sensitivity properties of miniaturized plasmonic structures, paving the way for the realization of compact and powerful sensor configurations. Various plasmo-photonic architectures using resonators or interferometric structures have been proposed such as ring resonators [15],[17], bimodal interferometers [18],[19] and MZIs [20],[24]. The MZI configuration probably offers the most powerful plasmo-photonic sensor layout, as it offers a high degree of parametrization together with experimentally reported sensitivity values up to 4764 nm/RIU [23] and a simulated and theoretically formulated potential for up to 60000 nm/RIU [22], [25]. Almost all plasmo-photonic MZI architectures reported so far employ non-identical waveguide branches, [21] resulting to a differential phase change between the two branches even when the same environmental change is experienced by both arms. This increases its sensitivity to environmental perturbations and obviously degrades its Limit-of-Detection (LoD) performance characteristics. In this paper, we present experimentally a balanced plasmophotonic MZI refractive index sensor that employs identical plasmonic waveguides at both its reference and sensor arms, © 2025 The Authors. This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ 4800107 IEEE PHOTONICS JOURNAL, VOL. 17, NO. 5, OCTOBER 2025 Fig. 1. (a) Conceptual 3D design of the dual plasmonic MZI sensor, (b) cross-section of the tapered Si3N4waveguide, (c) cross-section of the aluminum plasmonic stripe, (d) microscope photo of the fabricated chip and (e) simulated transmission values concerning the vertical offset. allowing for both high bulk sensitivity and improved LoD performance. We extend our previous work reported in [26] and validate experimentally its improved LoD performance against state-of-the-art imbalanced plasmo-photonic MZI sensor architectures that use only a single plasmonic waveguide at their sensing arm. The proposed balanced MZI device is integrated onaSi 3N4platform with an identical aluminum (Al) stripe placed on both arms of the MZI. Experimental evaluation of the sensitivity performed with different water-based solutions reveals a high bulk sensitivity of 8801 nm/RIU, increased by almost 84% compared to state-of-the-art plasmo-photonic MZI sensors [23]. Its noise resiliency is experimentally compared against an optimized reference plasmo-photonic MZI sensor that has a single plasmonic stripe at its sensor arm and an even higher sensitivity of 12,093 nm/RIU, revealing a LoD of 4.4 ×10−6 that is improved by one order of magnitude compared to the LoD of 5 ×10−5measured by the reference MZI sensor. Given the significantly improved noise resiliency and LoD performance of the proposed sensor configuration even when operating as a bare die without any thermal stabilization mechanism, it is expected that this could allow for LoD values well below 10−7 when produced in a packaged and thermally stabilized setup. These enhancements will reduce potential external noise factors, leading to a decreased noise standard deviation improving the LoD of the proposed sensor. II. SENSOR LAYOUT AND PRINCIPLE OF OPERATION A 3D schematic of the proposed sensor layout is illustrated in Fig. 1(a).Fig.1(b) shows the cross-section of the tapered Si3N4photonic waveguide along with the dimensions of the materials used, while Fig. 1(c) present the Aluminum (Al) plasmonic waveguide part of the configuration. Fig. 1(d) depicts a microscope photo of the fabricated chip on which several variations of the proposed configuration are implemented. The proposed sensor architecture comprises a plasmo-photonic MZI integrated on a Si3N4platform from Ligentec. The MZI consists of 800 ×800 nm Si3N4waveguides. A 1 ×2 Multi-Mode Interference (MMI) splitter was used to split the input light equally between the two branches of the MZI. Tapered Si3N4 waveguides with a size of 7.5 μm×800 nm were used to achieve the coupling of the photonic modes with the plasmonic modes of the Al stripes that were implemented at the two MZI branches. Two identical Al stripes were deposited in respective etched cavities with a Vertical Offset (V.O) of 350 nm. The thickness and width of the metal stripes were set to 80 nm and 7μm, respectively. The plasmonic length of the stripes (Lplasm) was set to 70 μm, while longer plasmonic stripes can lead to higher sensitivities the losses of the device will also increase. The 70 μm of plasmonic length provide a high sensitivity value while the losses of the device are within reasonable limits. An inverted taper was used to couple the plasmonic mode back into a photonic mode, so that the two spatially separated light beams propagate through the Si3N4waveguides of the two MZI branches and are forced to interfere via a 2 ×2 MMI coupler. A photonic differential length of ΔL was added in one of the two MZI arms to define the Free Spectral Range (FSR) value of the interferometer. Identical thermo-optical phase shifters were also employed at the two MZI branches in order to spectrally tune the interferometer resonance within a detectable spectral window. One of the two identical Al plasmonic stripes serve as the sensor arm, where different liquid analytes are used to cover the plasmonic region and the SPP mode is entirely exposed to the overlying analyte. The second Al stripe serves as the reference arm and is always covered by the same aqueous solution throughout the entire measurement process, so that the guided SPP mode is constantly exposed to the same surrounding medium. Hence, any difference in the constitution between the liquid analytes that cover the sensor and reference arms, respectively, will result in different refractive indices for the two analytes and will be translated into a differential phase change for the two SPP modes guided at the two branches. This will SIMOS et al.: IMPROVING LIMIT-OF-DETECTION IN HIGH-SENSITIVITY PLASMO-PHOTONIC MACH-ZEHNDER 4800107 TABLE I SIMULATED FSR AND SENSITIVITY VALUES FOR DIFFERENT ΔL cause a MZI resonance wavelength shift that is directly related to the refractive index difference between the two MZI branches. Taking into account that the environmental conditions can be considered identical at the two MZI branches, a noise resilient performance requires that the differential phase change and as such the resonance wavelength shift remains constant as long as the two liquid analytes remain the same, irrespective of any environmental changes. III. DESIGN ANALYSIS A design analysis has been carried out for optimizing sensor performance with respect to its bulk sensitivity and noise resiliency performance using the FDE, 3D FDTD and INTERCONNECT modelling tools of the Ansys Lumerical software suite. Fig. 1(b) and (c) depict the cross-sections of the Si3N4 photonic and plasmonic waveguides, respectively. We introduce the Vertical Offset (V.O.) parameter that corresponds to the height of the bottom oxide layer that needs to be etched away to form the cavity where the plasmonic waveguide is deposited in order to achieve best interface coupling. Fig. 1(e) illustrates the power coupling between the photonic and plasmonic waveguide for a V.O. ranging from 0 nm up to 1000 nm, showing that a maximum transmission of ∼0.4 is observed at a V.O. of 350 nm. This transmission value was then employed as the interface loss in a MZI circuit level simulation analysis, where also different photonic ΔL values for the reference waveguide arm were utilized to allow for MZI sensors with different FSR values. Taking into account the well-known dependence of the MZI sensor sensitivity on its FSR value [25], a selection of rather small ΔL values ranging from 5–16 um was used leading to FSR values between 70 and 186 nm and to respective sensitivity values ranging from 3037 to 9323 nm/RIU, as shown in Table I. The sensitivity performance for each different ΔL was captured via the circuit level simulations by applying different refractive indices on the sensor arm while retaining a constant refractive index value the at the reference arm. A similar procedure has been followed for modelling also the reference plasmo-photonic MZI sensor that contains only a single plasmonic stripe placed at its sensing arm, with its reference arm employing a photonic waveguide. The same V.O. of 350 nm was assumed also here for optimal coupling between the photonic and plasmonic waveguide sections. Considering the relationship between sensitivity and FSR, a FSR value of 310 nm was opted in order to allow for a fair comparison with the balanced MZI sensor with the highest sensitivity, leading to aΔL value of 20.4 μm. IV. FABRICATION The plasmonic sensors were fabricated on Ligentec’s AN800 (all nitride 800 nm thick Si3N4) platform [27]. The photonic stackwasfabricatedon100mmSiwaferandconsistsofabottom thermal silicon oxide, 800 nm thick Si3N4, and a silicon dioxide top cladding. In addition, aluminum-based heaters are integrated in the stack. Local openings above the pads are defined allowing for electrical probing. The plasmonic sensor is defined in two processes: first, a local opening of the cladding is etched down to the target V.O. (350 nm below the SiN bottom level). Next, an aluminum stripe is defined at the bottom of the cavity using a lift-off technique. The lift-off process prevents aluminum from depositing on the cavity sidewall, which would contribute to coupling losses at the photonic-plasmonic interface. The lift-off pattern was defined via stepper lithography, as it offers high alignment accuracy (reducing die-to-die variation) and highthroughput, paving the path towards scalability. In view of CMOS compatibility aluminum is the primary metal of interest in this study. However, it is worth to note that other plasmonic metals can be deposited as well on the patterned chips featuring the photonic circuitry and local opening in the cladding. In this context an overlayed PMMA-based electron-beam lithography process exploiting alignment marks etched in the cladding has been successfully demonstrated. In particular, gold and aluminum stripes have been deposited by lift-off process at the bottom surface of sensing areas. This versatile chip-level process allows for a positioning of metal stripes ends with respect to input and output Si3N4waveguides with an accuracy down below 100 nm in spite of the electron-beam exposure being performed in a several micron-deep openings. A close-up image and a SEM image of the fabricated sensor are presented in Fig. 2(a) and (b) respectively. A total of 11 dies originate from different wafers have been evaluated and present a fabrication yield of 100%. V. EXPERIMENTAL RESULTS The experimental setup used to characterize and validate the proposed configuration is shown in Fig. 3(a). Two Santec Tunable Semiconductor Lasers (TSL) 570 have been utilized. One of the TSLs operates in the range of 1355 nm–1485 nm while the second TSL operates in the range between 1480 nm to 1640 nm. The TSL outputs are connected to a Santec Optical Switch Unit (OSU) via a Polarization Maintaining Fiber (PMF) in order to allow tunability over a spectral window of 285 nm. A Lensed PMF (LPMF) was used at the output of the OSU. The LPMF is placed on a 562 Newport XYZ stage along with a fiber rotator to align the input LPMF with the Spot Size Converters (SSC) of the chip. The fiber rotator is used to ensure that TM polarized light is injected into the chip. A Peltier module is placed together with a thermistor below the DUT holder to control the temperature on the chip. The Thermo-Electric Controller (TEC) is controlled by an ILX LDT 5910C thermoelectric temperature controller. A Single Mode Fiber (SMF) and a second XYZ stage were used to align the SMF to the output SSC of the chips and collect the light. The SMF is connected to a Santec MPM 210 optical power meter (PM). The experimental set up is placed on a vibration-isolation 4800107 IEEE PHOTONICS JOURNAL, VOL. 17, NO. 5, OCTOBER 2025 Fig. 2. (a) Close-up image of the sensing area (SiN ph-WG: photonic waveguide and Al pl-WG: aluminum plasmonic waveguide), (b) SEM image of the fabricated sensor. table along with a humidity sensor located close to the chip to measure the environmental conditions. Fig. 3(b) depicts the spectral response of the dual plasmonic MZI sensor configuration when both plasmonic stripes are exposed to air as the surrounding medium. The measured FSR of the detected spectrum is 180.3 nm, which is in good agreement with the 186 nm FSR value (186 nm) predicted by the simulations. A high extinction ratio (ER) of up to 22.5 dB was also measured. In the same structure, we apply voltage to the thermo-optic phase shifters (PS) located on the arms of the MZI to characterize their performance. The PS allow us to shift the resonance of the structure into the region of our optical window. By using the PS, we can use configurations with high FSR values and thus avoid the problem of the narrow optical window. Fig. 3(c) shows the spectral response of the MZI when different voltage levels are applied at its phase shifter (PS). As can be observed, sweeping the voltage level of the PS forces the MZI resonance to tune, allowing inthis waytobeing the MZI within adetectable spectral window even when MZIs with high FSR values are employed. A 2πphase shift was obtained by applying power equal to 312 mW. R.I. experiments were subsequently performed in order to evaluate the bulk sensitivity of the proposed sensor configuration. A droplet of distilled water was placed on the plasmonic stripe of the reference arm of the MZI configuration and served as a reference liquid. In parallel, another droplet of the same distilled water was also placed on the plasmonic stripe of the sensor arm, so that both the reference and sensor arm host the TABLE II EXPERIMENTAL FSR AND SENSITIVITY VALUES FOR DIFFERENT ΔL same overlying liquid medium. As a result, the two spatially separated optical beams experience again the same optical paths, as was also the case when both arms were exposed to air environment, allowing in this way to have an unchanged differential phase change and the same spectral location of the MZI resonance. The sensitivity performance was evaluated by using then different water-based solutions with different refractive indices ranging between 1.3322–1.3411, which were used to cover the sensor arm while the reference arm was constantly covered with pure distilled water. The obtained spectral responses and sensitivity curves for 3 different MZI sensor structures are shown in Fig. 4(a)–(f).Fig.4(a) and (b) illustrate the acquired spectra for different aqueous solutions and the sensitivity curve, respectively, for the MZI 1 structure that has an expected FSR value of 186 nm, revealing a bulk sensitivity of 8801 nm/RIU with a fitted R-squared of 0.99. The spectra of MZI 2 that has an experimentally measured FSR of 99.7 nm are depicted in Fig. 4(c), while the achieved sensitivity of 3248 nm/RIU is illustrated in Fig. 4(d). The spectra of MZI 3 that has an experimentally measured FSR of 63.8 nm and the respective sensitivity curve are shown in Fig. 4(e) and (f), respectively, declaring a sensitivity of 2603 nm/RIU. The experimental results for the three different MZI structures are summarized in Table II. The measured bulk sensitivity values are in good agreement with the corresponding sensitivity values predicted by simulations and presented in Table I.Thisisalso verified in Fig. 5, which shows the sensitivity versus FSR curve obtained both through simulations and experiments for the three MZI structures, revealing a close matching between simulations and experiments both with respect to the absolute sensitivity values as well as to the slope of the curves. An error analysis has also been performed on MZI 1 with FSR value of 186 nm. The analysis on the experimental results of the sensitivity tests presents a standard deviation of 286.1 nm/RIU and a relative error of the simulation and experimental sensitivity calculated approximately at 9.41%. To evaluate the thermal stability properties of the MZI sensor, the spectral output of the MZI 1 with an FSR close to 186 nm was constantly monitored for a total duration of 30 minutes at room temperature without using a TEC. A zoomed-in picture of 20 different recorded spectra over this 30 min timespan is shown in Fig. 6(a), revealing that almost no resonance shift takes place within this time period. The same experiment was then repeated when temperature conditions are controlled by means of a TEC placed below the photonic chip holder. Fig. 6(b) shows the recordedspectralresponseswhenthechip temperature increases gradually from 28.9 to 42.3 °C, depicting again a minor SIMOS et al.: IMPROVING LIMIT-OF-DETECTION IN HIGH-SENSITIVITY PLASMO-PHOTONIC MACH-ZEHNDER 4800107 Fig. 3. (a) Experimental set up used to evaluate the fabricated sensors, (b) experimental transfer function for both output ports of the MZI sensor with FSR value of 186 nm and (c) MZI sensor transfer function wavelength shift by applying voltage to thermos-optic phase shifters. Fig. 4. (a) Refractive index experiment spectrums in MZI 1 with expected FSR 186 nm, (b) captured resonance shifts and linear fit for MZI 1 spectrums, (c) Refractive index experiment spectrums in MZI 2 with expected FSR 100 nm, (d) captured resonance shifts and linear fit for MZI 2 spectrums, (e) Refractive index experiment spectrums in MZI 3 with expected FSR 70 nm and (f) captured resonance shifts and linear fit for MZI 3 spectrums. Fig. 5. Sensitivity versus FSR acquired through simulation and experimental measurements. variation in the resonance dip with a maximum wavelength shift of Δλmax =1.3 nm and a standard deviation of 0.508 nm over almost 13.5 °C temperature. This proves that the temperature variations result to almost identical thermo-optically induced phase shifts at the two MZI branches, since the MZI exploits a balanced configuration with almost identical plasmo-photonic branches, allowing for high tolerance to temperature drifts. The experimental characterization of the LoD was carried out for the MZI sensor with the highest bulk sensitivity of 8801 nm/RIU and relies on the use of the well-known formula LoD =3σsensor S(1) where σsensor is the standard deviation of the resonance and S is the bulk sensitivity of the sensor. Based on the experimental results obtained for a duration of 30 min at room temperature when no thermal control is used, 4800107 IEEE PHOTONICS JOURNAL, VOL. 17, NO. 5, OCTOBER 2025 Fig. 6. (a) Spectrum resonance over time in room temperature and (b) captured spectrums while increasing the temperature. the standard deviation of the resonance drift was calculated to be σt=0.022 nm. This measurement includes the resonance drift experienced due to thermal variations but also includes the drifts due to fiber I/O coupling vibrations and additional noise factors that originate from the characterization setup of the non-packaged device. In order to remove these additional noise components that would not be present in case a fiber-pigtailed and packaged device is employed, a similar spectral response monitoring process was repeated for a simple straight Si3N4 waveguide by measuring the spectral shift of the resonances generated due to the birefringence of the waveguide. The standard deviation σI/O of these spectral variations was found to be 0.018 nm and includes the noise of the fiber interface vibrations and any additional noise factors that affect the single waveguide response. Using then the formula: σ2 t=σ2 sensor +σ2 I/O (2) the noise originating from the measurement setup can be subtracted and the sensor noise is calculated to equal σsensor =0.0129 nm. By applying S =8801 nm/RIU and σsensor =0.0129 nm to (2) the LoD of the proposed configuration was calculated to be 4.4 ×10-6 RIU. The same procedure was also repeated for the reference plasmo-photonic TABLE III COMPARISON OF DUAL PLASMONIC MZI AND SINGLE PLASMONIC MZI TABLE IV PLASMO-PHOTONIC SENSORS COMPARISSON TABLE MZI configuration where only a single plasmonic stripe was employed at the MZI sensor branch. The sensitivity was found to be 12093 nm/RIU and the standard deviation of the reference MZI noise was found to be 0.2024 nm, resulting to a LoD of 5×10-5 RIU, i.e., one order of magnitude lower than the LoD obtained in the case of the balanced plasmo-photonic MZI. The trade-off for the enhancement of the LoD value for the dual plasmonic configuration is the increased fabrication complexity which leads to a higher sensitivity to fabrication errors. Table III summarizes the performance characteristics of the balanced and non-balanced reference MZI sensors. A comparison table of the state of the art for the plasmo-photonic sensors is presented in Table IV. VI. CONCLUSION We demonstrate a novel, balanced plasmo-photonic MZI refractive index sensor configuration on a Si3N4platform. The proposed configuration utilizes two identical plasmonic metal stripes one in each arm of the MZI, and the phase changes due to environmental perturbations are the same in both arms, compensating for external noise. This improves the stability of the sensor and increases its resilience to environmental perturbation. The proposed plasmo-photonic MZI configuration achieves the highest reported bulk sensitivity of up to 8801 nm/RIU and can be further improved by changing the FSR of the architecture. Additionally, the sensor has a very good LoD of 4.4 ×10−6 RIU. The architecture and the CMOS compatible materials used in the design can lead to mass production sensor chips capable of performing multiple examinations on a single die. REFERENCES [1] M. A. Butt, N. L. Kazanskiy, S. N. Khonina, G. S. Voronkov, E. P. Grakhova, and R. V. Kutluyarov, “A review on photonic sensing technologies: Status and outlook,” Biosensors, vol. 13, 2023, Art. no. 568, doi: 10.3390/bios13050568. SIMOS et al.: IMPROVING LIMIT-OF-DETECTION IN HIGH-SENSITIVITY PLASMO-PHOTONIC MACH-ZEHNDER 4800107 [2] R. Janeiro, R. Flores, and J. Viegas, “Silicon photonics waveguide array sensor for selective detection of VOCs at room temperature,” Sci. Reports, vol. 9, no. 1, Nov. 2019, Art. no. 17099, doi: 10.1038/s41598-01952264-9. [3] M. Vlk et al., “Extraordinary evanescent field confinement waveguide sensor for mid-infrared trace gas spectroscopy,” Light Sci. Appl., vol. 10, no. 1, Jan. 2021, Art. no. 26, doi: 10.1038/s41377-021-00470-4. [4] G. A. Rodriguez, S. Hu, and S. M. Weiss, “Porous silicon ring resonator for compact, high-sensitivity biosensing applications,” Opt. Exp., vol. 23, no. 6, pp. 7111–7121, Mar. 2015, doi: 10.1364/OE.23.007111. [5] H.-T. Kim and M. Yu, “Cascaded ring resonator-based temperature sensor with simultaneously enhanced sensitivity and range,” Opt. Exp., vol. 24, no. 9, pp. 9501–9510, May 2016, doi: 10.1364/OE.24.009501. [6] S. M. Lo et al., “Photonic crystal microring resonator for label-free biosensing,” Opt. Exp., vol. 25, no. 6, pp. 7046–7056, Mar. 2017, doi: 10.1364/OE.25.007046. [7] A. Bastos et al., “Integrated optical Mach-Zehnder interferometer based on organic-inorganic hybrids for photonics-on-a-chip biosensing applications,” Sensors, vol. 18, no. 3, Mar. 2018, Art. no. 840, doi: 10.3390/s18030840. [8] A. M. Taha, S. Yousuf, M. S. Dahlem, and J. Viegas, “Highly-sensitive unbalanced MZI gas sensor assisted with a temperature-reference ring resonator,” IEEE Photon. J., vol. 14, no. 6, Dec. 2022, Art. no. 6657709, doi: 10.1109/JPHOT.2022.3215713. [9] H. Niu et al., “Mach-Zehnder interferometer-based integrated-photonic acetone sensor approaching the sub-ppm level detection limit,” Opt. Exp., vol. 30, no. 16, pp. 29665–29675, Aug. 2022, doi: 10.1364/OE.459805. [10] V. Passaro et al., “Recent advances in integrated photonic sensors,” Sensors, vol. 12, no. 11, pp. 15558–15598, Nov. 2012, doi: 10.3390/ s121115558. [11] E. Luan, H. Shoman, D. M. Ratner, K. C. Cheung, and L. Chrostowski, “Silicon photonic biosensors using label-free detection,” Sensors, vol. 18, no. 10, Oct. 2018, Art. no. 3519, doi: 10.3390/s18103519. [12] T. Chung, S.-Y. Lee, E. Y. Song, H. Chun, and B. Lee, “Plasmonic nanostructures for nano-scale bio-sensing,” Sensors, vol. 11, no. 11, pp. 10907–10929, Nov. 2011, doi: 10.3390/s111110907. [13] I. Choi and Y. Choi, “Plasmonic nanosensors: Review and prospect,” IEEE J. Sel. Topics Quantum Electron., vol. 18, no. 3, pp. 1110–1121, May/Jun. 2012, doi: 10.1109/JSTQE.2011.2163386. [14] O. Limaj et al., “Infrared plasmonic biosensor for real-time and label-free monitoring of lipid membranes,” Nano Lett., vol. 16, no. 2, pp. 1502–1508, Feb. 2016, doi: 10.1021/acs.nanolett.5b05316. [15] X. Sun, D. Dai, L. Thylén, and L. Wosinski, “Double-slot hybrid plasmonic ring resonator used for optical sensors and modulators,” Photonics,vol.2, no. 4, pp. 1116–1130, Dec. 2015, doi: 10.3390/photonics2041116. [16] M. A. Butt, N. L. Kazanskiy, and S. N. Khonina, “Modal characteristics of refractive index engineered hybrid plasmonic waveguide,” IEEE Sens. J., vol. 20, no. 17, pp. 9779–9786, Sep. 2020, doi: 10.1109/JSEN.2020.2991215. [17] M. A. Butt, “Racetrack ring resonator-based on hybrid plasmonic waveguide for refractive index sensing,” Micromachines, vol. 15, no. 5, May 2024, Art. no. 610, doi: 10.3390/mi15050610. [18] K. Fotiadis et al., “High-sensitivity bimodal plasmo-photonic refractive index sensor,” ACS Photon., vol. 10, no. 8, pp. 2580–2588, Aug. 2023, doi: 10.1021/acsphotonics.3c00290. [19] K. Fotiadis et al., “Theoretical and experimental analysis of single-arm bimodal plasmo-photonic refractive index sensors,” Sensors, vol. 24, no. 12, 2024, Art. no. 3705, doi: 10.3390/s24123705. [20] X. Sun, D. Dai, L. Thylén, and L. Wosinski, “High-sensitivity liquid refractive-index sensor based on a Mach-Zehnder interferometer with a double-slot hybrid plasmonic waveguide,” Opt. Exp., vol. 23, no. 20, pp. 25688–25699, Oct. 2015, doi: 10.1364/OE.23.25688. [21] X. Sun, L. Thylén, and L. Wosinski, “Hollow hybrid plasmonic Mach– Zehnder sensor,” Opt. Lett., vol. 42, no. 4, pp. 807–810, Feb. 2017, doi: 10.1364/OL.42.000807. [22] A. Manolis et al., “Plasmonics co-integrated with silicon nitride photonics for high-sensitivity interferometric biosensing,” Opt. Exp., vol. 27, no. 12, pp. 17102–17111, Jun. 2019, doi: 10.1364/OE.27.17102. [23] A. Manolis et al., “Ultra-sensitive refractive index sensor using CMOS plasmonic transducers on silicon photonic interferometric platform,” Opt. Exp., vol. 28, no. 14, pp. 20992–21001, Jul. 2020, doi: 10.1364/OE.28.20992. [24] G. Wang and W. Feng, “On-chip Mach-Zehnder interferometer sensor with a double-slot hybrid plasmonic waveguide for high-sensitivity hydrogen detection,” Opt. Exp., vol. 31, no. 24, pp. 39500–39513, Dec. 2023, doi: 10.1364/OE.31.39500. [25] E. Chatzianagnostou et al., “Scaling the sensitivity of integrated plasmo-photonic interferometric sensors,” ACS Photon., vol. 6, no. 7, pp. 1664–1673, Jul. 2019, doi: 10.1021/acsphotonics.8b01683. [26] S. Simos et al., “Temperature-tolerant and high-sensitivity plasmophotonic Mach-Zehnder refractive index sensor,” in Proc. Conf. Lasers Electro-Opt., 2025. [27] LIGENTEC, “LIGENTEC | photonic integrated circuits manufacturing,” (n.d.). [Online]. Available: https://www.ligentec.com/