key: cord-0982877-3fpualuh
authors: Esposito, Flavio; Sansone, Lucia; Srivastava, Anubhav; Baldini, Francesco; Campopiano, Stefania; Chiavaioli, Francesco; Giordano, Michele; Giannetti, Ambra; Iadicicco, Agostino
title: Long period grating in double cladding fiber coated with graphene oxide as high-performance optical platform for biosensing
date: 2020-10-22
journal: Biosens Bioelectron
DOI: 10.1016/j.bios.2020.112747
sha: f06c0fd7535777561adb499a9d27eae9d7cad677
doc_id: 982877
cord_uid: 3fpualuh

In this work, the development and testing of a novel fiber-optic based label-free biosensor is presented, whose performance were verified through the detection of C-reactive protein (CRP) in serum. The device is based on a long period grating fabricated in a double cladding fiber with a W-shaped refractive index (RI) profile. As a result, the working point of the device was tuned to the mode transition region by chemical etching of the outer fiber cladding, obtaining a significant enhancement of the RI sensitivity and an excellent visibility of the grating resonances due to the mode transition in an all-silica structure. The fiber transducer was coated with a nanometric thin layer of graphene oxide in order to provide functional groups for the covalent immobilization of the biological recognition element. A very low limit of detection of about 0.15 ng/mL was obtained during the detection of CRP in serum, and a large working range (1 ng/mL – 100 μg/mL) of clinical relevance has been also achieved.

In the great many of biosensors based on fiber-optics, the engineering of the devices able to maximize the refractive index (RI) sensitivity, and hence to enhance the limit of detection (LOD), is often achieved impairing signal stability, repeatability and signal-to-noise ratio (SNR). One of the most appropriate example is given by long period grating (LPG) based devices which played a leading role in the scenario of fiber-optic biosensors (Chiavaioli et al., 2017a; Esposito et al., 2020a ). An LPG consists of a few centimeter long optical fiber with periodic perturbation in its J o u r n a l P r e -p r o o f refractive index and/or geometry, where period typically ranges from 100-1000 μm (James and Tatam, 2003) . The grating induces the coupling between the light from the fundamental core mode (of a single mode fiber) to forward propagating cladding modes. As a result, discrete attenuation bands are visible in the fiber transmission spectrum, located at those wavelengths , satisfying the phase-matching condition given by (James and Tatam, 2003; Shu et al., 2002) :

where , and , ( ) are the effective refractive index of the core mode and the i th cladding mode, respectively, while Λ is the grating period. The LPG finds large interest in bio-chemical sensing due to the intrinsic sensitivity to the surrounding refractive index (SRI) of the cladding modes. The sensing principle behind this kind of technology platforms lies in the measurement of changes in the surface refractive index stemming from the binding interaction between a biological recognition element (BRE) grafted onto the fiber surface and the analyte under investigation (Chiavaioli et al., 2017b) . However, suitable strategies are necessary in order to increase the sensitivity and selectivity (Chiavaioli et al., 2014a (Chiavaioli et al., , 2015 Del Villar, 2015; Shu et al., 2002; Śmietana et al., 2016) . The mode transition phenomenon promotes high SRI sensitivity, in turn achieving very low limits of detection. However, it demands the deposition of a high RI (HRI) overlay onto LPG region, which can affect the visibility of the attenuation bands, long term stability and repeatability of the device. In particular, the resonance depth can be reduced even to few dB during mode transition (Biswas et al., 2016; Del Villar et al., 2005) .

To address the following challenges, a device based on an unconventional long period grating fabricated in a double cladding fiber (DCF) with a W-shaped refractive index profile has been proposed here. As exemplary analyte to verify its performance, attention was focused on C-reactive protein (CRP), a relevant biomarker involved in several pathologies, which has been widely investigated. It is a homopentameric protein composed by five-polypeptide subunits (23 kDa each) non-covalently associated in a ring configuration. The level of CRP in serum in healthy Caucasian people is estimated to be around 0.8 µg/mL (Sproston and Ashworth, 2018) ; nevertheless, many factors can alter this value, such as age, gender, smoking status, weight, lipid levels, and blood pressure (Hage and Szalai, 2007) . CRP is then considered a serum biomarker for infection or inflammation during which its concentration can rise up to 1000-fold as a consequence of some bacterial infections (Aray et al., 2016; Sproston and Ashworth, 2018) . Moreover, constantly high concentration of CRP, rather than spikes of it, has been associated to cardiovascular diseases (CVDs) and to atherosclerosis (Nathan and Ding, 2010) . The clinical significance of CRP has been also evaluated for type 2 diabetes mellitus, age-related macular degeneration, hemorrhagic stroke, J o u r n a l P r e -p r o o f Alzheimer's disease, and Parkinson's disease (Luan and Yao, 2018) . Finally, very recently, it has been demonstrated that CRP levels could reflect COVID-19 disease severity and could be used as a key indicator for the monitoring of such disease (Wang, 2020) . In this frame of diseases, in which increasing concentration of CRP is involved, the importance of having a reliable, affordable, and sensitive device for CRP quantification became evident.

The commercially-available optical system for CRP measurements, at the moment, is the well consolidated enzyme-linked immunosorbent assay (ELISA) method, with which a LOD of about 10 pg/mL has been reached (Thermofisher Scientific, n.d.) . Other methods are based on different techniques such as immunoturbidimetric or immunonephelometry assay reaching 0.03 µg/mL and 0.2 µg/mL, respectively (Clarke et al., 2005; Dominici et al., 2004; Parra et al., 2005) . Different fiber-optic biosensors have been also reported until now for the detection of CRP (Salvo et al., 2017) . Among them, those based on surface plasmon resonance (SPR) are widely employed: for example, in 2016 a low limit of detection of 0.009 µg/mL was obtained in serum with a plastic optical fiber-based device (Aray et al., 2016) , whereas one year later a single-ended device was reported, which was tested for CRP concentrations down to 0.01 µg/mL in serum (Wang et al., 2017) . Fiber grating sensors have been also employed for this purpose. In 2015, an etched fiber Bragg grating (eFBG) coated with anti-CRP antibody -graphene oxide (GO) complex, was reported with a LOD of 0.01 µg/mL in serum (Sridevi et al., 2015) . More recently, with the same eFBG technology, a very low LOD of 0.82 fg/mL was obtained in modified buffer solution (Schulze et al., 2018) . Furthermore, the physical phenomenon of lossy mode resonance (LMR) has been applied for the detection of CRP, where an Indium Tin Oxide (ITO)-coated D-shaped fiber device attained a LOD of 0.0625 µg/mL in tri-buffer solution (Zubiate et al., 2017) . Finally, other technologies have also been exploited achieving remarkable LODs, such as, for example, chemiluminescence-based sensor (Islam and Kang, 2011) , where a LOD of 0.2 pg/mL in serum was attained.

The latest trends in nanomaterials for biosensing see the investigation of the properties of graphene oxide (Hernaez et al., 2017) . The hydrophilic nature of GO, deriving from the oxygencontaining functional groups, gives good water dispersibility and biocompatibility, which are highly important features in bio-applications. Due to its favorable intrinsic properties, it has been employed for the development of various types of biosensors based on electrochemistry, optics (fluorescence, colorimetry, and Raman), and mass analysis (Liu et al., 2012; Wang et al., 2011) .

Currently, only few papers have reported about the integration of GO with LPG technology (Esposito et al., 2018b; Liu et al., 2018; Xu et al., 2020) .

Very recently, an unconventional device working in mode transition region and capable of providing both high sensitivity and excellent SNR has been proposed (Esposito et al., 2020b; Giordano et al., 2020) . On these bases, in this work an LPG in a DCF with W-shaped RI profile empowered to induce the mode transition phenomenon in all-silica fiber, thus improving the grating spectral features, has been designed, developed and characterized. The grating was coated with a nanometric thin layer of GO exhibiting functional groups (such as carboxylic groups) for the covalent immobilization of the capture antibody. Finally, the developed system was tested towards the detection of CRP in serum using CRP concentrations of clinical relevance. 

Graphene oxide is a chemically modified graphene that is typically prepared by oxidation and exfoliation of graphite-bearing oxygen functional groups, such as carboxyl (-COOH), hydroxyl (-OH), or epoxy (-O-), on their basal planes and edges with the modified Hummers' method (Hummers and Offeman, 1958; Rourke et al., 2011) .

Here, a small amount of flake graphite (average particle diameter of 5 nm, 99.75% purity, Asbury Carbons, U.S.A.) was vigorously stirred for 5 days in a solution of NaNO 3 and KMnO 4 in concentrated H 2 SO 4 , washed with H 2 SO 4 (5% wt) and subsequently with a solution of H 2 O 2 (30% wt) to complete the oxidation. Inorganic anions and other impurities were removed through washing cycles that included centrifugation, discarding supernatant liquid and suspending the solid in an aqueous mixture of 3% wt H 2 SO 4 and 0.5% wt H 2 O 2 using stirring and ultrasonication. For GO layer fabrication, 10 mg of powder was first suspended in 5 ml of water by ultrasonication and then J o u r n a l P r e -p r o o f centrifuged at 14000 rpm to remove multilayer species. Finally, the GO solution was diluted at a concentration of 2 mg/mL. XE-100 Atomic Force Microscopy (AFM, Park Systems) was used for the imaging of GO sheets deposited on glass substrate. Surface imaging was obtained in noncontact mode using silicon/aluminum coated cantilevers (PPP-NCHR 10M; Park Systems) 125 μm long with resonance frequency of 200 to 400 kHz and nominal force constant of 42 N/m. The scan frequency was typically 1 Hz per line. Fig. 1(a) shows the AFM image of GO deposited on glass slide and the corresponding profile measurement. Nanosheets with different sizes ranging from few tens of nanometers to hundreds of nanometers can be observed. Profile measurement demonstrated the presence of sheets up to 4 nm thick.

The nanostructure morphology analysis of GO deposited on an optical fiber tip (as similar substrate but easier in handling during the imaging) was conducted by means of scanning electron microscopy (SEM FEI Quanta 200 FEG). From the SEM images showed in Fig. 1(b) , it is evident that GO has a multiple lamellar layer structure and it is possible to distinguish the edges of individual sheets. The films are stacked one above the other and show wrinkled areas. The individual GO sheets were found to have a width of 1-2 µm. Moreover, from the EDX analysis, the graphene oxide contains around 66.52 @% C, 32.72 @% O and 0.76 @% S.

The nanostructure of GO was also confirmed by Raman Spectroscopy, through a Renishaw inVia Reflex Raman spectrometer (UK) using a 514.5 nm excitation source. The Raman spectrum of GO film on a glass slide substrate is detailed in Fig. 1(c) , and it displays a D band at ∼1340 cm −1 and a broad G band at ∼1580 cm −1 . The D peak generates from the structural imperfections created by the attachment of hydroxyl and epoxide groups on the carbon basal plane. The G peak duly corresponds to the first order scattering of the E 2 G phonon mode of sp 2 C atoms. The intensity of 2D-band at ∼2700 cm −1 with respect to the D and G peaks is small, due to disorder and multilayer materials (Zhu et al., 2010) .

For the purpose of the work, the 2 mg/mL GO water dispersion, prepared according to the mentioned procedure, was deposited on the transducer surface by dip coating technique using a withdrawal speed of 100 mm/min (Esposito et al., 2018b) . 

The optical transducer proposed in this work consists of an unconventional LPG written in DCF (model S1310, manufactured by Nufern, USA) with W-type RI profile exploiting the mode transition phenomenon (Esposito et al., 2020b) , coated with nanometric layer of graphene oxide, as schematically reported in Fig. 2 .

The W-type fiber presents an outer cladding with refractive index higher than inner one: some cladding modes can be guided in the outer ring-shaped cladding without interacting with the core mode. However, through the chemical etching of the fiber (i.e. decreasing of the external cladding thickness) the transition phenomenon from outer-to inner-cladding of those modes can be induced.

As a matter of fact, the mode transition phenomenon is achieved in all-silica fiber, achieving a significant enhancement of the SRI sensitivity while preserving the band visibility, with consequent improvement in the resonance wavelength detection and SNR.

The device is subsequently coated with a thin layer of GO in order to provide functional groups on the sensor surface for the BRE grafting process.

J o u r n a l P r e -p r o o f 

The GO thin layer, which has been deposited onto the fiber surface, directly provides the -COOH functional groups, which are used for the covalent immobilization of the capture antibody, without the need for any additional functional layer (i.e. silane, polymer, etc.). In fact, with outstanding chemical stability and large surface area, GO consists of carbon atoms featured in two dimensional honeycomb lattice structure, containing both sp 2 -and sp 3 -hybridized carbon atoms as well as functional groups such as hydroxyl, epoxy, carboxyl on its basal plane and sheet edges, which can be used for immobilization of biomolecules. The enriched functional groups can interact in an ionic, J o u r n a l P r e -p r o o f electrostatic or, the strongest way, in a covalent manner, so that in principle they provide the highest extraction efficiency of biomolecules per unit area (Loh et al., 2010) . The -COOH groups of GO were activated by cross-linking chemistry with 200 mM EDC and 50 mM NHS for 30 minutes. The covalent immobilization of the receptor for CRP (monoclonal antibody clone C5, 500 µg/mL) was allowed to react for 1 hour at a very slow flow rate of 15 µL/min. The unreacted antibodies were removed washing for 10 minutes with PBS at a faster flow rate of 150 µL/min. The surface passivation was obtained with 1% (w/v) of BSA in PBS, treating the surface for 20 minutes and then washing for 10 minutes in PBS. The passivation step is useful in order to block the remaining activated carboxylic groups, so as to prevent non-specific adsorption onto the surface (Berneschi et al., 2017; Giannetti et al., 2020; Lichtenberg et al., 2019; Pisco and Cusano, 2020; Vaiano et al., 2016; Zubiate et al., 2019) .

The CRP calibration curves were obtained performing analyte binding steps, with increasing concentrations of CRP within the range 0.01-100 µg/mL in real samples of human serum (diluted 1:10 (v/v) in PBS) for 20/30 minutes at a slow flow rate of 25 µL/min, followed by a washing step with PBS for 10/15 minutes. In both cases, the duration of each step has been adjusted depending on the time necessary to reach the steady state in chemical kinetics during the binding interaction.

During the entire assay protocol, the temperature of the control system described in the previous section was set at (26±0.05) °C.

The 

For the purpose of the work, Nufern S1310 single-mode DCF was selected to host the LPG with (Iadicicco et al., 2007) . Moreover, slight changes are evident in the resonance bandwidths, except for some regions where an increasing behavior of these is observed. All figures reveal that 6 th cladding mode moves towards outside of the investigated wavelength range, when the fiber diameter reaches 120 µm (with slight disagreement in comparison to numerical result), whereas the 4 th cladding mode becomes visible for diameters lower than 117 µm. Moreover, it is worth pointing out that the shift rate is maximum when a modal transition phenomenon occurs. Two transition regions can be identified in Fig. 5(a) ; as far as the 5 th cladding mode is concerned, they occur for diameter values around 116 µm and 107 µm, represented by the shadowed regions. Here the resonance wavelength of the i-th order cladding mode shifts versus the one of the (i+1)-th and so on (Cusano et al., 2006; Esposito et al., 2020b) . Moreover, during the modal transition, the resonance amplitude ( Fig. 5(b) ) and bandwidth (Fig. 5(c) ) experienced a decrease and a widening, respectively, even if the attenuation band remained very well visible. This behavior was not observed in the numerical curves and can be attributed to irregularities in the grating structure.

Moreover, soon after transition (for instance when diameter is around 113 µm) the resonance amplitude and bandwidth of 5 th cladding mode (red line) changed to match those of the 6 th cladding mode (green curve). An unexpected effect (but repeatable for all investigated gratings) is visible far from the transitions: a decrease and a widening of the resonance amplitude and bandwidth occur, respectively, when fiber diameter is around 118 µm and 110 µm. Even if the reason is still under investigation, it can be ascribed to defects of the grating structure due to the fabrication process.

However, such regions are far from the operating point selected in this work, and hence they did not affect the final results.

The etching process was thus stopped at the mode transition around 107 µm to enhance the SRI sensitivity in water-like environment. The spectra after the etching process are plotted in Fig. 4 : in water environment was also evaluated according to (Esposito et al., 2019) and a maximum polarization dependent loss, PDL max , not higher than 5 dB was found. Subsequently, the grating was coated with a few nanometer thin layer of GO for the creation of functional groups on its surface, by using the procedure reported in section 2.2 (Esposito et al., 2018b ). Fig. 4 shows the spectra of the LPG in air and water after GO deposition with blue and orange solid lines, respectively. Regarding the latter, a resonance band centered at 1533 nm with depth of 10 dB and bandwidth of 43 nm can be observed: the peak depth in water environment decreased significantly. Moreover, we believe that the GO layer slightly affects the SRI sensitivity, even if it has refractive index of 1.8-2.0, because of the low thickness.

For comparison, a second LPG was fabricated in a similar manner, but in this case the etching process was stopped for a fiber diameter slightly higher (108 µm), which resulted in a working point less close to transition region. In the following, the devices are named LPG-B and LPG-A for the former and the latter reported, respectively. The final SRI sensitivities (before GO) in water environment were found to be equal to −1035 nm/RIU (13 times sensitivity gain due to etching) and −670 nm/RIU (10 times sensitivity gain) for LPG-B and LPG-A, respectively, by considering a linear approximation in the range 1.33-1.37. After the GO deposition, no further SRI characterization was carried out to preserve the GO surface properties. In fact, it is important to observe that it is advantageous to keep the GO layer as thin as possible because its non-negligible absorption coefficient affects the visibility of the attenuation band.

Finally, a third LPG (hereinafter named LPG-C) was fabricated with the same parameters and procedures used for LPG-B. This device played the role of assessing the repeatability and reproducibility of the entire process as previously mentioned.

In the following steps, such as biosensing layer preparation and CRP detection, the devices were placed inside the microfluidic system as described in section 2.4. (Chiavaioli et al., 2014a (Chiavaioli et al., , 2018 (Chiavaioli et al., , 2015 . It is worthy to observe that, except for the change of the signal due to a higher RI of BSA with respect to that of PBS, the surface passivation seems to be not so effective. In fact, the signal does not change before and after the PBS injection, which means that the available surface for the immobilization of the receptor is totally covered with a very low probability of non-specific binding interaction. This is also proved in Fig. 7 where a solution of diluted human serum is used as specificity test with a negligible change of the signal before and after the PBS injection; the residual shifts in all cases were of the order of the maximum experimental error (0.016 nm).

It should be pointed out that the sudden change in the signal before the immobilization of the capture antibody (anti-CRP) was due to the fact that the sensor was exposed to air for a few seconds in order to remove the air bubbles formed during the previous step. After that, everything took place regularly with the sensor always immersed in solution at a constant flow rate according to the values provided in Section 2.5.

J o u r n a l P r e -p r o o f 

The response of the three devices during the calibration phase with CRP solutions at increasing concentrations as specific target analyte is shown in Fig. 7 . In the sensorgrams, the first step refers to the specificity test with a solution of free-CRP human serum used as a negative control (blank measurement). Human serum is a highly complex biological matrix constituted by more than 10,000 different kinds of proteins, and is routinely used as clinically-validating procedure (Nanjappa et al., 2014) . In this case, the variation of the signal measured in PBS after rinsing was negligible and within the limits of variation of the signal (standard deviation) considering the SNR of the system, thus guaranteeing a good specificity of the biosensor (Chiavaioli et al., 2017b) . After that, the residual wavelength shifts measured in PBS after rinsing reflect the amount of CRP that binds when increasing concentrations of the target analyte are injected into the microfluidic system.

It could be noted from Fig. 7 that, differently from the other experiments, the LPG-B resonance wavelength increased immediately following by a slowly decrease when CRP solutions were injected. This might be relied on the effect of temperature stabilization inside the microfluidic system that implied a longer time for the injected solutions to reach the temperature set by the control system, since the CRP solutions were stored in the refrigerator and taken out just before their use. As mentioned in section 2.5, to overcome this issue, the duration of the steps for the analyte injection and the subsequent rinsing in PBS was lengthened with respect to LPG-A. In particular, the time to reach the stable condition during the injection of the CRP solution increased J o u r n a l P r e -p r o o f of about 10 minutes. Moreover, during the experiments with LPG-C, to avoid any problem, CRP solutions were taken out from the refrigerator some minutes before starting the detection procedure.

Another issue to observe is the fact that the LPG-A resonance wavelength experienced a slowly increase during the washing steps in PBS, while this didn't happen with the other sensors where a stable condition (i.e. steady-state) was attained. It can be ascribed to flow-induced bending on the grating due to high flow rate used during the washing steps. Furthermore, the gluing procedure of LPG-B and LPG-C into the microfluidic system was optimized by applying a slight pre-strain state along the fiber. Since the temperature was the same for all the LPGs, a comparison between the performances of the three sensors could be effectively performed. However, it is worth pointing out that the experiment with LPG-A (first experiment) can be considered as a proof-of-concept for the use of DCFs coated with GO as biosensors. In all cases, it is evident that the sensor response moved monotonically towards shorter wavelengths with increasing CRP concentration, thus in agreement with the sensing mechanism of the optical transducer used (Chiavaioli et al., 2017a; Esposito et al., 2018b; James and Tatam, 2003; Shu et al., 2002) .

The calibration curves for the three devices drawn in semi-log scale are shown in Fig. 8a Fig. 8(a) for LPG-A and in Fig. 8(b) for the other two LPGs). It should be pointed out that the experimental points showed in the calibration curves (Fig. 8) refer to a single independent experiment without any regeneration. The fitting of the experimental data by using the Logistic function is also reported as dashed curves. It was extrapolated from the Langmuir isotherm model, which is a well-accepted mathematical model capable to take into account the degree of interaction between ligand binding sites (Dudley et al., 1985) , as also detailed in the literature (Chiavaioli et al., 2014b) . Here, the Logistic function is profitably expressed in terms of wavelength shift ∆ as:

where ∆ accounts for the difference between the asymptotes of the sigmoidal curve, and hence represents the dynamic signal change (also named dynamic range), x 0 is the value of the analyte concentration for which the value of LPG resonance wavelength shift is equal to 50% of the dynamic range and p is a coefficient that is related to the slope of the sigmoidal curve for x = x 0 . As a matter of fact, it is expected that the first parameter of equation (2) changes proportionally to the sensitivity of the sensor, whilst x 0 and p should be comparable for all LPGs, since they depend on the binding interaction between anti-CRP and CRP (i.e. on the assay protocol). For a direct comparison and in order to assess the biosensor repeatability and reproducibility, Fig. 8 (b) accounts for the histogram of the results extrapolated from Fig. 8(a) for the detection of CRP with two independent experiments using LPG-B (blue) and LPG-C (grey). Each column represents the mean value of the signal, together with the respective standard deviation within subsequent 20 measurements (black error bars). Table 1 gathers all the fitting parameters for the three devices. The high correlation coefficient R 2 proves the effectiveness of the fitting procedure. As expected, x 0 and p are quite similar (within the error) for all LPGs, thus testifying the same interaction mechanism, whereas ∆ is significantly higher for LPG-B and LPG-C than that for LPG-A due to the improved sensitivity. Those results can be considered very remarkable in biosensing, taking into account all the involved variables, and underpin the effectiveness of the proposed biosensing platform.

J o u r n a l P r e -p r o o f According to the literature (Chiavaioli et al., 2017b) , a series of parameters of interest can be derived from the calibration curve to assess the performance of the biosensor. Among them, the LOD, defined as the variation of the signal relating to the concentration 0 of the analyte (blank measurement) plus three times its standard deviation, is certainly one of the most important. In this case, a LOD of about 1.1 ng/mL (3% ≅ 0.026 nm) is attained with LPG-A. Despite a slightly larger standard deviation (3% ≅ 0.042 nm for LPG-B and 3% ≅ 0.037 nm for LPG-C) as expected from the spectral features of the resonance, a LOD of about 0.32 ng/mL and 0.15 ng/mL is attained with LPG-B and LPG-C, respectively, a value roughly 60-fold lower (LPG-C) than that reported in the literature using the same assay protocol based on the anti-CRP/CRP interaction, but making use of an optical transducer based on the SPR phenomenon (Aray et al., 2016) . While the LOD attained is not the lowest reported in the literature (Schulze et al., 2018) , it should be pointed out that the respectively. In addition, the surface sensitivity can be calculated (Chiavaioli et al., 2017b) if the CRP concentration is reported in linear scale like the classical Langmuir isotherm model, by applying a linear regression approach from the blank measurement up to 0.01 µg/mL CRP where the linearity is maintained. This calculation leads to the values of -19.22 nm/(µg/mL), -18.63 nm/(µg/mL) and -8.58 nm/(µg/mL) for LPG-C, LPG-B and LPG-A, respectively. Finally, another interesting parameter was evaluated, which is the working range of the biosensor defined as the CRP concentration range between 10% and 90% of the dynamic range (∆ ) (Chiavaioli et al., 2014b) . The proposed biosensing device can cover a large working range (1 ng/mL -100 µg/mL) of clinical relevance.

Given the reported values of LOD and working range, it is important to point out that a common laboratory practice is to dilute the body-fluid sample (i.e. serum, plasma, etc.) in order to minimize the matrix-effect of not specific adsorption onto the surface of the sensing area. The applied dilution is generally 1:10, but can be even down to 1:100, leading, consequently, to a decrease of one (1:10) or two (1:100) orders of magnitude of the clinical range. Taking into account that the level of CRP in serum in healthy Caucasian people is estimated to be around 0.8 µg/mL, a LOD of 0.08 µg/mL or 0.008 µg/mL should be considered in case of dilution of 1:10 and 1:100, respectively. With this in mind, the low LOD and the wide working range attained with the proposed devices are perfectly justified.

In this work, we have proposed a novel fiber-optic based biosensor which was tested for the realtime label free detection of C-reactive protein in serum. It is based on an unconventional LPG fabricated in W-type double cladding fiber by EAD technique. Here, the working point of the device was tuned to mode transition region by slight etching of the fiber outer cladding, in order to attain significant enhancement of the sensitivity and at the same time preserving the visibility of the grating spectral features (and thus SNR). In addition, due to absence of additional buffer layers, the so-prepared LPGs, consisting of all-silica structure tuned in modal transition, guarantee stability and long-term use with similar spectral and sensitivity features. Moreover, fiber transducer was coated with a nanometric thin layer of GO exhibiting oxygen-containing functional groups, which were used for the covalent immobilization of the capture antibody. The device was tested using a stateof-the-art microfluidic system for fiber-optic biosensors and the achieved results were analyzed by considering a series of parameters of interest. The device was applied for the detection of CRP concentrations in serum of clinical relevance. In the best case, a LOD of about 0.15 ng/mL was J o u r n a l P r e -p r o o f

SPR-based plastic optical fibre biosensor for the detection of C-reactive protein in serum

Fluorescence biosensing in selectively photo-activated microbubble resonators

Realization of high sensitive refractive index sensor with enhanced contrast using an over coupled long period fiber grating operating at mode transition

Towards sensitive label-free immunosensing by means of turn-around point long period fiber gratings

Biosensing with optical fiber gratings

Sol-Gel-Based Titania-Silica Thin Film Overlay for Long Period Fiber Grating-Based Biosensors

Towards a Uniform Metrological Assessment of Grating-Based Optical Fiber Sensors: From Refractometers to Biosensors

Femtomolar Detection by Nanocoated Fiber Label-Free Biosensors

Characterisation of a label-free biosensor based on long period grating

Comparison of differing C-reactive protein assay methods and their impact on cardiovascular risk assessment

Mode transition in high refractive index coated long period gratings

Ultrahigh-sensitivity sensors based on thin-film coated long period gratings with reduced diameter, in transition mode and near the dispersion turning point

Nanodeposition of materials with complex refractive index in long-period fiber gratings

Measurement of C-reactive protein: Two high sensitivity methods compared

Guidelines for immunoassay data processing

Arc-Induced Long Period Gratings in Erbium

Arc-Induced Long Period Gratings from Standard to Polarization-Maintaining and Photonic Crystal Fibers

Ultrasensitive biosensor based on long period grating coated with polycarbonate-graphene oxide multilayer

Label-free Biosensors based on Long Period Fiber Gratings: A Review

Sensitivity Enhancement in Long Period Gratings by Mode Transition in Uncoated Double Cladding Fibers

Towards an Integrated System as Point-of-Care Device for the Optical Detection of Sepsis Biomarkers

Trasduttore in fibra ottica chimicamente funzionalizzato per la sensoristica. IT Patent pending nr

C-Reactive Protein Gene Polymorphisms, C-Reactive Protein Blood Levels, and Cardiovascular Disease Risk

Optical Fibre Sensors Using Graphene-Based Materials: A Review

Preparation of Graphitic Oxide

Spectral behavior in thinned long period gratings: effects of fiber diameter on refractive index sensitivity

Chemiluminescence detection of label-free C-reactive protein based on catalytic activity of gold nanoparticles

Optical fibre long-period grating sensors: characteristics and application

Non-Specific Adsorption Reduction Methods in Biosensing

Graphene oxide functionalized long period fiber grating for highly sensitive hemoglobin detection

Biological and chemical sensors based on graphene materials

Graphene oxide as a chemically tunable platform for optical applications

The Clinical Significance and Potential Role of C-Reactive Protein in Chronic Inflammatory and Neurodegenerative Diseases

Plasma Proteome Database as a resource for proteomics research: 2014 update

Nonresolving Inflammation

Use of a time-resolved immunofluorometric assay for determination of canine C-reactive protein concentrations in whole blood

Lab-On-Fiber Technology: A Roadmap toward Multifunctional Plug and Play Platforms

Long period fiber grating nano-optrode for cancer biomarker detection

The Real Graphene Oxide Revealed: Stripping the Oxidative Debris from the Graphene-like Sheets

Sensors and Biosensors for C-Reactive Protein, Temperature and pH, and Their Applications for Monitoring Wound Healing: A Review

Femtosecond-Pulsed Laser Written and Etched Fiber Bragg Gratings for Fiber-Optical Biosensing

Sensitivity characteristics of long-period fiber gratings

Towards refractive index sensitivity of long-period gratings at level of tens of µm per refractive index unit: fiber cladding etching and nano-coating deposition

Role of C-Reactive Protein at Sites of Inflammation and Infection

Sensitive detection of C-reactive protein using optical fiber Bragg gratings

Lab on Fiber Technology for biological sensing applications

C-reactive protein levels in the early stage of COVID-19

A label-free fiber optic SPR biosensor for specific detection of C-reactive protein

Graphene and graphene oxide: Biofunctionalization and applications in biotechnology

Graphene oxide-functionalized long period fiber grating for ultrafast label-free glucose biosensor

Graphene and Graphene Oxide: Synthesis, Properties, and Applications

Fiber-based early diagnosis of venous thromboembolic disease by label-free D-dimer detection

High sensitive and selective Creactive protein detection by means of lossy mode resonance based optical fiber devices

• Fiber biosensor based on long period grating (LPG) in double cladding fiber (DCF)

GO) coating with functional groups for covalent binding of antibody • High sensitivity and signal visibility due to mode transition in all-silica structure • Real-time and label-free detection of C-reactive protein (CRP) in human serum • Low limit of detection (LOD) of 0.15 ng/mL and large working range