1. Introduction

Serum albumin plays an important role in the normal functioning of the human body. It also balances the osmotic pressure of plasma and other colloids [1,2]. The amount of serum albumin in blood is an indicator of the nutritional status of an organism. The circulation of abnormal amounts of serum albumin in the blood may be harmful to health. An excess amount of serum albumin may cause heart or kidney problems, and a very low level of serum albumin in the blood may cause edema [3]. Therefore, the development of a precise, fast, and sensitive detection system for serum albumin in medical and enzymatic reaction applications is highly demanded. Bovine serum albumin (BSA) is one of the most widely investigated proteins from the serum albumin series [4]. It is a cow protein having structural homology with human serum albumin (HSA) and is readily accessible [5]. So BSA is frequently studied as a model protein instead of HSA. This protein has a molecular weight of 66.5 kDa and is used in biomedical applications and enzymatic reactions [6,7]. Moreover, BSA is used in many immunochemical assays as a target for analyzing and designing purposes. Therefore, the detection of BSA has attracted increasing attention in immunological and bio-analytical studies [4]. Determination of micro-quantities of BSA is possible with methods like enzyme-linked immunosorbent assay (ELISA) [8], surface-enhanced Raman spectroscopy (SERS) [9], electrochemical [10,11] FT-IR spectroscopy [12], and fluorimetric measurements [13]. Although these techniques provide significant sensitivity and selectivity, they have several limitations, such as complex procedures, time-consuming, and poor reproducible performance. Some of these methods require a labeled reagent like enzyme-labeled antibody/antigen [14]. The literature search revealed that most of the fluorescence-based BSA sensor suffers from drastic reductions in their fluorescence emission peaks due to the quenching effect [15]. The traditional impedance-based electrochemical sensors require a longer time for the measurement under a steady-state frequency response. The recording of a full impedance spectrum makes the measurements time-intensive. In addition, the long-term test may lead to an increase in non-specific adsorption or distortion in the low-frequency region [16]. Therefore, a hypersensitive, stable, compact, label-free, and accurate measuring system can address the above-mentioned limitations.

The integrated technique of modal interferometry and MIPs would be a promising technique for its compact miniaturized size, ultra-sensitivity, and hassle-free fast performance [17–19]. MIPs are a kind of synthetic polymeric receptors that possess excellent selectivity and high binding performance comparable to natural biomolecules, like enzymes-substrates, antigens-antibodies, and receptor-hormones [20–22]. The recognition sites formed by embedding the target molecules into the polymeric structure are complementary to the target molecules in shape, size, and functional groups. Molecularly imprinted technology (MIT) is gradually becoming a research hotspot because of its excellent stability, easy synthesis, fast response, and compatibility [23,24]. These imprinted polymeric receptors are immobilized on interferometric transducers for the development of biosensors. Different processed fiber structures are mainly used in the fabrication of transducing platforms. The inline splicing of photonic crystal fiber (PCF) with normal single-mode fiber (SMF) forming an SMF-PCF-SMF hybrid structure has the potential of a compact, stable, and inline interferometric sensing transducer [25]. In this transducer, the superposition of core and excited cladding modes produces a stable interference spectrum. The sensitivity of this transducer depends on the significant influence of cladding modes for the binding of target analytes. The immobilization of different 2D materials on the transducer surface having high absorption potential can increase the intensity of the excited cladding modes significantly and hence the sensitivity of these sensors [26,27].

Here an interferometric biosensor has been fabricated for the detection of serum protein BSA with high sensitivity. The transducer of the proposed probe is made using an SMF-PCF-SMF hetero-fiber structure and functionalized with a nanocomposite of MIP-MoS₂ as a recognizer. 2D material like MoS₂ has a high specific surface area and strong absorption making a potential way to enhance the intensity of interfering cladding modes of this transducing platform. The sensor shows an unprecedented sensitivity of 2.34 × 10⁷ pm/µg L^-1. The optimization of performance parameters like template concentration, dipping, and elution periods is carried out to achieve efficient performance. The sensor has working potential over a wide pH range. The proposed sensor shows a wide dynamic detection range of 10⁻⁵-10³µg/L with an extremely low (sub-femtomolar) detection limit. The compact structure, easy fabrication, high sensitivity, large operating range, and stable performance of this sensor show the potential for long-distance sensing and online monitoring of biological entities.

2. Fabrication of sensing probe

The sensing probe is made using a fiber optic transducer functionalized with the recognizing unit. The modal interferometry technique is utilized to fabricate the transducing platform. This transducer is functionalized by a nanocomposite of MIP-MoS₂ for selective detection of BSA protein.

2.1. Fabrication of transducer

The transducer of this photonic biosensor is made using a hetero-fiber structure. A solid core PCF (NKT, LMA-8) is used in the fabrication process, having a clad diameter of 125 µm with a crystal defect at the center forming the solid core of diameter 8.5 µm. The PCF is designed in such a way that the core is surrounded by 6-fold symmetric arrangements of air holes having a diameter of 2.32 µm and a pitch of 5.6 µm. These air channels combine with solid silica effectively forming the cladding region of this fiber. The fabrication process is illustrated schematically in Fig. 1(a). In the first step, a section of PCF is cleaved using a commercial fiber cleaver (Sumitomo, FC-6RS) and spliced with an SMF (9/125 µm, SMF-28) pigtail using a fusion splicer (Fujikura 80s) [28]. The arc time and power of the splicing process are kept at ∼ 3000 ms and STD + 20 bit respectively. During the splicing process, the applied arc fused the air channels of the PCF around the SMF-PCF splicing point. This fused portion is called the collapse region as denoted by CR-1. In step 2, another end of the PCF section is cleaved at a distance of ∼ 2 cm from the first splicing point. This cleaved end of PCF is again spliced with another SMF pigtail in a similar way as shown in step 3 of Fig. 1(a). The splicing of this PCF-SMF junction produced another collapse region denoted as CR-2. This completes the transducer fabrication process, producing an SMF-PCF-SMF or simply an SPS structure. The microscopic image of the PCF cross-section is shown in Fig. 1(b) indicating the solid core surrounded by air holes. The enlarged view of these air holes is displayed in Fig. 1(c) showing the average diameter (d) and pitch (Λ) of air holes. The microscopic images of the two collapse regions are shown in Fig. 1(d) and (e) respectively. The length of both collapse regions is around ∼180 µm. The simulation work is carried out using the finite element method (FEM) in COMSOL Multiphysics 5.6 to find the intensity distribution for this sensing transducer. The intensity distribution of the fundamental core mode (LP₀₁) in PCF is shown in Fig. 1(f) and the propagation of light in the proposed fiber optic transducer is shown in Fig. 1(g). This figure shows that the light in the leading SMF propagates through the core as fundamental core mode. At the first collapse region (CR-1) the light diffracts and excites the higher-order cladding modes in the PCF. These excited modes propagate through the cladding region and at the second collapse region (CR-2) again coupled with the core mode eventually. The superposition of these modes produces an interference pattern in the output.

 

Fig. 1. (a) Schematic image of the transducer fabrication process. (b) Microscopic image of PCF cross-section. (c) Enlarged view of PCF cross-section. (d) Microscopic image of first collapse region (CR-1). (e) Microscopic image of second collapse region (CR-2). (f) Simulative intensity distribution profile of PCF cross-section showing fundamental core mode. (g) Intensity distribution in SMF-PCF-SMF based transducer.

Download Full Size | PDF

2.2 Synthesis of MIP@MoS₂ nanocomposite

The MIP@MoS₂ nanocomposite is used as a recognizing material in this sensor which is synthesized following a two-step process. In the first step, the MoS₂ nanosheet is prepared using a hydrothermal process followed by the synthesis of imprinted polymer, and subsequently, MIP@MoS2 nanocomposite is prepared. The MoS₂ is synthesized by dissolving 1 g of sodium molybdate and 1.2 g of thioacetamide in 80 ml of distilled water (DI water) under a rigorous stirring condition until a homogeneous and transparent solution is obtained. This solution is made acidic in nature by adding 0.4 g of oxalic acid. Finally, the solution is kept in a 100 ml stainless steel air-tight autoclave and heated in a hot oven at a constant temperature of 180 °C for 24 h without stirring. This completed the MoS₂ synthesis process producing a black colored homogeneous dispersion of MoS₂ nanosheet. In the second step, the BSA imprinted polymer is synthesized. In this case, 80 mg BSA (Template) 315 µl of methacrylic acid (MAA) and 1.52 ml ethylene glycol dimethacrylate (EGDMA), and 32 mg of azobisisobutyronitrile (AIBN) is dissolved in 80 ml of dimethyl sulfoxide (DSMO) solvent followed by the mixing using a magnetic stirrer. The dissolved O₂ is removed from this solution by purging N₂ into it for 20 min. Finally, the container is sealed properly under an N₂ gas environment and heated in a water bath at 70 °C for 24 h. This completed the polymer synthesis process resulting in the formation of a dispersed solution containing the polymer. This polymer freezes the BSA molecules and is termed a non-imprinted polymer (NIP). This NIP is mixed with a previously synthesized MoS₂ solution at a 4:1 v/v ratio. The final mixture is sonicated for 4 h completing the synthesis of NIP-MoS₂ nanocomposite. The BSA molecules are removed from this polymer structure in the subsequent probe fabrication process and it is then called MIP-MoS₂ nanocomposite.

2.3 Functionalization of transducer

The fabricated transducer is functionalized by MIP-MoS₂ nanocomposite to make it sensitive for BSA detection. This functionalization process is illustrated schematically in Fig. 2(a). In this process, initially, the transducer is fixed in a lab-made sample holder. The probe region is thoroughly cleaned using acetone and methanol successively to remove any contaminants from the fiber surface. The fiber is dried in a hot oven at 60 °C for 1 h. The interference spectrum of this bare transducer is shown in Fig. 2(b) by the solid violet line. The polymer solution containing the recognizer element is given to the sample holder using a microfluidic syringe. The NIP-MoS₂ nanocomposite binds to the fiber surface through a physical adsorption process. This immobilized probe is placed into the hot oven for 1 h for drying. The interference spectrum of this coated probe is shown by the brown dotted line. These two curves imply that the transmitted power decreases slightly for the absorption of light by the immobilized NIP-MoS₂ layer. To create the imprinted sites on the immobilized NIP nanolayer, the probe is eluted. This elution process removes the bounded BSA molecules from the polymer structure and creates complementary binding sites. The resulting polymer is termed as MIP. The removal of BSA molecules from the polymer increases the transmitted power slightly as shown by the pink dashed curve in Fig. 2(b). This completed the probe functionalization process.

 

Fig. 2. (a) The representation of the probe functionalization process. (b) Output characteristic curves at each step of the probe functionalization process. (c) The sensing mechanism of this proposed sensor.

Download Full Size | PDF

3. Theory and sensing principle

This sensor uses the SPS-based hetero-fiber structure as the transducer. In this fiber configuration, the superposition of core mode and excited cladding modes produces an interference at the output. The intensity of this spectrum can be expressed using Mach-Zehnder double beams interference as governed by the following relation [29]:

Here, ${I_{core}}$ and ${I_{clad}}$ are the intensities of core and higher-order cladding modes propagating through the PCF. The phase difference between these two interfering modes at the point of interference is given by

Here, $\delta {n_{eff}}$ is the change in the effective index of cladding mode for changing external RI. The BSA molecules from the applied sample solutions bind selectively with the imprinted sites of functionalized MIP. This alters the effective index of the cladding modes and hence produces wavelength shifts. Thus, by precisely measuring the wavelength shift and calibrating, the concentration of external analyte (BSA) can be measured. The sensing mechanism of this sensor is explained schematically in Fig. 2(c).

4. Optimization of sensor’s performance

The performance of the sensor depends on multiple parameters namely the dipping time, elution time, template concentration, amount of cross-linker, polymerization temperature, amount, and nature of solvent. Among them, the template concentration, dipping time, and elution time play significant roles in the efficient performance of the sensor. So, these factors are optimized to make the sensor efficient.

4.1 Optimization of template concentration

In this sensor, the MIP acts as a receptor for selective detection of BSA. This MIP is synthesized experimentally, where BSA is used as the template. The amount of template plays a significant role in the efficient performance of MIP. This determines the number of specific binding sites in the polymer matrix. The template (BSA) amount is optimized by synthesizing MIP of varying template concentrations ranging from 0 g/L to 2 g/L. The synthesized polymers are used subsequently to make sensing probes. The dipping and leaching times of all these probes are taken constant. The performance of these probes is checked using a BSA sample of concentration 10³mg/L. The wavelength shifts produced by these probes are plotted in Fig. 3(a) showing the wavelength shift for increasing template amount giving maximum shift. For higher templates, the ratio of template and functional monomer is high producing a large number of imprinted sites on the polymer matrix. These sites are formed very closely to each other and hence the binding of a BSA molecule is affected by the surrounding sites. Therefore, these sites cannot work effectively and hence produce a smaller wavelength shift. On the other hand, for a small amount template, the number of binding sites is small giving smaller shifts. The probe immobilized with NIP (without any template) produces a negligible wavelength shift due to the absence of any binding sites in the immobilized polymer

 

Fig. 3. Optimization of (a) template (BSA) concentration during polymer synthesis, (b) dipping time during probe immobilization process, and (c) the leaching time to create imprinted binding sites on the polymer surface.

Download Full Size | PDF

4.2 Optimization of dipping time

The recognizing material MIP-MoS₂ nanocomposite is immobilized on the transducing surface through the dip coating method. The dipping time of this process plays an important role as it determines the thickness of the immobilized MIP layer on the transducer. The thickness of the recognizing material controls the change of the effective index of interfering cladding modes for binding. Hence, the thickness of this recognizing material is optimized by controlling the dipping period. The sensing probes are prepared by immobilizing with a nanocomposite of varying dipping periods ranging from 15 min to 45 min. From the transmitted spectrum of these probes, the wavelength shift of a specific dip has been plotted in Fig. 3(b). The figure shows that the wavelength shift is maximum for 25 min. For an immobilization with a lower dipping time, the thickness of the sensing region is small which cannot bind a significant number of BSA molecules. On the other hand, for a longer time of dipping the thickness is higher which allows the binding of a larger number of analyte molecules. However, such binding cannot alter the mode propagation for longer distances from the fiber surface. Hence, in both cases, the sensor produced a comparatively smaller wavelength shift. This characterization shows the optimized dipping time is 25 mins.

4.3 Optimization of leaching time

The template (BSA) is removed from the NIP-MoS₂ nanocomposite through a leaching process. This creates complementary binding sites in the polymer surface. The leaching time controls the number of binding sites in a polymer matrix. Thus, this leaching time is optimized by making sensing probes immobilized with optimized template concentration and dipping time. The probes are leached for varying durations ranging from 0 min to 7 min. These probes are then subsequently characterized using a BSA sample of concentration 10³µg/L. The wavelength shift produced by the probes of varying leaching times is plotted in Fig. 3(c) showing the maximum wavelength shift is for a leaching time of 3 min. The lower leaching times are not sufficient to remove a significant number of BSAs from the polymer matrix. The probe with 0 min leaching produced a negligible wavelength shift for the absence of any binding sites in the polymer matrix. On the other hand for a longer time, leaching affected the polymer structure. Thus, in both cases, the sensor produced lower wavelength shifts. The obtained optimized template concentration, dipping time, and leaching time as obtained from the optimization process is 1 g/L, 25 min, and 3 min respectively. These optimized parameters are used to fabricate the sensing probe and used for the subsequent characterization.

5. Characterization of sensor

5.1 Characterization with BSA sample

The fabricated probe is characterized using wavelength interrogation technique with the help of BSA sample solutions of concentration range 10⁻⁵mg/L to 10³mg/L. The samples are injected into the sample holder using the microfluidic syringe. The corresponding transmission curves of the sensor are recorded in the detector. The variation of specific dip wavelength with the BSA sample concentration is plotted in Fig. 4(a). This shows the dip wavelength shifts towards the higher wavelength region for increasing the concentration. This wavelength shift is calibrated with sample concentration and plotted in Fig. 4(b). The calibration shows the response of this sensor with the applied sample concentration. In the lower concentration regime of the sample, the sensor gives a linear shift with concentration. However, at higher concentrations, this linearity deviates slightly as clear from this graph. This saturation behavior arises due to the presence of a finite number of imprinted sites in the polymer matrix. Hence, with increasing the concentration almost all sites occupied with BSA molecules. Therefore, the number of binding sites per BSA molecule decreases with concentration. The wavelength shift increment rate with sample concentration tells about the sensitivity. The sensitivity of this sensor is calculated from the calibrated relation for all concentration values of the sample. The logarithmic sensitivities of this sensor have been plotted in Fig. 4(c). The maximum sensitivity of this sensor is found 2.34 × 10⁷ pm/µg L^-1 for a concentration of 10⁻⁵mg/L. The detection limit of this sensor is calculated using three times of standard deviation of wavelength shift and sensitivity of the sensor at near zero concentration of the sample. The LOD of this sensor is found to be 0.65 fM. Most of the biosensors belonging to different categories possess a sensitive detection performance for their highly sensitive transducing platforms. However, a large number of such sensors suffer from poor recognition ability. These sensors cannot differentiate the target analyte of interest and show responses for other structurally analogous elements. Such non-specific binding makes these sensors poor in selective and creates doubt on its reliability. This serious issue prevents their applications despite of having high sensitivity. This major drawback can be solved by functionalizing the transducers with highly selective, stable recognizer elements. Here, in this sensor, the MIP has been utilized as a recognizing agent. It is an artificial molecular receptor that shows highly selective detection potential with good chemical stability. So we strongly believe that this sensor will show high selectivity and recognizing power over the traditional sensors. This selectivity performance can be checked in the future by observing the performance of this sensor for sample solutions of different analogous chemical compounds.

 

Fig. 4. (a) Characteristic curve of the sensor for the variation of BSA concentration. (b) Calibration of the wavelength shifts with the variation of the BSA concentration. (c) Sensitivity curve of the sensor. (d) The response of the sensor for variation of sample pH.

Download Full Size | PDF

5.2 Characterization of the sensor with sample pH

In this sensor, the nanocomposite MIP-MoS₂ has been used as a molecular recognizer for BSA detection. The BSA molecules fit inside the imprinted binding sites of the MIP-MoS₂ nanocomposite through the formation of hydrogen bonds. So, the pH of this sample has an active role in this bond formation and hence the performance of this sensor depends on the sample’s pH. This effect is checked by using aqueous BSA samples with equal concentrations but different pHs ranging from 5 to 9. The wavelength shift produced by the sensor for these samples of different pH levels is plotted in Fig. 4(d). This figure indicates that the sensor produced the maximum wavelength shift for a sample having pH 7. The higher and lower pH of the analyte sample affects the formation of hydrogen bonds. Hence, the sensor produced a comparatively smaller wavelength shift.

5.3 Reproducible performance of the sensor

The reproducible performance of a biosensor is an important characteristic feature that ensures its commercial production and applications. The reproducibility of this sensor is checked by fabricating five identical sensing probes on the same day using a similar experimental technique. The performance of these fabricated probes is checked using the BSA sample of a particular concentration (10³µg/L). The deionized water of BSA concentration 0 µg/L is taken as a reference sample for all these probes. The wavelength shift produced by these probes for the given concentration change is measured and plotted in Fig. 5. The figure shows that all these probes produce almost equal amounts of wavelength shifts for this BSA concentration. This confirms the reproducibility of the fabrication process. Moreover, the repeatability of each probe is also checked by measuring the wavelength shift produced by each probe for the application of the BSA sample (10³µg/L) multiple times. The shift produced by each probe for this repeating measurement is expressed by the error bars as shown in Fig. 5. The small variation in shifts confirms the repeatable performance of these sensors.

 

Fig. 5. The reproducible performance of the sensor.

Download Full Size | PDF

6. Discussions and conclusion

In this work, an optical biosensor is fabricated using PCF based modal interferometric transducer functionalized with a nanocomposite of MoS2 and MIP for precise detection of BSA. The performance of the proposed photonic biosensor is compared with other reported BSA sensors that have been developed utilizing different physical phenomena to establish its advanced characteristic features. This critical comparison expressed in Table 1 is made according to the category, working principle, probe configuration, operating range and detection limit of the sensors. The table shows that the proposed sensor holds the highest operating range (10⁻⁵-10³µg/L) as compared to other reported BSA sensors to the best of our knowledge except the sensor in Ref. [30]. Further, the current BSA sensor possesses a very low detection limit (up to sub-femtomolar) except few electrochemical sensors presented in Refs. [30], [39], and [43]. However, these sensors suffer from some limiting issues, like bulky devices, time-consuming measurements, complex probe fabrication processes etc. Moreover, these sensors are not suitable for online long-distance monitoring and remote biosensing applications.

In summary, a novel fiber optic modal interferometry-based biosensor has been fabricated using a nanocomposite of MIP and 2D material MoS₂ for trace-level detection of serum protein BSA. The transducer is fabricated using an SMF-PCF-SMF heterostructure that produces a stable interference spectrum for modal interferometry of core and cladding modes. The transducer is functionalized with MIP-MoS₂ nanocomposite facilitating complementary imprinted binding sites for precise detection of BSA. The 2D material MoS₂ with high absorption power provides an enhanced sensitivity of the sensor by increasing the cladding mode intensity. The performance of the sensor has been optimized by optimizing the control parameters: template concentration, dipping, and elution period. The characterization results show that the sensor has a wide detection range of 10⁻⁵µg/L to 10³µg/L with a trace-level detection potential of BSA up to sub-femtomolar concentration. It shows a working potential for a wide pH variation of sample solution. Easy fabrication, stable structure, wide detection range, and high sensitivity of this sensor show a new path for the development of 2D material and MIP-based biosensors in different clinical application fields.

Table 1. Comparison of characteristic features with the reported sensors^a

View Table