Label-Free Optical Biochemical Sensors via Liquid-Cladding-Induced

Research Article www.acsami.org

Label-Free Optical Biochemical Sensors via Liquid-Cladding-Induced Modulation of Waveguide Modes Nhu Hoa Thi Tran,†,‡ Jisoo Kim,†,‡ Thang Bach Phan,§,∥,⊥ Sungwon Khym,‡,# and Heongkyu Ju*,†,‡,∇ †

Department of Nano-Physics and ‡Gachon Bionano Research Institute, Gachon University, Seongnam, Gyeonggi-do 461-701, South Korea § Center for Innovative Materials and Architectures, ∥Faculty of Materials Science, University of Science, and ⊥Laboratory of Advanced Materials, University of Science, Vietnam National University, HoChiMinh, Vietnam # Department of Science, Hongik University, 94 Wausan-ro, Mapo-gu, Seoul, Republic of Korea ∇ Neuroscience Institute, Gil Hospital, Incheon 405-760, South Korea ABSTRACT: We demonstrated modulation of the waveguide mode mismatch via liquid cladding of the controllable refractive index for label-free quantitative detection of concentration of chemical or biological substances. A multimode optical fiber with its core exposed was used as the sensor head with the suitable chemical modification of its surface. Injected analyte liquid itself formed the liquid cladding for the waveguide. We found that modulation of the concentration of analyte injected enables a degree of the waveguide mode mismatch to be controlled, resulting in sensitive change in optical power transmission, which was utilized for its real-time quantitative assay. We applied the device to quantitating concentration of glycerol and bovine serum albumin (BSA) solutions. We obtained experimentally the limit of detection (LOD) of glycerol concentration, 0.001% (volume ratio), corresponding to the resolvable index resolution of ∼1.02 × 10−6 RIU (refractive index unit). The presented sensors also exhibited reasonably good reproducibility. In BSA detection, the sensor device response was sensitive to change in the refractive indices not only of liquid bulk but also of layers just above the sensing surface with higher sensitivity, providing the LOD experimentally as ∼3.7 ng/mL (mass coverage of ∼30 pg/mm2). A theoretical model was also presented to invoke both mode mismatch modulation and evanescent field absorption for understanding of the transmission change, offering a theoretical background for designing the sensor head structure for a given analyte. Interestingly, the device sensing length played little role in the important sensor characteristics such as sensitivity, unlike most of the waveguide-based sensors. This unraveled the possibility of realizing a highly simple structured label-free sensor for point-of-care testing in a real-time manner via an optical waveguide with liquid cladding. This required neither metal nor dielectric coating but still produced sensitivity comparable to those of other types of label-free sensors such as plasmonic fiber ones. KEYWORDS: optical biochemical sensor, waveguide mode mismatch, label-free sensors, multimode optical fiber, refractometer, remote sensing

1. INTRODUCTION

optical waveguides. Use of optical waveguides as biosensor heads, whereby light guidance ensured efficient coupling of optical signal with analyte, favors miniaturization of a sensing platform. This could find potential applications for a multichanneled structure of a tiny sensor and remote sensing platform. Optical-waveguide-based biosensors made use of a change, upon analyte injection, in the guided mode properties such as optical phase,25−27 polarizations,28 optical spectra,29 and transmitted power for sensor signal transduction.30−32 It was also seen that use of smaller number of waveguides in optical biosensors could benefit the sensor stabilities due to the

Label-free biosensors are devices capable of quantitative assay of target analyte in a sample without using labels such as fluorophores and radioactive isotopes that need to be prelabeled on capture probes as in label-aided assays.1−3 The label-free devices that could allow real-time monitoring of timedependent binding kinetics as well as the time-efficient and cost-effective sensing operation has led to development of a variety of technologies for clinical diagnosis of diseases, ranging from those by cyclic voltammetry techniques,4−6 impedance spectroscopy,7,8 optical absorption spectroscopy,9,10 and methods based on optical refractometry.11−24 Among devices of aforementioned technologies, label-free optical biosensors relied on refractive index change induced by immobilization of analyte molecules on sensing surface of a sensor head made up of optical components such as prisms and © 2017 American Chemical Society

Received: June 27, 2017 Accepted: August 29, 2017 Published: August 29, 2017 31478

DOI: 10.1021/acsami.7b09252 ACS Appl. Mater. Interfaces 2017, 9, 31478−31487

Research Article

ACS Applied Materials & Interfaces

Figure 1. Waveguide mode mismatch for an optical biosensor. Variation of an index of liquid cladding controls the mode mismatch degree, leading to the power transmission change.

head that would suffer no wear and tear of such overlayers. However, a complex detection unit such as a lock-in detection device was employed to detect the relevant phase (for the purpose of avoiding noise possibly arising from ambient light fluctuations). The resolution of glucose concentration of about 0.1 mg/dL has been demonstrated,44 known to be better than the fiber optical SPR glucose sensors.45−47 However, the relevant working principles of such fiber-based sensors without metals has not been clearly manifested. Furthermore, a bulky complex detection system such as a lock-in amplifier needed to be replaced by a somewhat simpler and compact detection scheme which compromised no degraded sensitivity of the sensor system for point-of-care-testing applications. In this paper, we presented a label-free optical biosensors made up of a multimode optical fiber, the plastic cladding of which was removed along a certain length (5 cm) for the silica core exposure into which analyte liquid is injected to form the “liquid cladding”. Over the region of the so-called liquid cladding, we modulated the refractive index of the cladding as a function of injected analyte concentration. This lead to the concentration-controlled modulation of degree of waveguide mode mismatch across the interface between the waveguides of the liquid cladding and the solid plastic one. We performed the quantitative detection of concentration of glycerol and protein, i.e., bovine serum albumin (BSA), with appropriate chemical modification of the sensing surface. For sensing, we simply measured optical power at the fiber output as a function of concentration of analyte injected. We exploited availability of higher output power of visible light propagating along the multimode fiber than that with a single mode, thus escalating possibility that it could overshadow the ambient light fluctuation for enhanced signal-to-noise ratio. We found that reasonably good agreement with the experimental measurement was possible through application of the theoretical model presented to include modulation of waveguide mode mismatch and evanescent field absorption. This model provided a reasonable nonlinear fit to the measurement, permitting a proper design of the presented sensor with sufficiently high sensitivity for a specific concentration range. We obtained the glycerol detection limit of ∼1.02 × 10−6 RIU (∼0.001% glycerol v/v) which was comparable to that reported by the SPR fiber sensor that used the same multimode

limitation of eliminating unwanted external disturbance incoherently exerted on multiple waveguides. Optical waveguide biosensors where total internal reflection (TIR) excited guided modes used interaction of TIR-produced evanescent fields with analyte injected.33−35 A high refractive index cladding media such as TiO2 enhanced the evanescent field strength, aiming at increased sensitivity. The enhanced evanescent coupling with analyte, however, encountered the decreased mode confinement factor, thus giving rise to decreased output power at the waveguide output. This compromised the signal-to-noise ratio, partly accounting for the sensor sensitivity being hardly better than that of the order of magnitude of 10−3 RIU (a minimum resolvable refractive index).35 Surface plasmon resonance (SPR) excited via overlaying thin layers of noble metals on a waveguide core have been utilized to build optical refractometers, aiming at label-free biosensors.36−39 Metals with thicknesses on the order of tens of nanometers coated on a waveguide core exposure allowed the SPR waves to couple evanescently with injected analytes, and this waveguide based SPR technique (no prism coupled SPR) has demonstrated the detection limits of the order of magnitude of 10−4−10−6 RIU.40−43 These sensitivities being higher than those obtained by TIR evanescent field coupling was attributed to the sensitive condition for the plasmonic resonance and the enhanced strength of SPR evanescent fields. However, unexpected variation of deposited metal thickness on a scale of nanometers over the surface area of sensors might hamper reproducibility and reliable calibration of the sensor signals. Meanwhile, it was reported that a single mode fiber has been used for glucose sensors using enzymes (glucose oxidase) preimmobilized on its core exposure obtained by removal of part of the fiber silica cladding (using toxic acids).44 The injected glucose then reacted with the aid of an enzyme to generate the secondary products, leading to change in a refractive index immediately above the core. The polarizationdependent phase change in the fiber mode could then be detected as a function of injected glucose concentration with a lock-in detection technique. This fiber had no metal coating on its core exposure, i.e., no use of SPR phenomenon, for sensing, being favorable for simplified fabrication of the relevant sensor 31479


Research Article


Figure 2. Schematic of silica surface treatment for BSA bonding: (a) Creation of hydroxyl functional groups on a silica substrate by a piranha process. (b) Production of amine-modified surface by subsequent treatment with APTES. (c) Creation of carboxyl functional groups by SA on the amine-modified surface.

fiber platform even with the longer sensor head.48 In addition, upon BSA injection the detection limit of ∼1 ng/mL was demonstrated. It was revealed that the presented optical biosensors need neither plasmonic metal coating nor high index dielectric film coating for enhanced evanescent field strength49,50 and did not compromise the detection sensitivity (comparable to or even better than the fiber based SPR sensors). The presented fiber sensor exhibited reasonably good reproducibility of the sensor signals, e.g., coefficient of variation (CV) of less than 4%. More interestingly, the device sensing length played little role in the sensor characteristics such as the sensitivity, due to the signal transduction only relying on the degree of the waveguide mode mismatch across the interface in the absence of significant absorption of light. This lead to a great possibility of fabricating label-free biochemical sensors that could be made to fit nearly any size desired for a given application in a highly simple format at an inexpensive cost, still ensuring reasonably sufficient sensitivity. We observed that the sensor response was subject to change in both the refractive indices of the liquid bulk and the layers immobilized immediately above the sensing surface. It was seen that the latter index effects could dominate over the former ones as would be shown in the BSA detection. Thus, the presented sensing platform enabled the real-time monitoring of the relevant binding kinetics information to be accessible.

n12 − n22 , where n1 and n2 are the refractive indices of the core and cladding of the multimode fiber, respectively. As optical modes of waveguide A enter waveguide B that has the different cladding (air or liquid), optical power of those modes in waveguide A is redistributed amid the eigen-modes of waveguide B without attenuation of total power. This is due to the fact that waveguide B has an NA larger than that of waveguide A.51,52 However, this is not the case where the optical modes excited in waveguide B enter waveguide C which has a smaller NA. The optical modes in waveguide B cannot couple to the modes in waveguide C without power loss, which is attributed to the mode mismatch created radiation modes. This leads to the drop in optical power transmission at the output of waveguide C in accordance with the ratio between the squared NAs of two waveguides, i.e., (NAC)2/(NAB)2.51−53 The control of the mode mismatch degree, which dominates the power transmission, plays a key role for the sensing principle and signal transduction occurring in the presented sensor platform. As shown in Figure 1, analyte liquid is injected onto the core of waveguide B to substitute for air cladding. This results in formation of liquid cladding and thus reduces gap between two different NAs of two waveguides, B and C (= A), toward a reduced degree of mode mismatch. This increases optical power at the output of waveguide C with increasing concentration of the analyte, allowing quantitative detection. A further increase in analyte concentration may decrease the output power due to evanescent light absorption by analyte liquid during the propagation along waveguide B. Thus, the transmission of optical power at the output of waveguide C can be given in the form of

2. SENSING PRINCIPLE Figure 1 shows three different waveguides serially interconnected through sharing the same core, i.e., waveguides A, B, and C. We assume that waveguides A and C have the same solid cladding while waveguide B has air cladding. One can inject liquid onto the core surface of waveguide B to replace its air cladding, thus forming the liquid cladding around the core. Different concentrations of liquid yield different refractive indices, permiting modulation of the liquid cladding index. Let us assume electromagnetic fields propagate along a waveguide, with their core confined component of the electric fields being a linear superposition of a number of modes as follows: Ei(x , y , z) =

∑ Ci ,m fi ,m (x , y) e−izβ

i ,m

T = T0 + e−(γ + αc)L

0.372 [n12 − nlc2(c)]

(2)

where T0 is the background transmission of optical power through incomplete mode coupling, γ is responsible for the light scattering loss possibly due to roughness of core exposure surface induced by imperfect removal of the plastic cladding, and α accounts for power attenuation during the light propagation along waveguide B of length L via absorption of evanescent light as a function of analyte liquid concentration c. Waveguide C (or A) has an NA of 0.37, while waveguide B has

(1)

where Ci,m is the coefficient of the eigen-mode function f i,m(x, y) for the mth mode, and βi,m is the propagation constant for the mth mode in waveguide i (i = A, B, and C). The number of modes excited along propagation is determined by the waveguide geometry and its numerical aperture, i.e., NA =

an NA of n12 − nlc2 where nlc(c) is the refractive index of the liquid cladding as a function of c. It is noted that use of the highest possible slope of differential power change with respect to concentration near zero facilitates minimization of minimum 31480


Research Article


Figure 3. FTIR spectroscopic transmission: (a) SiO2 substrate, and (b) SiO2 substrate with its surface chemically modified.

Table 1. Peaks of the FTIR Spectra for Various Functional Groups wavenumbers (cm−1) mode of vibrations

465

680

798

942

1088

1577

3380−3500

3380

1658

δ Si−O−Si

δout‑plane N−H

υ O−Si−O

υ Si−OH

υas Si−O−Si

δin‑plane N−H

−OH stretch

−NH stretch

−COOH stretch

Figure 4. (a) Clad-free fiber core as a sensing surface. (b) Ring-shaped flow cell. (c) Schematic of the experimental setup for the real-time continuous monitoring of label-free optical biosensor.

3. EXPERIMENTAL APPARATUS AND TECHNIQUES

detectable concentration of analyte. This optimum slope can be found at a certain range of a mode mismatch degree, for which increasing concentration leads to the drastic power increase.

3.1. Materials and Reagents. Deionized water with a resistivity of 18 MΩ·cm was produced by a Millipore milli-Q (MQ) water 31481


Research Article


liquid of 400 μL volume to be filled around a cylindrical surface of the 5 cm long sensor head. Care was taken to make sure that we had no air bubbles in the solution which was injected into the sensor surface by a pump (EYELA, SMP-21). The pump flow rate used to inject solutions of the capture probe (antibody) or the analyte (glycerol or BSA) was 5 μL/s, while that for surface rinsing with PBS was 10 μL/s. Figure 4c shows a schematic of an optical biosensing setup. We coupled a 5 mW He−Ne laser (λ = 632.8 nm, HNL050L, Thorlabs) into one end of the fiber. This was fulfilled by free-space end-fire coupling into the fiber via a three-dimensional translational stage where an antireflection (AR)-coated lens of 0.25 NA (16.5 mm focal length) was mounted. Optical power at the other end of the fiber was collected by an AR coated aspheric lens of 0.53 NA, for continuous monitoring by a power meter interfaced with a computer. The standard deviation of the laser power which must have played a role in estimating the LOD for the presented sensor system was approximately 2.0145 × 10−6 W over 5 h. 3.4. Preparation of Capture Probe and Analyte. We used the linear relationship between concentration of glycerol solution and its refractive index to characterize the presented fiber sensor as a refractometer. The refractive index of the glycerol solution, ng has the form of

purification system for preparation of aqueous solutions and cleaning of the silica surface. Methanol, (3-aminopropyl)-triethoxysilane (APTES), succinic anhydride (SA), 1,4-dioxane, toluene, 1-ethyl-3(3-(dimethylamino)propyl) carbodiimide (EDC), N-hydroxysuccinimide (NHS), BSA, and phosphate-buffered saline (PBS) were purchased from Sigma-Aldrich Co. (St Louis, MO, USA). Sulfuric acid (H2SO4 98%), hydrogen peroxide (30%), and glycerol (99%) were purchased from Duksan Pure Chemicals Co. (LTD, Korea). Heavily p-doped silica wafers of (110) orientation (Waferpro. Co., San Jose, CA, USA), 1 Ω·cm resistivity, and 275 μm thickness were cut into 1.5 cm × 1.5 cm square pieces. 3.2. Surface Modification for BSA Bonding. Figure 2a−c illustrates chemical modification of the fiber core surface of silica for BSA bonding. On its surface, hydroxyl groups (−OH) were produced in a standard process with piranha solution (H2SO4/H2O2 = 4:1 v/v) as shown in Figure 2a. Then the surface was immersed in APTES solution of 2% in toluene on a shaker for 24 h at room temperature to generate amine groups with silanization as shown in Figure 2b. The amine-modified silica surface was washed with toluene and methanol before being cured under argon gas for 30 min at 100 °C and was immersed in the MQ water for 2 h at 40 °C to remove unreacted ethoxy groups (−O−CH2−CH3) in APTES. The surface was washed again with methanol and dried in a vacuum oven at 60 °C for 30 min. Figure 2c shows the generation of carboxyl groups by subsequent immersion of the modified surface in succinic anhydride (SA) solution of 20 mg/mL in 1,4-dioxane for 30 min at 80 °C. The surface was then washed with methanol and dried in vacuum oven for 30 min at 60 °C (with 1,4-dioxane heavily used as a stabilizer for the acid anhydrides organic solvent). Aforementioned functional groups produced step-bystep were confirmed by spectroscopic measurement of molecular absorption via the instrument, Fourier transformation infrared spectroscopy (FTIR) (JASCO FT/IR-4700) with the scanning range of 4500−400 cm−1. Figure 3a,b shows the FTIR spectra of the silica surface without and with its chemical modification, respectively, while Table 1 offers the reference data for absorption modes of functional groups at infrared light wavenumber. We observed light absorption of O−Si−O by way of its symmetric stretch vibrational mode and the bending vibrational mode at ∼800 and 461−467 cm−1, respectively.54,55 In addition, absorption of the asymmetric stretch vibrational mode of Si−O−Si was observed at 1088 cm−1. All these absorption bands indicated the presence of silicon dioxide. Comparison of Figure 3a,b revealed that the chemical surface treatment caused substantial absorption over the spectral band of 3380−3500 cm−1 due to both the stretch vibrational mode of the hydroxyl group found in water hydrogen-bonds and the Si−OH stretch vibrational mode of the silanol hydrogen bonded to a water molecule (SiO−H···H2O).55−57 The −NH group generated by the APTES treatment on the surface of SiO2 could be confirmed by the bands at 680 and 1577 cm−1 that indicated the presence of the N−H out-of-plane and the N−H in-plane bending vibrations, respectively.58 In addition, the absorption band that ranged from 3200 to 3500 cm−1 unfolded the formation of the −NH stretch vibrational mode.59 At around 942 cm−1, we observed absorption due to Si−OH in-plane vibration of a silanol group.55,60 The presence of a carboxylic acid group introduced as described in Figure 2c was confirmed by the pronounced absorption band at around 1658 cm−1 as shown in Figure 3b.61,62 3.3. Label-Free Fiber Optical Sensor System. We used a multimode optical fiber of silica core with a plastic cladding (JFTLHPolymicro Technologies, USA) for a sensor head. It had an NA of 0.37 (n1 = 1.457, ncl = 1.41), and a core diameter of 200 μm. We removed part of the cladding layer using a soldering iron. This was followed by the subsequent cleansing with solution of acetone and ethanol to obtain the 5 cm long cladding-free core exposure to be used as the surface of the sensor head, as shown in Figure 4a. The fiber sensor head was encompassed by a ring-shaped flow cell molded out of polydimethylsiloxane (PDMS, Sylgard 184, Dow Corning Co., USA) with two ports for the inlet and outlet through which plastic tubing was connected (see Figure 4b). The flow cell structure allowed the

ng = n w + ηC

(3)

where nw is the refractive index of water, C is the concentration of the glycerol solution in volume-to-volume ratio [%], and η is the proportionality coefficient. Through a digital magnetic stirrer (1000 rpm spin), we diluted glycerol with deionized water to generate various concentrations from 0 to 30% and checked the corresponding refractive indices by an Abbe refractometer (Digital Abbe Refractometer, Atago, DR-A1) as shown in Figure 5. The linear fit was used to convert its concentration to the corresponding refractive index for estimation of NA of waveguide sensor with liquid glycerol cladding.

Figure 5. Refractive indices of glycerol at various concentrations. We also applied the presented sensor platform to specific sensing of BSA molecules by modifying the core surface chemically with prior immobilization of the BSA antibody. The surface modification included generation of the carboxyl groups on the core surface (in Figure 2c), and the subsequent activation by EDC and NHS. The BSA concentrations used for detection were from 10 to 1000 ng/mL at pH 7.4.

4. RESULTS AND DISCUSSION Figure 6 shows optical power transmission as a function of glycerol concentration and its nonlinear fit with eq 2. It rapidly increased with increasing concentration near zero. This was 31482


Research Article


we attributed to enhancement of absorption of evanescent light confined within about 200 nm above the fiber core surface by the liquid cladding. This resulted in higher loss of propagating optical mode through waveguide B. This eventually led to the signal decrease. As seen in Figure 6, at low concentrations, transmission increase by mode mismatch reduction was dominant over the evanescent light absorption effects. These absorption effects, however, became substantial at high concentrations, yielding transmission decrease down to a certain level which could be determined by a mode confinement factor of a waveguide B. The reasonably good agreement between the measured data and the nonlinear fit based on eq 2 indicated that one can design the presented waveguide sensor with sufficiently high sensitivity for a specific range of low concentrations of analyte, taking into account their absorption and refractive indices. We experimentally obtained the LOD of 0.002% glycerol concentration, defined as the concentration that caused the signal to change three times as much as the standard deviation of the signal measured at a blank analyte.48 The minimum detectable refractive index change corresponding to the LOD was then 1.02 × 10−6 RIU, which turned out to be lower than that obtained from the SPR fiber sensor of the same multimode fiber.48 The fiber sensor characterization with glycerol was preceded by the quantitative immune-assay of a BSA as shown in Figure 7a,b. We started with injection of PBS, yielding a time-stable baseline signal, followed by the EDC-NHS injection for carboxyl group activation. To explain the rapid increase in transmission upon EDC-NHS injection, we invoked the fact that the refractive index of the EDC-NHS, i.e., 1.3563 was higher than that of PBS, i.e., 1.3373.64 We checked the absorption properties of BSA solution, EDC-NHS, and PBS by a UV−vis spectrophotometer. It was found that the negligible absorption of light at the wavelength (632.8 nm) occurred in all the PBS, EDC-NHS, and BSA solutions, compared to that of the glycerol solution. This indicated the fact that modulation of the mode mismatch degree in the presented fiber device played a dominant role in determining the sensor response, through the index engineering only.

Figure 6. Fiber sensor response to the change in glycerol concentration.

followed by the rapid decrease of the transmission, which was then saturated as the concentration further rises above 5%. The inset shows the quite steep slope of the transmission increase with concentration from 0 to 0.1% due to the reduction of mode mismatch mentioned above. From the fitting, we obtained the loss parameter of γ ≈ 0.134 m−1 that accounts for the scattering loss during propagation along 5 cm sensor length and absorption coefficient of α ≈ 0.026 m−1 %−1 for evanescent light absorption of glycerol per its unit concentration on the sensing surface. According to eq 2, increasing liquid concentration could contribute to transmission change in both directions, i.e., the increase and decrease in transmission. Its increasing concentration would increase the effective index of the liquid cladding and thus alleviate mode mismatch between waveguides B and C, resulting in raising transmission of optical power through the interface between two waveguides. This is described by the denominator of the second term of eq 2. However, a higher concentration of glycerol caused the signal to decrease, which

Figure 7. Time-dependent optical power transmitted through the fiber sensor upon injection of serial BSA solutions of various concentrations: (a) Concentration range of 10−100 ng/mL. (b) Concentration range of 100−1000 ng/mL. 31483


Research Article

ACS Applied Materials & Interfaces We used the two different ranges of the BSA concentrations for its quantitative assay, i.e., 10−100 ng/mL and 100−1000 ng/mL, as shown in Figure 7a,b. Within each concentration range, we injected the BSA solution of several concentrations in a row, with each injection of the BSA concentration being accompanied by the prior rinsing of the surface by PBS. This serial BSA injection implied that the resultant BSA concentration at the point when BSA of a given concentration was injected was the sum of all the concentrations of injected BSA including those preceding it. After probing the sensor signals upon injection of a series of BSA concentrations in each range, we replaced the fiber sensor head by a new one of the same type, which had its surface chemically ready (the same as the previous one) for capturing BSA of serial concentration in another range. Prior to BSA injection, we used an Abbe refractometer with the index resolution of 10−4 RIU to measure the refractive indices of BSA solutions at all the concentrations used. They were all determined to be the same, 1.3352, implying the indiscernibility between the different concentrations due to its limited resolution. In contrast, injection of the additional BSA solution of various concentrations into the presented fiber sensor yielded the immediate transmission increase, indicating distinguishability between concentrations probed as shown in Figure 7a,b. The transmission increase became more remarkable at a range of the higher concentrations as compared between Figure 7a,b, unveiling possibility of quantitating the immune protein assay. In each range of concentrations probed, the foremost injection of BSA that followed EDC-NHS injection lead to an immediate increase in transmitted powers, despite the fact that the refractive index of EDC-NHS (1.35) was higher than that of the BSA solution (1.3352). Given the BSA refractive index of 1.38,65,66 this implied that the injected BSA molecules immediately immobilized on the surface via amino-carboxyl coupling and then occupied the space immediately above it (the signal increase occurred over about 1 min despite the fact that it appeared to increase abruptly as shown in Figure 7a,b). This lead to the instantaneous increase in the effective index of the liquid cladding. It could be inferred that the sensor device with its index sensitivity was stronger in a region in close proximity to the surface than that in bulk and could support real-time monitoring of kinetics of time-dependent biochemical bindings such as immuno and bioaffinity reactions. We observed the signal saturation that followed its rapid increase at the lower concentrations of BSA (Figure 7a), while at the higher concentrations the gradual decrease in the signal (Figure 7b) was preceded by its rapid increase. This was possibly due to the fact that higher concentration of BSA caused to increase the BSA−BSA coupling strength and weaken amino-carboxylic coupling, thus reducing the number of BSA molecules immobilized on the surface. Figure 8 shows the normalized transmitted power versus concentration of the BSA solution. They were obtained by normalizing the transmitted power change with respect to the baseline of the transmitted power. Care had to be taken to set the power baseline from which the quantitative power change induced by the BSA injection was estimated, taking into account the bulk refractive index given by the buffer of the BSA solution. The bulk index set by the BSA buffer was quite similar to that of PBS permitted us to use the power level measured at the foremost PBS injection as the power baseline for both such estimation of the BSA induced power change and for further

Figure 8. Normalized change in the transmitted power at injection of BSA concentrations (10−1000 ng/mL).

normalization. The normalized power change increased monotonically with the BSA concentration, as shown in Figure 8, revealing the capability of quantitating BSA immune assay. The experimentally measured standard deviation of the signal led us to reach the BSA LOD of ∼3.7 ng/mL. The corresponding mass coverage was estimated as ∼30 pg/mm2, taking into account the structural information including the sensing surface, i.e., the area of the cylindrical surface of the fiber sensor head and the volume of the liquid injected into the flow cell. The mass coverage proved to be only about 1/50 as large as the monolayer coverage of BSA molecules (∼1.4 ng/ mm2),67 reflecting the sufficient sensitivity of the presented device to the index change in a region adjacent the sensing surface. The sensitivity obtainable without any additional coating of metal or dielectric layers on the fiber core was seen to favor fabrication of a sensitive mini-sized, compact label-free biosensor that would suffer no wear and tear of the additional overlayers. We checked the dependence of the sensor characteristics including sensitivity on its length using a chemical solvent with its concentration low enough to absorb little light. It was revealed that varying the sensor length influenced little of its performance. This was attributed to the fact that sensor signal transduction occurred only across the interface where a decrease in numerical apertures of interconnected waveguides took place along the light propagation. This length independence could enhance flexibility of the device dimensional design into a size desired for a given application. We also checked the coefficient of variation of the sensor signals, which at less than 4% was reasonably good reproducibility of the presented fiber sensor signal.

5. CONCLUSIONS We used a multimode optical fiber whose plastic cladding had been removed along a certain length to form liquid cladding of a waveguide for quantitative assay of concentrations of chemical and biological substance in a real time format. A multimode fiber was adopted as a basic platform of the sensor device due to the availability of higher optical power transmitted through it at visible wavelength for higher signal-to-noise ratio (thus suppressing coefficient of variation of the sensor). The 31484


Research Article


free Real Time Protein Detection. Sens. Actuators, B 2013, 177, 848− 855. (4) Wei, F.; Patel, P.; Liao, W.; Chaudhry, K.; Zhang, L.; ArellanoGarcia, M.; Hu, S.; Elashoff, D.; Zhou, H.; Shukla, S.; Shah, F.; Ho, C.M.; Wong, D. T. Electrochemical Sensor for Multiplex Biomarkers Detection. Clin. Cancer Res. 2009, 15, 4446−4452. (5) Ikebukuro, K.; Kiyohara, C.; Sode, K. Novel Electrochemical Sensor System for Protein Using the Aptamers in Sandwich Manner. Biosens. Bioelectron. 2005, 20, 2168−2172. (6) Luo, X.; Davis, J. J. Electrical Biosensors and the Label Free Detection of Protein Disease Biomarkers. Chem. Soc. Rev. 2013, 42, 5944−5962. (7) Affanni, A.; Corazza, A.; Esposito, G.; Fogolari, F.; Polano, M. Protein Aggregation Measurement through Electrical Impedance Spectroscopy. J. Phys.: Conf. Ser. 2013, 459, 012049. (8) Andrade, C. A. S.; de Oliveira, H. P.; Oliveira, M. D. L.; Correia, M. T. S.; Coelho, L. C. B. B.; de Melo, C. P. Protein Unfolding Studied by Fluorescence Methods and Electrical Impedance Spectroscopy: The Cases of Cratylia mollis and Concanavalin A. Colloids Surf., B 2011, 88, 100−107. (9) Xia, F.; Zuo, X.; Yang, R.; Xiao, Y.; Kang, D.; Vallée-Bélisle, A.; Gong, X.; Yuen, J. D.; Hsu, B. B. Y.; Heeger, A. J.; Plaxco, K. W. Colorimetric Detection of DNA, Small Molecules, proteins, and Ions Using Unmodified Gold Nanoparticles and Conjugated Polyelectrolytes. Proc. Natl. Acad. Sci. U. S. A. 2010, 107, 10837−10841. (10) Cupane, A.; Bologna, C.; Rizzo, O.; Vitrano, E.; Cordone, L. Local Dynamics of DNA Probed with Optical Absorption Spectroscopy of Bound Ethidium Bromide. Biophys. J. 1997, 73, 959−965. (11) Sharma, P.; Sharan, P. Design of Photonic Crystal-Based Biosensor for Detection of Glucose Concentration in Urine. IEEE Sens. J. 2015, 15, 1035−1042. (12) Jung, D. Biosensor Based on Distributed Bragg Reflector Photonic Crystals for the Detection of Protein A. J. Chosun Nat. Sci. 2010, 3, 33−37. (13) Kehl, F.; Bischof, D.; Michler, M.; Keka, M.; Stanley, R. Design of a Label-Free, Distributed Bragg Grating Resonator Based Dielectric Waveguide Biosensor. Photonics 2015, 2, 124−138. (14) Chen, J.; Shi, S.; Su, R.; Qi, W.; Huang, R.; Wang, M.; Wang, L.; He, Z. Optimization and Application of Reflective LSPR Optical Fiber Biosensors Based on Silver Nanoparticles. Sensors 2015, 15, 12205− 12217. (15) Mohanty, G.; Sahoo, B. K.; Akhtar, J. Comparative Analysis for Reflectivity of Graphene Based SPR Biosensor. Opt. Quantum Electron. 2015, 47, 1911−1918. (16) Kidd, A.; Guieu, V.; Perrier, S.; Ravelet, C.; Peyrin, E. Fluorescence Polarization Biosensor Based on an Aptamer Enzymatic Cleavage Protection Strategy. Anal. Bioanal. Chem. 2011, 401, 3229− 3234. (17) Zhang, X.; Daaboul, G. G.; Spuhler, P. S.; Dröge, P.; Ü nlü, M. S. Quantitative Characterization of Conformational-Specific protein− DNA Binding Using a Dual-Spectral Interferometric Imaging Biosensor. Nanoscale 2016, 8, 5587−5598. (18) Lee, M.; Fauchet, P. M. Two-Dimensional Silicon Photonic Crystal Based Biosensing Platform for Protein Detection. Opt. Express 2007, 15, 4530−4535. (19) Khan, M.; Khan, A. R.; Shin, J.-H.; Park, S.-Y. A Liquid-CrystalBased DNA Biosensor for Pathogen Detection. Sci. Rep. 2016, 6, 22676. (20) Kim, D.-K.; Kerman, K.; Saito, M.; Sathuluri, R. R.; Endo, T.; Yamamura, S.; Kwon, Y.-S.; Tamiya, E. Label-Free DNA Biosensor Based on Localized Surface Plasmon Resonance Coupled with Interferometry. Anal. Chem. 2007, 79, 1855−1864. (21) Pollet, J.; Delport, F.; Janssen, K. P.; Jans, K.; Maes, G.; Pfeiffer, H.; Wevers, M.; Lammertyn, J. Fiber Optic SPR Biosensing of DNA Hybridization and DNA-Protein Interactions. Biosens. Bioelectron. 2009, 25, 864−869. (22) Sepúlveda, B.; Carrascosa, L. G.; Regatos, D.; Otte, M. A.; Fariña, D.; Lechuga, L. M. Surface Plasmon Resonance Biosensors for

concentration changes of the liquid cladding gave rise to redistribution of the waveguide mode power, leading to a change in the degree of the waveguide mode mismatch between waveguides of liquid and solid claddings. Experimental results with the presented device used as a refractometer showed good agreement with the theoretical model which included transmitted power change induced both by the waveguide mode mismatch and by evanescent field absorption. This enabled a suitable design of the waveguide device with liquid cladding, which could be used as a liquid refractometer or label-free optical biosensor for a specific range of concentrations of a given biological analyte. We demonstrated the LOD of glycerol concentration, 0.001% (v/v), which corresponded to the minimum resolvable index of ∼1.02 × 10−6 RIU, being comparable to those reported in the plasmonic fiber sensors that used the same fibers. The LOD of BSA concentration was experimentally obtained as ∼3.7 ng/mL (corresponding mass coverage of ∼30 pg/mm2). It was also revealed that the device sensitivity arose from changes in the refractive indices both of liquid bulk and layers just above the sensing surface, with the latter effects dominant over the former. Application of the presented sensor device could be extended to the immunoassay of other kinds of proteins unless their strong absorption of the visible light occurred. The sensing length independence of the sensor characteristics such as sensitivity could facilitate design of the device dimensions into a desired size for given applications. The presented device permitted us to find its use in a label-free biochemical sensors of a miniaturized format which was possibly suited to the point-of-care testing. This was due to the highly simple structure which facilitated its fabrication by excluding coating of plasmonic metal film or additional dielectric one.

■

AUTHOR INFORMATION

Corresponding Author

*Tel./Fax: +82 31 750 8552. E-mail: [email protected]. ORCID

Heongkyu Ju: 0000-0003-0223-7864 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (NRF2017R1D1A1B03033987) and also supported by Materials and Components Technology Development Program program of MOTIE/KEIT [10053617, Development of the POCT platform and disposable PCR Chip of semiconductive elements of optical source and detection].

■

REFERENCES

(1) Guo, Y.; Ye, J. Y.; Divin, C.; Thomas, T. P.; Myc, A.; BersanoBegey, T. F.; Baker, J. R., Jr.; Norris, T. B. Label-Free Biosensing Using a Photonic Crystal Structure in a Total-Internal-Reflection Geometry. Proc. SPIE 2009, 7188, 71880B−71880B12. (2) Daniels, J. S.; Pourmand, N. Label-Free Impedance Biosensors: Opportunities and Challenges. Electroanalysis 2007, 19, 1239−1257. (3) Esfandyarpour, R.; Esfandyarpour, H.; Javanmard, M.; Harris, J. S.; Davis, R. W. Microneedle Biosensor: A Method for Direct Label31485


Research Article

ACS Applied Materials & Interfaces Highly Sensitive Detection in Real Samples. Proc. SPIE 2009, 7397, 73970Y. (23) Hossain, B.; Rana, M. DNA Hybridization Detection Based on Resonance Frequency Readout in Graphene on Au SPR Biosensor. J. Sens. 2016, 2016, 1. (24) Nguyen, T. T.; Han, G.-R.; Jang, C.-H.; Ju, H. Optical Birefringence of Liquid Crystals for Label-Free Optical Biosensing Diagnosis. Int. J. Nanomed. 2015, 10, 25−32. (25) Wu, D.; Zhu, T.; Wang, G.-Y.; Fu, J.-Y.; Lin, X.-G.; Gou, G.-L. Intrinsic Fiber-Optic Fabry-Perot Interferometer Based on Arc Discharge and Single-Mode Fiber. Appl. Opt. 2013, 52, 2670−2675. (26) Zhu, T.; Wu, D.; Liu, M.; Duan, D.-W. In-Line Fiber Optic Interferometric Sensors in Ssingle-Mode Fibers. Sensors 2012, 12, 10430−10449. (27) Sharma, U.; Wei, X. In Fiber Optic Sensing and Imaging; Kang, J. U., Ed.; Springer: New York, 2013; Chapter 2, pp 29−54. (28) Thompson, R. B. In Advanced Concepts in Fluorescence Sensing: Part B: Macromolecular Sensing; Geddes, C. D., Lakowicz, J. R, Eds.; Springer: New York, 2005; Chapter pp 1−19. (29) Magnusson, R.; Wawro, D.; Zimmerman, S.; Ding, Y. Resonant Photonic Biosensors with Polarization-Based Multiparametric Discrimination in Each Channel. Sensors 2011, 11, 1476−1488. (30) Cennamo, N.; Massarotti, D.; Conte, L.; Zeni, L. Low Cost Sensors Based on SPR in a Plastic Optical Fiber for Biosensor Implementation. Sensors 2011, 11, 11752−11760. (31) Gouveia, C. A. J; Baptista, J. M.; Jorge, P. A. S. In Current Developments in Optical Fiber Technology; Harun, S. W., Arof, H., Eds.; InTech: Rijeka, Croatia, 2013; Chapter 13, pp 345−373. (32) Miao, Y.-p.; Liu, B.; Zhao, Q.-d. Refractive Index Sensor Based on Measuring the Transmission Power of Tilted Fiber Bragg Grating. Opt. Fiber Technol. 2009, 15, 233−236. (33) Kwon, S. W.; Yang, W. S.; Lee, H. M.; Kim, W. K.; Son, G. S.; Yoon, D. H.; Lee, S.-D.; Lee, H.-Y. The Fabrication of Polymer-Based Evanescent Optical Waveguide for Biosensing. Appl. Surf. Sci. 2009, 255, 5466−5470. (34) Canning, J.; Padden, W.; Boskovic, D.; Naqshbandi, M.; de Bruyn, H.; Crossley, M. J. Manipulating and controlling the evanescent field within optical waveguides using high index nanolayers. Opt. Mater. Express 2011, 1, 192−200. (35) Son, G.-S.; Kwon, S.-W.; Kim, W.-K.; Yang, W.-S.; Lee, H.-M.; Lee, H.-Y.; Lee, S.-D.; Lee, S.-S. Refractive-Index Sensor Incorporating a Polymeric Waveguide Overlaid with TiO2 Thin Films. Proc. SPIE 2009, 7218, 721818−1−721818−10. (36) Homola, J.; Vaisocherová, H.; Dostálek, J.; Piliarik, M. MultiAnalyte Surface Plasmon Resonance Biosensing. Methods 2005, 37, 26−36. (37) Homola, J.; Koudela, I.; Yee, S. S. Surface Plasmon Resonance Sensors Based on Diffraction Gratings and Prism Couplers: Sensitivity Comparison. Sens. Actuators, B 1999, 54, 16−24. (38) Homola, J.; Yee, S. S.; Gauglitz, G. Surface Plasmon Resonance Sensors: Review. Sens. Actuators, B 1999, 54, 3−15. (39) Namira, E.; Hayashida, T.; Arakawa, T. Surface Plasmon Resonance Study for the Detection of Some Biomolecules. Sens. Actuators, B 1995, 24, 142−144. (40) Debackere, P.; Taillaert, D.; De Vos, K.; Scheerlinck, S.; Bienstman, P.; Baets, R. Si Based Waveguide and Surface Plasmon Sensors. Proc. SPIE 2007, 6477, 647719. (41) Homola, J.; Č tyroký, J.; Skalský, M.; Hradilová, J.; Kolárǒ vá, P. A Surface Plasmon Resonance Based Integrated Optical Sensor. Sens. Actuators, B 1997, 39, 286−290. (42) Prieto, F.; Sepúlveda, B.; Calle, A.; Llobera, A.; Domínguez, C.; Abad, A.; Montoya, A.; Lechuga, L. M. An Integrated Optical Interferometric Nanodevice Based on Silicon Technology for Biosensor Applications. Nanotechnology 2003, 14, 907−912. (43) Č tyroký, J.; Homola, J.; Lambeck, P. V.; Musa, S.; Hoekstra, H. J. W. M.; Harris, R. D.; Wilkinson, J. S.; Usievich, B.; Lyndin, N. M. Theory and Modelling of Optical Waveguide Sensors Utilising Surface Plasmon Resonance. Sens. Actuators, B 1999, 54, 66−73.

(44) Lin, T.-Q.; Lu, Y.-L.; Hsu, C.-C. Fabrication of Glucose Fiber Sensor Based on Immobilized GOD Technique for Rapid Measurement. Opt. Express 2010, 18, 27560−27566. (45) Verma, R.; Gupta, B. D. A Novel Approach for Simultaneous Sensing of Urea and Glucose by SPR Based Optical Fiber Multianalyte Sensor. Analyst 2014, 139, 1449−1455. (46) Srivastava, S. K.; Arora, V.; Sapra, S.; Gupta, B. D. Localized Surface Plasmon Resonance-Based Fiber Optic U-Shaped Biosensor for the Detection of Blood Glucose. Plasmonics 2012, 7, 261−268. (47) Singh, S.; Gupta, B. D. Fabrication and Characterization of a Surface Plasmon Resonance Based Fiber Optic Sensor Using Gel Entrapment Technique for the Detection of Low Glucose Concentration. Sens. Actuators, B 2013, 177, 589−595. (48) Nguyen, T. T.; Lee, E.-C.; Ju, H. Bimetal Coated Optical Fiber Sensors Based on Surface Plasmon Resonance Induced Change in Birefringence and Intensity. Opt. Express 2014, 22, 5590−5598. (49) Chiu, N.-F.; Tu, Y.-C.; Huang, T.-Y. Enhanced Sensitivity of Anti-Symmetrically Structured Surface Plasmon Resonance Sensors with Zinc Oxide Intermediate Layers. Sensors 2014, 14, 170−187. (50) Quigley, G. R.; Harris, R. D.; Wilkinson, J. S. Sensitivity Enhancement of Integrated Optical Sensors by Use of Thin HighIndex Films. Appl. Opt. 1999, 38, 6036−6039. (51) Okamoto, K. Fundamentals of Optical Waveguides, 2nd ed.; Academic: San Diego, CA, 2006. (52) Saleh, B. E. A.; Teich, M. C. Fundamentals of Photonics; John Wiley: Chichester, U.K., 1991. (53) Esposito, J. J. In Handbook of Fiber Optics, Theory and Applications; Wolf, H. F., Ed.; Garland STPM Press: New York, 1979; pp 241−303. (54) Al-Oweini, R.; El-Rassy, H. Synthesis and Characterization by FTIR Spectroscopy of Silica Aerogels Prepared Using Peveral Si(OR)4 and R″Si(OR′)3 Precursors. J. Mol. Struct. 2009, 919, 140−145. (55) Bertoluzza, A.; Fagnano, C.; Antonietta Morelli, M.; Gottardi, V.; Guglielmi, M. Raman and Infrared Spectra on Silica Gel Evolving Toward Glass. J. Non-Cryst. Solids 1982, 48, 117−128. (56) Brinker, C. J.; Haaland, D. M. Oxynitride Glass Formation from Gels. J. Am. Ceram. Soc. 1983, 66, 758−765. (57) Orcel, G.; Phalippou, J.; Hench, L. L. Structural Changes of Silica Xerogels During Low Temperature Dehydration. J. Non-Cryst. Solids 1986, 88, 114−130. (58) Anandha Babu, G.; Ramasamy, P. Growth and Characterization of 2-Amino-4-Picolinium Toluene Sulfonate Single Crystal. Spectrochim. Acta, Part A 2011, 82, 521−526. (59) Bordbar, A. K.; Rastegari, A. A.; Amiri, R.; Ranjbakhsh, E.; Abbasi, M.; Khosropour, A. R. Characterization of Modified Magnetite Nanoparticles for Albumin Immobilization. Biotechnol. Res. Int. 2014, 705068. (60) Gopal, N. O.; Narasimhulu, K. V.; Rao, J. L. EPR, Optical, Infrared and Raman Spectral Studies of Actinolite Mineral. Spectrochim. Acta, Part A 2004, 60, 2441−2448. (61) Shi, Y.; Zhang, H.; Yue, Z.; Zhang, Z.; Teng, K.-S.; Li, M.-J.; Yi, C.; Yang, M. Coupling Gold Nanoparticles to Silica Nanoparticles through Disulfide Bonds for glutathione detection. Nanotechnology 2013, 24, 375501. (62) Benetti, E. M.; Acikgoz, C.; Sui, X.; Vratzov, B.; Hempenius, M. A.; Huskens, J.; Vancso, G. J. Nanostructured Polymer Brushes by UVAssisted Imprint Lithography and Surface-Initiated Polymerization for Biological Functions. Adv. Funct. Mater. 2011, 21, 2088−2095. (63) Liu, Y.; Gan, L.; Carlsson, D. J.; Fagerholm, P.; Lagali, N.; Watsky, M. A.; Munger, R.; Hodge, W. G.; Priest, D.; Griffith, M. A Simple, Cross-linked Collagen Tissue Substitute for Corneal Implantation. Invest. Ophthalmol. Visual Sci. 2006, 47, 1869−1875. (64) Berlind, T.; Pribil, G. K.; Thompson, D.; Woollam, J. A.; Arwin, H. Effects of Ion Concentration on Refractive Indices of Fluids Measured by the Minimum Deviation Technique. Phys. Status Solidi C 2008, 5, 1249−1252. (65) Hand, D. B. The refractivity of protein solution. J. Biol. Chem. 1935, 108, 703−707. 31486


Research Article

ACS Applied Materials & Interfaces (66) Nguyen, T. T.; Bea, S. O.; Kim, D. M.; Yoon, W. J.; Park, J.-W.; An, S. S. A.; Ju, H. A Regenerative Label-Free Fiber Optic Sensor Using Surface Plasmon Resonance for Clinical Diagnosis of Fibrinogen. Int. J. Nanomed. 2015, 10, 155−163. (67) Lahiri, J.; Isaacs, L.; Grzybowski, B.; Carbeck, J. D.; Whitesides, G. M. Biospecific Binding of Carbonic Anhydrase to Mixed SAMs Presenting Benzenesulfonamide Ligands: A Model System for Studying Lateral Steric Effects. Langmuir 1999, 15, 7186−7198.

31487


Label-Free Optical Biochemical Sensors via Liquid-Cladding-Induced

Recommend Documents