1. Introduction

Hydrogen peroxide (H₂O₂), a non-radical reactive oxygen species (ROS) produced under physiological and pathophysiological conditions, can act as a key redox signaling molecule in diverse biological processes. Growing evidence suggests that abnormal perturbations in hydrogen peroxide levels are linked to a variety of human diseases including cardiovascular diseases [1], diabetes mellitus [2], gout [3], ulcerative colitis [4], sepsis [5], neurodegenerative diseases [6], and cancer [7]. In addition, H₂O₂ participates in a variety of cellular processes. Being diffusible, H₂O₂ molecules can have extended spatial and temporal distribution in the extracellular space. Relatively large amounts of H₂O₂ are present in blood plasma, body fluids, and even exhaled air. H₂O₂ is known to be involved in renal function adaptation, where it acts as an antibacterial agent in the urine [8]. Accurate, quantitative detection of H₂O₂ should help improve our understanding of how ROS impact human health and disease in processes as simple as diet, breathing, and excretion. To date, various analytical techniques have been developed for H₂O₂ detection or quantitation, which include fluorometry [9], spectrophotometry [10], and electrochemical methods [11,12]. While these approaches have their own benefits, their practical use are hampered by time or reagent costs, implementational complexity, and other technical demands on instrumentation and trained personnel.

Recently, fiber-optic based sensors have emerged as an attractive detection strategy, being recognized for their extremely sensitivity, and resistance to electromagnetic and chemical interferences. So far, this strategy has been explored in the design of physical, chemical and biological sensors [13]. As an operating principle, fiber-optic sensors leverage refractive index variations near the sensor interface, which can be monitored as a sensing transduction signal during optical detection. Fiber-optic sensors have been developed to detect ROS such as H₂O₂, with remarkable sensitivity [14,15]. Thus far, most fiber-optic H₂O₂ sensors operate on the basis of reactions of redox pair involving H₂O₂ [16–20] or of enzymatic catalysis [21,22]. While a particular redox and enzymatic reaction may confer specificity to a sensor for detecting H₂O₂, certain limitations remain, which include costly or time-consuming preparation, cumbersome operation, poor reproducibility, and low chemical and biological stability of active enzymes [23–25]. In particular, only a very limited number of fiber-optic sensors can be used for complex biological sample detection, traditional fiber-optic sensors are often susceptible to non-target interference during the detection process, the lack of sensing specificity limits its practical application. In recent years, nanomaterials and chemical techniques have attracted much attention in the field of optical sensing, which also inspired us to devise a simple yet effective way to prepare fiber-optic sensors that offer sufficient selectivity and sensitivity in practical applications.

Here, we report a microfiber mode interferometer by modifying non-enzymatic H₂O₂-sensitive materials, and its application to selective ROS sensing. The sensing device consists of a uniform waist region with an intermediate diameter of 8 µm and two tapering transition regions. Sensing material of the interferometer interface was designed to be switchable via hydrophobic-hydrophilic transformations on encountering H₂O₂, which in turn dictates changes in refractive index of interference instrument interfaces. The period of patterns depends strongly on the external H₂O₂-sensitive materials and is insensitive to non-target analytes. It is anticipated that the fiber-optic sensor reported here addresses some of the shortages of previously reported microfibers in terms of key properties such as selectivity, and thus opens up opportunities for biological and other applications requiring high-performance H₂O₂ sensors.

2. Experimental details

2.1 Materials and methods

2.1.1 Materials

We purchased sulfuric acid (H₂SO₄, 98%) and hydrogen peroxide (H₂O₂, 30%) from Guangzhou Chemical Reagent Factory (Guangzhou, China), anhydrous ethanol (EtOH) with concentration≧99.7% from Guangdong Guanghua Sci-Tech Co., L-cysteine (L-Cys), dichloromethane (DCM), potassium chloride (KCl), urea, L-Arginine (L-Arg), L-Alanine (L-Ala), L-Histidine (L-His), Glycine (Gly) from Meryer Chemical Technology Co. Ltd., sodium chloride (NaCl), methacryloyl chloride, ascorbic acid (AA), magnesium chloride hexahydrate (MgCl₂·6H₂O) from Aladdin, Inc., triethylamine from Macklin Biochemical Co. Ltd, lithium phenyl (2,4,6-trimethylbenzoyl) phosphinate (LAP), (4-(4,4,5,5-tetramethy l-1,3,2-dioxaborolan-2-yl)phenyl) methanol, glutathione (GSH) from Bide Pharmatech Ltd.. N-(2-aminoethyl)-2,2,4-trimethyl-1-aza-2-silacyclopentane from Gelest, Inc., Glucose from Xiya Chemical Technology Co. Ltd. (Shandong, China). 3-Hydroxytyramine hydrochloride (DA) from Shanghai Energy Chemical Co. LTD (Shanghai, China). All reagents were of analytical grade and were not subjected to further purification.

A solution of piranha was prepared by mixing sulfuric acid at 98% and hydrogen peroxide at 30% at a ratio of 7:3 by volume. Silane solution (2.4%, v: v) was made by adding 100 µL of the silane coupling agent to 4 mL of anhydrous dichloromethane and stirring evenly [26]. Cross-linking solution configuration was performed as follows: To 4 mg of the photo-initiator LAP, add 4 mL of ethanol to dissolve it; add 100 µL of phenylboronic ester monomer, followed by ultrasonic dispersion for 10 min.

2.1.2 General characterization and equipment

Wettability characterization experiments were performed by static contact angle measurements of surfaces with a contact angle meter (DSA-100, kruss, Germany) on flat glass, which was retouched and soaked in water or H₂O₂ solutions. Scanning electron microscopy (SEM) images of the microfibers were acquired by SEM (Thermo Scientific, FEI, Czechia) at 15 kV after 20 s of gold coating. Fourier infrared spectrum (FTIR) spectra were recorded with an FTIR spectrometer (Vertex 70, Bruker) for the samples in a wavelength range of 3500 to 400 cm^-1.

2.2 4-(4,4,5,5-Tetramethyl-1,3,2-dioxaborolan-2-yl) benzyl acrylate (phenylboronic ester monomer) synthesis

(4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl) phenyl) methanol (5 g, 21.25 mmol) was solved in anhydrous dichloromethane (30 mL), then triethylamine (5 g, 49 mmol, 6. 87 mL) was added, stirring was continued, and the ice water bath was cooled to 0 °C, methacryloyl chloride (5 g, 48 mmol, 2.7 mL) was solved in 3 ml of anhydrous dichloromethane, and diluted methacryloyl chloride was added dropwise to the solution at 0 °C. The ice-water bath was removed, and the reaction mixture was stirred at room temperature for 12 h. The anhydrous dichloromethane was then removed by rotary evaporation. The resultant solid-liquid mixture was redissolved in ethyl acetate, and then washed three times with brine. After drying over anhydrous sodium sulfate, the organic solvent was removed by evaporation. The compound was purified by silica gel column chromatography, by using ethyl acetate and petroleum ether (v/v = 1/30) as eluents to obtain the phenylboronic ester monomer in the form of a colorless oily liquid [27]. ¹H NMR (600 MHz, chloroform-d) δ 7.81 (d, J = 7.7 Hz, 2H), 7.37 (d, J = 7.7 Hz, 2H), 6.15 (s, 1H), 5.59 (s, 1H), 5.20 (s, 2H), 1.97 (s, 3H), 1.34 (s, 12H). ¹³C NMR (151 MHz, CDCl₃) δ 167.19, 139.11, 136.18, 134.99, 129.01, 127.15, 125.87, 83.86, 77.24, 77.02, 76.81, 66.30, 24.86, 18.35. HRMS (EI): m/z calcd for C₁₇H₂₀N⁺: 302.1689 [M]⁺, found: 302.1454. Detailed spectrum can be found in the supplemental document.

2.3 Microfabrication of fiber-optic sensor

In this experiment, a melt-drawn cone method was used to prepare the micro-nano fiber. A butane flame injector with a flame width of about 5 mm was used as the high-temperature heat source, and the outer flame of the flame was aimed at the part of the highly germanium-doped photosensitive fiber, where the coating layer had been removed, and the fiber was heated to a molten state. The fiber is kept in the molten state for better elongation, the displacement stage is controlled by the program to apply a uniform force along the fiber in two opposite directions in parallel, the flame and the displacement stage are controlled to stop working at the same time, resulting in a microfiber with a uniform diameter waist bundle in the middle and a tapered transition part on both sides. By observing the interferometric spectrum of the spectrometer and adjusting the movement time and speed of the displacement stage, the microfiber sensor required for high sensitivity and sensing was obtained. We prepared the optical fiber with a waist region diameter of 8 µm and a total length of approximately 10 mm for the tapered transition region and waist beam region.

The microfiber was soaked in piranha solution, removed and rinsed with water, ethanol and anhydrous dichloromethane. It was then immersed in silane solution for 2 h. The fibers were recovered and rinsed with ethanol, water, ethanol and anhydrous dichloromethane [26], 100 µl of methacryloyl chloride was dissolved in 2 ml of anhydrous dichloromethane, 40 µl of triethylamine was dissolved in 2 ml of dichloromethane, and the ice water bath was cooled to 0°C. The dilute triethylamine solution was added drop by drop to the methacryloyl chloride solution and stirred well. The fibers were recovered from anhydrous dichloromethane and immersed in the above solution, remove the ice water bath after 10 minutes, and the reaction mixture was stirred at room temperature for 20 min. The fiber was removed by washing with ethanol and water alternately to remove unreacted material from the fiber surface as well as to record the spectra. After that, the fiber was immersed in a cross-linking solution with UV lamp for 15 min. The cross-linked fiber was removed, placed in water for stabilization, it was then used for H₂O₂ detection.

2.4 Sensor setup

The optical setup of the hydrogen peroxide sensor is as follows. The tail fiber attached to the sensor was connected to a broadband source (BBS) at one end and to a spectrum analyzer (OSA) at the other end to record transmission spectra. Optical signals from the microfiber were excited by a broadband source (OS-EB-D-1250-1650, Golight). We chose a wavelength range of 1510 to 1610 nm. The transmission spectra were being monitored by an OSA (70D, Yokogawa), the signal distortion due to optical dispersion is minimal and the loss is lowest. The minimum wavelength resolution of the OSA was set to 0.02 nm, which was continuously measured and recorded at a rate of 1 spectrum each 10 seconds.

2.5 Stability tests of H₂O₂ sensors

The sensors were first immersed in deionized water and tested for stability by adding 200 µL of water dropwise. During the assay, H₂O₂ concentration was raised by adding 200 µL of different concentrations of H₂O₂ solutions once every 2 h. Transmission spectra of microfibers for different levels of H₂O₂ were recorded. The method for detecting H₂O₂ in urine was similar. All detections were performed at room temperature.

2.6 Tests on selectivity

We tested selectivity of the sensor for H₂O₂ against other analytes such as metal ions (Na⁺, K⁺, Mg²⁺), L-cysteine (L-Cys), urea, glutathione (GSH), glucose (Glu), ascorbic acid (AA), L-Arginine (L-Arg), L-Alanine (L-Ala), L-Histidine (L-His) and Glycine (Gly). The sensor piece was immersed in water (2 mL) and 200 µL of a solution containing NaCl, KCl, MgCl₂·6H₂O and L-cysteine (L-Cys), urea, glutathione (GSH), glucose (Glu), ascorbic acid (AA), L-Arginine (L-Arg), L-Alanine (L-Ala), L-Histidine (L-His), Glycine (Gly) at a concentration of 1 µM each. The interfering species mentioned was added dropwise. Transmission spectra of microfibers at different levels of H₂O₂ were recorded. All the detections were performed at room temperature.

3. Results and discussion

3.1 Sensor design

Here, a microfiber has been reported for use as an optical sensing structure, where the tapered fiber consists of a waist region with a uniform diameter in the middle and a reduced diameter region (transition region) on both sides, As show in Fig. 1(a). When the tapered region of the fiber is stretched thin, the waist diameter of the fiber is thinner than the wavelength of the guided light. The fiber transforms into a micron diameter waveguide coupled with an air cladding. As a result, part of the light is transmitted along the core and part of the light propagates swiftly into the air cladding, resulting in strong and fast near-field interactions between the air cladding environment and the guided light. This provides a higher sensitivity than the untreated Mach-Zehnder interferometer (MZI) [28].

 

Fig. 1. Representative scheme illustrating the preparation of H₂O₂-responsive optical micro-fiber sensor based on a phenylboronic ester derivative. (a) Schematic diagram of the sensor fabrication. (b) Chemical structure of phenylboronic ester polymer, and formation of hydrolysis products by H₂O₂. (c) Illustration for switching of the chemical structure and wettability properties under oxidative conditions.

Download Full Size | PDF

The sensor design is based on microfiber mode interferometer, using the modified cladding technique by chemical material as shown in Fig. 1. Specifically, the phenylboronic ester monomer is introduced in the process of cladding functionalization because there are methyl methacrylate groups in the surface of fiber, and the free radical polymerization can be achieved (Fig. 1(a)). The detection behavior of the sensor is derived from the switching of the chemical structure and wettability of the sensing material under H₂O₂ conditions (Fig. 1(b)). In addition, according to our previous report [29], he sensing device does not satisfy the condition of total internal reflection of light and has a strong evanescent field in the sensing region. However, the water-wet cladding boundary supports internal reflection conditions, which can reflect the leaked light back to the core of the optical fiber, leading to transmission through the modification region. Any variation in the refractive index of the sensing region (defined as sensing interface refractive index, SRI) caused by the analyte can change the refractive condition, and resulting in a different interference spectrum [30]. Here, once the sensor is exposed to H₂O₂, the wettability of sensing layer is converted from hydrophobic to hydrophilic (Fig. 1(c)). The rapid and stable chemical interaction between the material and water ensures excellent sensing performance of the sensing device.

3.2 Characterization of sensing material performance

This implemented approach was first evidenced on a model substrate, i.e. on a pretreated flat glass surface in wettability measurement test. Static contact angle tests of surfaces were performed on glass plates without the modified borate ester material and after the modified material being soaked in water and H₂O₂ solution for a period of time, to qualitatively show the reaction of the induction layer on the surface. The contact angle values of 20.9° were measured on unmodified glass plates and 57.4° after soaking the glass plate in water after modifying the material, which may be related to the hydrophobicity of these phenylboronic ester (Fig. 2(a)). Interestingly, the contact angle value after rinsing with H₂O₂ solution at a concentration of 300 g/L reached 51.6° (Fig. 2(b)), which is consistent with the assumption that hydrolysis of the borates occurred under oxidizing conditions. These results clearly demonstrate the characteristic behavior of the developed system in response to H₂O₂, whereby it becomes hydrophilic while the hydrophobic angle decreases on encountering H₂O₂.

 

Fig. 2. Performance characterization plots of sensing materials (a) Static contact angle measurements of glass plates, for blank glass plates, after soaking with water and after soaking with H₂O₂ after modification of sensing materials, respectively. Values are mean ± SD (n = 3). (b) Representative plots of static contact angle measurements for blank glass plate, modified sensing material after immersion in water and H₂O₂, respectively. (c) FTIR of the material for phenylboronic ester monomer, after cross-linking the material soaked in water, and H₂O₂. (d) Plots of the wave number intervals where the intercepted characteristic peaks are present. All the above steps were performed at room temperature.

Download Full Size | PDF

Hydrolysis process of the phenylboronic ester group can be visualized in the FTIR, as shown in Fig. 2(c) and Fig. 2(d). It can be seen that both monomer and polymer. FTIR spectra show strong band peaks caused by stretching vibrations of the benzene ring and B-O group at 860 and 1320 cm^-1, respectively, and band peaks of 1360 and 2980 cm^-1 for methyl, and two strong band peaks of 1089 and 1143 cm^-1 for carboxylic acid ester. It is evident that the material polymerization did not show any explicit changes in the H₂O₂ sensing material, which indicates the presence of sensing groups. After H₂O₂ immersion in the cross-linked material at a concentration of 300 g/L, all of the above absorption peaks either became weakened or disappeared, which further suggests that the sensing performance is based on changes in chemical structure. These results support feasibility of preparing H₂O₂ microfiber sensors in the next step.

3.3 Surface morphological characterization

For realizing in situ polymerization of phenylboronic ester on optical fiber, initial surface modification of optical fiber is a crucial step. After cleaning and silane modification in an organic solvent, the single molecular layer functionally surface appears smooth and resembles an unmodified fiber (Fig. 3(a)). After modification by in situ polymerization, the polymer gradually aggregated and attached to the surface of fibers as its solubility dropped in the solvent. As shown in Fig. 3(b) the fiber surface is covered with a large number of submicron globules-like structures that are uniform in size with diameters of about 400-500 nm. Hydrophobic globules and non-compactly arranged structures are the key to the hydrophobicity of the fiber surface. The appropriate amount of modified polymer can be monitored by a spectrometer, in order to avoid excessive polymerization causing a weakening of the interference spectrum.

 

Fig. 3. SEM image of the fiber surface. (a) Unmodified fiber surface. (b) After in-situ polymerization of the sensing material on the fiber.

Download Full Size | PDF

3.4 Stability and sensitivity of the phenylboronic ester modified micro-optical sensor

The transmission spectrum of a fabricated taper was measured by means of an optical spectrum analyzer (OSA) as shown in Fig. 4(a). Periodic fringes can be observed as a unique result of modal interference. Free spectral range (FSR) was about 25 nm and the extinction ratios were about 25 dB. After modifying the sensing material, the extinction ratio becomes 5 dB, as shown in the red interference fringe. Importantly, as anticipated, optical microfiber was remarkably sensitive to refractive index (RI) of the surrounding environment. As shown in Fig. 4(c), the measured shifts of selected dip(s) act as a function of the external RI. All the dips were red shift with increasing RI By fitting linearly to an external RI range from 1.3333 to 1.3374, the fiber device has a high measurement sensitivity up to 2352.376 nm/RIU (refractive index units). The subsequent detection was performed under constant-temperature conditions (25 °C) to avoid thermal expansion and thermal effects of the sensing materials. The device was almost insensitive to ambient temperature. In addition, transmission notches (∼1561.05 nm) of the modified fiber in water microfiber sensor was found to have negligible variations for a period of over 2 h (Fig. 4(b)), corroborating its high stability. This high stability is derived from the stable structure of the fiber optic device itself. However, detection accuracy and precision are still related to the repeatability of the device preparation, so we show the repeatability results for multiple fibers in SI figure S4. In summary, the observed stability and sensitivity of the sensor confer confidence on its potential use in H₂O₂ sensing.

 

Fig. 4. Characteristics of the sensor piece. (a) Comparison of interference spectra before and after modification. The black line indicates the interference spectra of bare fibers, and red line shows the interference spectra after phenylboronic ester modified. (b) Stability of the modified sensor device. The right drift of this peak at 1,561.05 nm was observed at 2 h. (c) Refractive index sensitivity measurement of the bare fibers. All the above steps were carried out at room temperature.

Download Full Size | PDF

3.5 Sensing performance of phenylboronic ester modified sensor in H₂O₂ detection

For characterization of the H₂O₂ sensor, the experimental setup used was the same as described above. The first H₂O₂ sample (0 g/L H₂O₂) was kept in the sample tube. Its interference spectra were recorded as a baseline (Fig. 5(a)). Generally, the signal became stable by 2 h, which may be taken as response time of the sensor (Fig. 5(b)). The H₂O₂ treated sensor underwent weak changes in surface morphology (Fig. 5(c), Fig. 5(d)). Although sensing materials underwent transformation from being hydrophobic to hydrophilic, they were still immobilized to the modified site. Water molecules were allowed to approach the cladding of the fibers.

 

Fig. 5. Characterization of H₂O₂ sensor. (a) Comparison of the spectra after adding H₂O₂ at a concentration of 0 g/L and 10⁻⁶ g/L for 2 h. (b) Comparison of the shift of the spectrum with time after adding 0 g/L of H₂O₂ and 10⁻⁶ g/L of H₂O₂ for 2 h. (c) Change of the surface morphology of the optical fiber after soaking with water (SEM). (d) Morphological changes of the fiber surface after soaking with hydrogen peroxide (SEM).

Download Full Size | PDF

Coherence spectra recorded for interference spectra from 0.9×10^−15–0.59×10⁻⁴ g/L of H₂O₂ are as shown in Fig. 6(a), where Fig. 6(b) also shows the right shift of the spectra over time for different concentrations. Based on the coherence spectra, an interference wavelength redshift can be noted with increasing concentrations of H₂O₂ (Fig. 6(b)). As described above, the reason for this redshift in interference wavelength is an increase in the effective RI of the sensing layer due to decomposition of sensing materials in the presence of H₂O₂ and the interaction of released water molecules with sensing surface. As the concentration of H₂O₂ increased its decomposition as the chemical reaction progressed, interaction of water molecules with sensing interface increases and hence the effective RI of the sensing layer increased. Augmentation in RI increased the resonance wavelength and hence the shift.

 

Fig. 6. Sensing performance of sensor in H₂O₂ detection. (a) Interference spectra recorded at different H₂O₂ concentrations, where the H₂O₂ concentrations were 0, 0.9×10⁻¹⁵, 0.8×10⁻¹³, 0.78×10⁻¹¹, 0.72×10⁻⁹, 0.67×10⁻⁷, 0.59×10⁻⁵ g/L, respectively; (b) Red-shift of the spectra with time for different H₂O₂ concentrations, where the H₂O₂ concentrations were 0, 0.9×10⁻¹⁵, 0.8×10⁻¹³, 0.78×10⁻¹¹, 0.72×10⁻⁹, 0.67×10⁻⁷, 0.59×10⁻⁵ g/L, respectively; (c) Linear fit of the sensor response to H2O2. Bar graphs of the response error for the detection of different H₂O₂ concentrations in multiple sensors. Values are mean ± SD (n = 3).

Download Full Size | PDF

Redshift of coherence wavelength with the concentrations of H₂O₂ (Fig. 6(c), Figure S5). A shift of around 1.4 nm in coherence wavelength was observed, when the H₂O₂ concentration changed from 0.9×10⁻¹⁴ g/L to 0.59×10⁻⁴ g/L. Sensitivity of the sensor defined as the redshift in the coherence wavelength per unit change in the concentration of H₂O₂ [31]. Based on the calibration curve of the sensor and the recommendations of IUPAC, we obtained the low limit of detection (LOD) from the calibration curve of the sensor based on the standard deviation of three blank measurements (3σ_blank) [32], which was found to be around 4.92×10⁻¹² g/L, in a trace concentration range. This extremely LOD is suitable for trace concentration detection and can improve operational viability for micro-samples that require dilution.

3.6 Selectivity and practical evaluation of sensor

Furthermore, selectivity of sensor for H₂O₂ was tested against other analytes that are biologically relevant [33–35], including metal ions (Na⁺, K⁺, Mg²⁺) and L-cysteine (L-Cys), urea, glutathione (GSH), glucose (Glu), ascorbic acid (AA), L-arginine (L-Arg), L-alanine (L-Ala), L-histidine (L-His) and Glycine (Gly). The concentrations of the above non-target analytes were 1 µM. As shown in Fig. 7, only the H₂O₂ group (1.9×10⁻³ µM (0.67×10⁻⁷ g/L), much lower than physiological concentration [8].) showed an obvious wavelength shift, while the influence of other analytes on wavelength of the sensor was negligible. The results show that the sensor has excellent selectivity to H₂O₂ over other interfering species, thus allowing accurate detection of H₂O₂ in complex biological environments. In comparison to existing H₂O₂ fiber optic sensors, this chemical material modified fiber optic sensor provides high specificity, see SI Table S1.

 

Fig. 7. The spectral response of H₂O₂ sensing and bare fiber, corresponding to H₂O₂ (1.9×10⁻³ µM) and Na⁺, K⁺, Mg⁺², L-cysteine (L-Cys), urea (Urea), glutathione (GSH), glucose (Glu) ascorbic acid (AA), L-arginine (L-Arg), L-alanine (L-Ala), L-histidine (L-His), glycine (Gly). The concentrations of non-target analytes were all 1 µM. Values are mean ± SD (n = 3).

Download Full Size | PDF

4. Conclusion

In this work, we demonstrate that fiber-optics device for polymer surface modification can be facilely fabricated, whose optical properties change dynamically in response to H₂O₂. These were subsequently applied to quantify H₂O₂ in solutions. The proposed sensor has many advantages over traditional optical sensors, which include high selectivity and sensitivity, stable response, broad detection range, ease of implementation, low costs, and real-time monitoring capability. The H₂O₂ microfiber sensor has a linear dynamic concentration range of 0.9×10⁻¹⁴ g/L to 0.59×10⁻⁴ g/L, and the LOD reached 4.92 × 10⁻¹² g/L. In addition, it operates on the basis of mode interference of wavelength; neither changes in light source intensity nor changes in ambient light affect measurements. The proposed sensor also demonstrated excellent selectivity for H₂O₂. This design principle for microfiber optic sensors based on hydrophobic-hydrophilic conversion is expected to be useful as a general sensing platform for monitoring of various biomarkers.