Open Access
Add to BibSonomy
Abstract
External temperature variations inevitably affect the accuracy of surface plasmon resonance (SPR) biosensors. To that end, we propose an ultra-compact label-free dual-channel SPR fiber sensor (DSPRFS) that can simultaneously measure the glucose concentration and ambient temperature in real-time. The proposed sensor is based on a unique dual-channel structure fabricated by etching a side-hole fiber (SHF), and has significantly higher spatial sensitivity than traditional SPR biosensors. After coating with silver and zinc oxide films, one channel was filled with polydimethylsiloxane (PDMS) to sense the ambient temperature, and the other channel was immobilized with glucose oxidase (GOx) enzyme for glucose sensing. The proposed sensor is analyzed theoretically, fabricated and characterized. Glucose concentration sensitivity and temperature sensitivity of the manufactured sensor sample were found to be as high as 6.156 nm/mMand -1.604 nm/°C with limits of detection (LOD) of 16.24 µM and 0.06 °C, respectively. The proposed sensor has an extremely compact structure, enables temperature compensation, and is suitable for in-situ monitoring and high-precision sensing of glucose and other biological analytes.
© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Numerous studies indicate that glucose concentration measurements are not only important for the diagnosis and treatment of diabetes [1], but also for early diagnosis of cancer [2] and schizophrenia [3], and, as a consequence, reliable glucose sensors are of importance in life science research and are increasingly used in the healthcare industry. Typically, blood or urine samples are used for glucose analysis, however these contain a lot of impurities. Alternatively, tears, which have a comparable glucose concentration to that in blood or urine, can be used for disease monitoring. As an example, a measured tear glucose concentration above 0.92 $mM$ indicate that the patient suffers from diabetes [4]. Tear analysis also has advantages over traditional methods in that it is non-invasive, reducing the risks of infection and patient discomfort in the case of blood sample collection. To date, various techniques based on glucose oxidoreductase (GOx) have been developed to detect glucose, including fluorescence-based sensors [5], electrochemical sensors [6,7], colorimetric methods [8], and optical sensors [9]. Fiber-based optical sensors have many advantages such as real-time detection, low cost, ability for remote sensing, small size, minimized sample quantity, and full biocompatibility [10,11]. A surface plasmon resonance (SPR) sensor is based on surface plasmon polaritons (SPPs) excited at the interface between a metal and a dielectric material when a wave vector matching condition is satisfied [12] and it can accurately monitor the change of the surrounding medium’s refractive index (RI). SPR fiber sensors combine the advantages of fiber optic sensing and SPR sensing, and have been widely used for glucose sensing and label-free biological sensing of numerous complex analytes.
Nano-porous and nanostructure ZnO films are used in biosensor applications due to advantages of biochemical stability, biocompatibility, excellent adsorption capacity and non-toxicity [13,14]. ZnO has a high isoelectric point (IEP) and large surface-to-volume ratio, which enables various low IEP biomolecules, such as enzymes and receptor proteins, to firmly couple to its surface through electrostatic interaction [15,16]. Furthermore, ZnO nanostructures can easily be manufactured in large quantities, making ZnO suitable as a biosensing platform.
The main purpose of a biosensor is to detect target biological molecules fast and accurately with high sensitivity and selectivity. For a GOx-based biosensor, the immobilization efficiency of biological enzymes to the sensor strongly affect the sensing performance. Commonly used immobilization methods include self-assembly [17], sol-gel [18], cross-linking [19,20], covalent binding [21] and physical adsorption [22,23]. The IEP of GOx (4.2) is lower than that of ZnO (9.5), which enables GOx to be directly immobilized on ZnO through electrostatic adsorption [24,25]. Coating of a ZnO film on top of a metal film also have several advantages; the ZnO film can prevent the metal film (e.g. Ag, Cu) from oxidizing; and, in the case of an SPR sensor, the high RI of ZnO can improve the sensor sensitivity as well as adjust the SPR resonance wavelength [26–29]. I.e., use of a ZnO film is overall beneficial for the performance of an SPR.
The optimal temperature window for the reaction between GOx and glucose molecules is 40-50 $^\circ C$, outside which the enzyme reaction rate will be reduced, and there is also a risk for GOx to permanently lose its biological activity [30]. As the biological recognition element of a glucose biosensor, ensuring maximum biological activity of GOx will greatly improve the sensitivity and accuracy of the sensor. Therefore, the ambient temperature is one of the most important factors affecting the performance of a glucose biosensor. However, in general, GOx based biosensors rarely considers the impact of ambient temperature, alternatively includes a cascaded external temperature sensor [31–34]. An external temperature sensor not only increases the complexity of the sensing system but also cannot measure the actual temperature at which GOx reacts with glucose molecules due to the physical separation between the temperature sensor and the biosensor.
In this paper, we propose an in-situ ultra-compact and highly sensitive dual-channel SPR fiber sensor (DSPRFS) that can simultaneously measure glucose concentration and the ambient temperature. We theoretically and experimentally evaluate optical and temperature sensing properties as well as the glucose concentration sensing property of the proposed sensor. The experimental results demonstrate that the optical and glucose concentration sensitivities of the biosensing channel are 2015.5 $nm/RIU$ and 6.156 $nm/mM$, respectively. The sensitivity of the temperature sensing channel is -1.604 $nm/^\circ C$. As temperature variations have a significant effect on the performance of the biosensing channel, we derive a calibrated sensing matrix that enables us to ensure maximum activity of GOx through real-time monitoring of ambient temperature variations as well as temperature compensation during the glucose concentration detection. Therefore, the sensitivity, accuracy and credibility of the glucose concentration measurement results are greatly improved as compared to previously reported sensors [7,8,10,35].
To compare various representative glucose concentration and temperature sensors with different optical configurations previously reported are listed in Table 1 and 2, respectively. As seen in Table 1, the linear range of the DSPRFS is larger than that of the other glucose sensor structures and its limit of detection (LOD) lower compared to all except the SPR fiber sensor of Ref. [10], which enhances the sensitivity by nearly 4 times compared to SPR fiber sensors without metal nanoparticles. However, metal nanoparticles increase both the manufacturing cost and the sensor fabrication complexity. The dynamic range of the representative temperature sensors listed in Table 2 are larger than that of the DSPRFS, however their sensitivities are smaller. The SLI sensor is the exception [37], however this is not suitable for compact biological sensing due its length of 22 $cm$.
Table 1. Performance comparison with previously reported glucose concentration sensors
Table 2. Performance comparison with previously reported temperature sensors
2. Sensor design
2.1 Operating principle
A schematic diagram of the DSPRFS is shown in Fig. 1(a). The sensor structure consists of two independent semi-open channels, which are formed by chemical etching of a side-hole fiber (SHF). As the remaining cladding outside the core is sufficiently thin, evanescent field energy of the light wave will leak out and interact with the external medium. Both channels are coated with a thin metal film (Fig. 1(b)). There exist a large number of free electrons in the metal and propagating charge waves localized at the metal-medium interface can interact with incident light [12,39]. The TM mode has a component in the direction perpendicular to the interface that can excite the SPR; while the TE mode cannot as the vibration direction is parallel to the interface. The propagation constant $\beta$ is:
where ${k_0} = 2\pi /\lambda$ is the free-space wavenumber, and $\lambda$ is the free-space wavelength [11,39]. ${\varepsilon _m}$ and ${\varepsilon _d}$ are the permittivities of the metal and the dielectric, respectively. Assuming the SPPs propagate along the x-axis, then the wave vector component of the evanescent field along the x-axis is: where ${n_p}$ is the RI of the SHF and $\theta$ is the angle of the incident light [12,39]. The coupling condition for SPR excitation is:
Fig. 1. (a) Schematic diagram of the dual-channel SPR fiber sensor and (b) cross-sectional view of etched side-hole fiber sensing area.
We immobilize GOx on one of the two semi-open channels (channel A in the following), which will act as the biosensor for glucose molecules. The other channel (channel B in the following), is filled with PDMS, thereby waterproofing it so that contact with the solution to be measured is avoided and allowing it to act as a temperature sensor. The SRI of channels A and B are approximately 1.3340 and 1.4145, respectively. A cross-sectional view of etched SHF sensing area is shown in Fig. 1(b). Using this structure, glucose concentration and ambient temperature can be measured simultaneously by exciting separate SPRs at different wavelengths in the two channels.
Glucose molecules flowing in channel A will chemically react with the GOx immobilized on the channel surface. GOx will catalyze β-D-glucose in the presence of water and oxygen to produce gluconic acid and hydrogen peroxide [41]:
The RI of gluconic acid (∼1.4161) and hydrogen peroxide (∼1.406) are higher than those of β-D-glucose (∼1.34), GOx (∼1.35), and phosphate buffer saline (PBS, ∼1.3340). Therefore, the SRI of channel A will increase, causing the wavelength of the SPR resonance dip to redshift. PDMS, which is filled into channel B, is a new type of non-toxic, hydrophobic and biocompatible organic polymer material that has a high negative thermo-optic coefficient (${\Delta }{n_{PDMS}}/{\Delta }T = - 4.2 \times {10^{ - 4}}/^\circ C$) [42] and the RI, measured by an Abbe refractometer (2WA-J), is 1.4145 at room temperature. When the ambient temperature increases, the RI of the PDMS will decrease, which will cause a blueshift of the SPR dip in channel B.
2.2 Simulations of the DSPRFS performance
The performance of the SPR sensor is related to the material and thickness of the metal layer that is coated on the channel walls. Performance optimization simulations have previously been reported for metals commonly used in SPR sensors such as Au, Ag, and Al at different thicknesses [43,44]. Ag has the highest sensitivity and has a suitably low price, e.g. being cheaper than Au. However, Ag is generally not used in SPR biosensors as it oxidizes easily and has poor biocompatibility. Our results showed a 45 nm Ag film to be the optimal solution, and to protect the Ag film against oxidation a 5 nm ZnO thin film is subsequently sputtered on. Three- and two-dimensional simulation results of hybrid plasmon modes excited in channels A and B at the respective SPR resonance wavelengths are shown in Fig. 2. In these simulations, the RIs of the core and cladding of the fiber, as well as water and PDMS are ${n_{core}} = 1.4500$, ${n_{cladding}} = 1.4450$, ${n_{water}} = 1.3330$ and ${n_{PDMS}} = 1.4145$. It is worth noting that only TM polarized light can excite the SPR, while TE polarized light shown in Fig. 2(e) cannot. At ambient temperature 20 °C, a NaCl solution with RI in the range from 1.3330 to 1.3612 is injected into the flow cell, and two SPR resonance dips appear in the transmission spectrum. The central wavelength of the SPR resonance dip excited in channel B is 737 nm. The SRI change does not impact on the plasmon condition in channel B due to the PDMS isolation. The influence of the SRI change on the transmission spectrum of the DSPRFS is shown in Fig. 2(f). When DI water (RI=1.3330) is injected into the flow cell, the SPR of channels A and B are also excited at the same time. Since the TOC of DI water and PDMS are both negative, the central wavelengths and of the SPR resonance dips are both blue shifted due to temperature increases. However, the shift of, which characterizes the temperature, is approximately ten times that of, which characterizes the SRI. The influence of ambient temperature change on the transmission spectrum of the DSPRFS is shown in Fig. 2(g). The simulation results thus show that the DSPRFS can simultaneously measure the SRI and the temperature of the surrounding environment, which enables real-time simultaneous biosensing and ambient temperature disturbance compensation.
Fig. 2. Three-dimensional simulation of hybrid plasmon modes excited in the dual-channel SPR fiber sensor: (a) channel A and (b) channel B. Two-dimensional simulation of the fundamental TM polarization mode in channel (c) A and channel B (d) as well as the fundamental TE polarization mode (e). Simulated transmission spectra of the sensor at varying RI (f) and temperature (g), with the remaining cladding thickness in the waist region of SHF is 0 µm. (h) Simulated transmission spectra of the sensor at varying remaining cladding thickness in the waist region, with ambient temperature at 20°C and a solution RI of 1.3330. (i) Glucose concentration sensitivities and temperature sensitivities at varying remaining cladding thickness in the waist region.
It is relevant to investigate the effect of variations of the remaining cladding thickness in the waist region after etching. Figure 2(h) shows simulated transmission spectra at varying remaining cladding thickness, with ambient temperature at 20 °C, and a solution RI of 1.3330 and Fig. 2(i) shows the dependence of glucose concentration and temperature sensitivities of the same. As the thickness of the remaining cladding increases, the depth of the SPR resonance dips increases. More importantly however, the glucose concentration sensitivities and temperature sensitivities of the sensor are largely unaffected by variations in the thickness of the remaining cladding.
3. Sensor fabrication
3.1 Materials and reagents
The DSPRFS was fabricated from an SHF manufactured by the Yangtze Optical Electronics Co., Ltd, China. Ag and ZnO targets used for magnetron sputtering were purchased from ZhongNuo Advanced Material Co., Ltd, China; polylactic acid (PLA) material for 3D printing from Shining3d Co., Ltd, China; and PDMS (Dow Corning Sylgard 184) from Dow Corning, USA. Hydrofluoric (HF) acid (≥ 40.0%), acetone, isopropanol, KCl, NaCl, NaH2PO4·2H2O and NaHPO4·12H2O were all procured from Sinopharm Chemical Reagent Co., Ltd, China. Deionized (DI) water is prepared in the laboratory. β-D-Glucose anhydrous, high purity (> 180 U/mg) GOx (from Aspergillus niger) freeze-dried powder was purchased from Aladdin Industrial Co., China.
3.2 Fabrication method
The fabrication process steps are shown in Fig. 3(a). First, using an optical fiber fusion splicer (S178A, FITEL, Co., Japan), both ends of the SHF that is about 1 cm long are spliced with a single-mode fiber (SMF) for subsequent experimental measurements. In the second step, the SHF is etched using HF acid diluted to 24% for a total of 45 minutes, creating a double semi-open channel structure. Immediately following the etching, the SHF is washed using a large amount of DI water to clean residual HF acid from the surface. Acetone and isopropanol are then used for further cleaning. In the third step, the etched fiber is fixed to a support using UV glue (A332, Ausbond, Co., USA). This serves two purposes: it reduces disturbance of the sensor due to external stress during measurements; and it ensures the subsequent fabrication steps can be done without risk of damage to the sensor. The fixed fiber-support structure is then placed in a base for sputtering. The support and the base used for fixing the sensor, shown in the Fig. 3(b), were designed by SolidWorks software and fabricated by a commercial 3D printer. In the fourth step, a 45 nm Ag film and a 5 nm ZnO film are sequentially sputtered on the surfaces of the two semi-open channels using a magnetron sputter (PRO Line PVD 75, Kurt J. Lesker, Co., USA). In order to enhance the adhesion of the Ag film to the silicon dioxide, the fiber surface is treated using a plasma cleaner (PDC-002, HARRICK PLASMA, Co., USA) for 5 minutes prior to sputtering. In the fifth step, channel B is filled with PDMS using a fiber tip with the help of a microscope. PDMS will quickly fill the entire semi-open channel as a result of capillary action. Due to the symmetrically opposing arc structures of the etched SHF and the fact that PDMS is a viscous liquid prior to curing, PDMS will not flow into the opposite semi-open channel, which is the key to fabricate the DSPRFS such that it can simultaneously measure glucose concentrations and ambient temperature. Filling the semi-open channel with PDMS is easily done, and due to its excellent elasticity, it also improves the mechanical strength of the sensor. The fiber is then dried in an oven at 60 °C for 3 hours to cure the PDMS. In the final step, GOx is immobilized on the ZnO film in channel A. 0.01 PBS is prepared with KCl, NaCl, NaH2PO4·2H2O and NaHPO4·12H2O [45,46]. When the concentration of GOx in the PBS solution is 4 mg/ml, pH value is 7.4, and the soaking time is 7 minutes, GOx has the best fixation effect on ZnO. Subsequently, GOx is electrostatically immobilized by drying in the air at room temperature for 30 minutes. Following the completed fabrication, the sensor is stored at 4 °C for glucose concentration measurement experiments. The fabrication process of the DSPRFS is standardized and repeatable. All fabrication parameters (i.e. etching time, etching temperature, sputtering time, chemical reaction condition et al.) can be controlled precisely. What’s more, in our design, a pair of semi-open channels with special structure can be obtained only by etching, which takes the advantages of simple operation, short time consumption, low cost, and does not require expensive equipment such as focused ion beam or side-polished machine. Subsequent sputtering, PDMS filling and immobilizing glucose oxidase (GOx) enzyme are also commonly used optical fiber SPR sensor fabrication processes. In addition, in large part due to the fiber-support and base structure, so that batches of fiber samples can be simultaneously etched and sputtered under the same operating parameters with high efficiency. The average losses of the SHF after being spliced with SMF and after being etched are 0.22 dB and 5.4 dB, respectively. The average insertion loss of the DSPRFS is 21.55 dB. All losses are measured using a laser source with a fixed wavelength of 1550 nm.
Optical microscopy and scanning electron microscopy (SEM) are used to confirm the structure and dimensions of the sensor, the results of which are shown in Fig. 4. Figures 4(a) and (b) show cross-sectional SEM images of the SHF before and after the HF hydrofluoric acid etching. The SHF contains two air holes with diameters of 43 µm, symmetrically located on both sides of the fiber core with a diameter of about 9 µm [47]. Following etching, two independent semi-open channels remain. The diameter and the thinnest waist of the central etched SHF are approximate 98 µm and 9 µm, respectively. Figures 4(c) and (d) show optical micrograph and SEM side views of the finished DSPRFS. The bright area seen in Fig. 4(c) results from scattering of guiding light in the core, and the granular matter shown seen in Fig. 4(d) is residue from the fabrication process.
Fig. 4. Cross-sectional SEM images of the SHF before (a) and after (b) HF acid etching. Optical micrograph (c) and SEM image (d) of the side view of the DSPRFS. (e) Cross-sectional optical micrograph of the dual-channel SPR fiber sensor. SEM images of the surface of channel A before (f) and after (g) immobilizing GOx.
Figure 4(e) shows a cross-sectional optical micrograph of the finished DSPRFS. The semi-open channel on the left is channel B filled with cured PDMS, and the semi-open channel on the right is channel A immobilized with GOx. Figure 4(f) and 4(g) show SEM images of the surface of channel A before and after immobilizing GOx, respectively. The surface of ZnO shown Fig. 4(f) can be seen to be smooth with excellent film-quality. After soaking in the GOx solution, granular substances with a diameter of about 120 nm appears on the surface of the ZnO (Fig. 4(g)), indicating that GOx is successfully immobilized. Similar to Liu’s work described in [48], in order to confirm the thickness of the Ag and ZnO film deposited onto the fiber, the thickness of a film deposited on a glass sheet placed next to the fiber on the base is measured by a step profiler (Dektak 150, Veeco Co., USA) after each sputtering step. Figure 5(a) shows that the film-thickness is 45.07 nm after sputtering Ag film and Fig. 5(b) shows the total film thickness to be 50.23 nm after ZnO film sputtering step, and the thickness of ZnO film can then be calculated to be 5.16 nm. Thus, we find the actual film-thickness obtained by sputtering is as expected.
Fig. 5. Thickness of the deposited film measured by a step profiler: (a) Ag film and (b) Ag film with a ZnO film on top.
4. Experiments
4.1 Experimental setup
The experimental setup used is shown in Fig. 6. The DSPRFS is placed inside a chamber made of PDMS that includes a flow cell enables full contact between the sensor and the sample to be measured. Waste liquid exits from the outlet to protect the sample from pollution. The chamber is placed on a thermo electric cooler (TEC), through which the temperature in the flow cell can be adjusted so as to measure the sensor temperature response to an accuracy of ±1°C. A high-power tungsten halogen lamp (HL 2000) is used as a light source. After propagation through the sensor, the light carrying the sensing information is fed into a spectrometer (Aurora 4000), which is connected to a computer that stores and displays the measured spectral information.
4.2 Optical and temperature sensing characteristics
Prior to being able to accurately measure glucose concentrations it is necessary to derive the optical and temperature sensing characteristics. Therefore, before immobilizing GOx in channel A, we evaluate the optical and temperature sensing characteristics of the DSPRFS. Characterizations were done using the experimental setup described in section 4.1. Keeping the ambient temperature at 20 °C, NaCl solutions with RIs of 1.3330, 1.3400, 1.3505, and 1.3612 are injected into the flow cell in turn. This RI range is representative for the most common biofluids. The measured transmission spectra, shown in Fig. 7(a), are consistent with the simulation results. For the temperature dependence, DI water is injected into the flow cell and the temperature is set by the TEC module to be 20 °C, 30 °C, 40 °C, and 50 °C in sequential order. The resulting experimental transmission spectra, shown in Fig. 7(b), also agree with the simulation results.
Fig. 7. Experimental transmission spectrum of the DSPRFS with surrounding (a) refractive index change and (b) temperature change. Experimental data of dual-channel SPR wavelength shift with surrounding (c) refractive index change and (d) temperature change.
Table 3 compare glucose concentration sensitivities and the temperature sensitivities calculated by the simulation and experimental results. We find the full width at half maximum (FWHM) of the SPR resonance dips measured in the experiments to be larger than in the simulations, which we attribute to fabrication irregularities and an uneven power distribution of the white light source. Another reason may be that both skew rays and guided rays propagate in the optical fiber. Due to the different incident angles of each ray, the penetration depth of the evanescent field varies, and, in addition, the optical power distribution of each ray is different. Multiple reflections of the transmitted beam within the sensor may also cause broadening of the spectrum.
Table 3. Sensitivities of the DSPRFS calculated by simulation and experimental results
Linear fits of the SPR resonance wavelengths of the two channels and, with respect to the SRI and temperature are shown in Fig. 7(c) and (d). We can calculate that the RI sensitivity of the DSPRFS channel A is 2015.5 nm/RIU within the range of 1.3330-1.3612. It is important to point out that filling channel B with cured PDMS almost makes it immune to SRI variations. The measured shows minute fluctuations in the wavelength shift, which may be attributed to temperature fluctuations of the TEC module and noise in the source and spectrometer. When the ambient temperature changes in the range of 20-50 °C, the temperature sensitivities of channels A and B are -0.173 nm/°C and -1.604 nm/°C, respectively, which indicates that the sensor is capable to compensate for temperature as intended. The LOD is an important characteristic that describes the ability of biosensors to detect the analyte. Following the International Union of Pure and Applied Chemistry’s definition, the LOD is the concentration of analyte derived from the smallest measure that can be detected with reasonable certainty [49]. Using to the calculation formula, where is the resolution of the spectrometer (0.1 nm in this work) and is the sensitivity of the sensor, we find the temperature and the RI LODs to be 0.06 °C and 5 × 10−5 RIU, respectively. The measured wavelength shift of the SPR excited in the two channels of the sensor displays a good linear relationship with SRI and temperature. According to both the simulation analysis and the experimental verification, channel A is sensitive to both SRI and temperature while channel B is sensitive only to temperature. Therefore, both SRI and temperature information can be deduced by:
4.3 Glucose and temperature simultaneous measurement
As discussed above, the DSPRFS enables real-time monitoring of glucose concentrations with no temperature crosstalk. Prior to the glucose solution measurement, the channel A surface needs to be functionalized by immersing the DSPRFS in a PBS solution with GOx. Figure 8(a) shows the transmission spectra evolution during the GOx immobilization. The inset in Fig. 8(a) shows the shift of, which is redshifted by about 11 nm after immobilizing. In order to calibrate the sensor, we fix the temperature at 40 °C using the TEC module and then sequentially inject glucose solutions with different concentrations in the range from 0 to 5 mM into the flow cell. The measured spectral data is generally stable after the glucose solution is added uniformly for about 30 s, which is similar to what was observed in a previous study [50,51]. Due to the enzymatic reaction of glucose molecules with GOx, the effective dielectric constant of the outer medium layer changes, which causes a redshift of while is stable at 704.12 ± 0.5 nm. A concentration of 0 mM corresponds to PBS with pH 7. It is worth noting that GOx has high biological activity at pH 7 and 40 °C [52,53], and thus this is a suitable operation point for glucose concentration sensing with high sensitivity. Figures 8(b) and 8(c) show the transmission spectrum and the SPR resonance wavelength shift of and as a function of the glucose solution concentration in the flow cell, respectively. It can be clearly seen that is significantly redshifted. At lower concentrations in the range 0-1.5 mM this redshift shows good linearity with the increase of the concentration. In this concentration range the GOx can fully react with the glucose molecules, however, the fixed amount of GOx limits the chemical reaction between glucose molecules and GOx at higher concentrations. Thus, above a glucose concentration of 1.5 mM, the wavelength shift of gradually levels off with no change in the redshift at concentrations greater than 4.0 mM. The maximum wavelength shift of is about 13 nM and we consider 4 mM to be the dynamic range of the glucose concentration of the sensor, indicating that it is most suitable for low-concentration glucose solution measurements.
Fig. 8. (a) Transmission spectra evolution of channel A during the GOx immobilization at room temperature. The inset show the SPR resonance wavelength shift. The concentration of GOx is 4 mg/ml, and pH value is 7.4. Experimental data of the dual-channel SPR fiber sensor in the range 0-5 mM under 40°C environment: (b) transmission spectrum, (c) wavelength shift, (d) linear part of wavelength shift and (e) selectivity experiment from 0 to 5 mM.
The dynamic range of the sensor covers that of the normal glucose concentration in tears (0.2 mM) and the higher glucose concentration in diabetic patients (0.92 mM) as reported in [4]. The dynamic range can be improved by using ZnO nanostructures to increase the efficiency of GOx immobilization 13. Figure 8(b) also shows the transmission spectrum evolution of the DSPRFS before and after being immobilized with GOx, where is redshifted. The characteristic equation obtained by polynomial fitting is:
where C is the glucose concentration. The glucose concentration of an unknown sample can thus be obtained by measuring the wavelength shift of at a constant ambient temperature. The linear part is enlarged in Fig. 8(d). The glucose concentration sensitivity of the DSPRFS is calculated to be 6.156 nm/mM in the low concentration range 0-1.5 mM at an ambient temperature of 40 °C.Channel A, which characterizes glucose concentration, is also affected by ambient temperature variations and thus accurate information cannot be obtained from measurement of the wavelength shift in channel A alone. Furthermore, temperature affects the activity of GOx, which reduces the sensitivity to the glucose concentration and the accuracy of the sensor. The output signal can be analyzed by the transfer matrix method to obtain the variations of glucose concentration and temperature:
In addition, we also investigated the specificity of the DSPRFS for glucose versus various typical substances with a high content in human blood: cholesterol, sucrose, KCl, and NaCl. Different kinds of solutions with the same concentration of 5 mM were injected into the flow cell successively. After each solution measurement, the sensor was washed with sufficient PBS to prevent mixing of different solutions causing inaccurate measurement results. The result of the selectivity experiment of the five solutions is shown in Fig. 8(e). It can clearly be seen that the only significant wavelength shift is for the glucose solution, indicating that the DSPRFS is indeed highly selective for glucose and can be used for glucose determination. In order to avoid non-specific adsorption on the sensor surface, i.e. non-specific recognition, several strategies are available; commonly used strategies include chemical antifouling (e.g., zwitterion- or PEG-based highly hydrated chemical surface modifiers), physical antifouling (e.g., substrate topography engineering or membrane filtration) and biological antifouling (e.g., bioaffinity depletion or enzyme catalyzed degradation) [54].
5. Summary
In this work we have theoretically and experimentally demonstrated an ultra-compact label-free in-situ dual-channel surface plasmon resonance fiber sensor that can simultaneously measure glucose concentration and the ambient temperature, which enables compensation of temperature variations yielding accurate glucose concentration measurement results. The experimental results show that the glucose concentration sensitivity and temperature sensitivity are as high as 6.156 nm/mM and -1.604 nm/°C, respectively. Glucose concentration and temperature LODs are 16.24 µM and 0.06 °C, respectively. Additionally, as the ambient temperature of the biological reaction can be monitored in real-time during the sensing process, maximum activity of the substances involved can be ensured. The proposed sensor has satisfactory selectivity, an extremely compact structure, high reliability, reasonable performance-price ratio, and is easy to fabricate, making it a very suitable for glucose sensing as well as for sensing of other biological analytes that requires temperature compensation.
Funding
National Natural Science Foundation of China (61774131, 91833303); National Key Research and Development Program of China (2018YFC1407503); China Postdoctoral Science Foundation (2018M642423); Fundamental Research Funds for the Central Universities (2019FZA5002, Zhejiang University NGICS Platform).
Disclosures
The authors declare no conflicts of interest.
References
1. “Diagnosis and Classification of Diabetes Mellitus,” Diabetes Care 33(Supplement 1), S62 LP-S69 (2010).
2. W. C. Liao, Y. K. Tu, M. S. Wu, J. T. Lin, H. P. Wang, and K. L. Chien, “Blood glucose concentration and risk of pancreatic cancer: systematic review and dose-response meta-analysis,” BMJ [Br. Med. J.] 349(jan02 3), g7371 (2015). [CrossRef]
3. T. Pillinger, K. Beck, C. Gobjila, J. G. Donocik, S. Jauhar, and O. D. Howes, “Impaired Glucose Homeostasis in First-Episode Schizophrenia: A Systematic Review and Meta-analysis,” JAMA Psychiatry 74(3), 261–269 (2017). [CrossRef]
4. D. K. Sen and G. S. Sarin, “Tear glucose levels in normal people and in diabetic patients,” Br. J. Ophthalmol. 64(9), 693–695 (1980). [CrossRef]
5. J. C. Pickup, F. Hussain, N. D. Evans, O. J. Rolinski, and D. J. S. Birch, “Fluorescence-based glucose sensors,” Biosens. Bioelectron. 20(12), 2555–2565 (2005). [CrossRef]
6. A. Heller and B. Feldman, “Electrochemical Glucose Sensors and Their Applications in Diabetes Management,” Chem. Rev. 108(7), 2482–2505 (2008). [CrossRef]
7. Q.-L. Liao, H. Jiang, X.-W. Zhang, Q.-F. Qiu, Y. Tang, X.-K. Yang, Y.-L. Liu, and W.-H. Huang, “A single nanowire sensor for intracellular glucose detection,” Nanoscale 11(22), 10702–10708 (2019). [CrossRef]
8. Q. Liu, Y. Yang, X. Lv, Y. Ding, Y. Zhang, J. Jing, and C. Xu, “One-step synthesis of uniform nanoparticles of porphyrin functionalized ceria with promising peroxidase mimetics for H2O2 and glucose colorimetric detection,” Sens. Actuators, B 240, 726–734 (2017). [CrossRef]
9. M. S. Steiner, A. Duerkop, and O. S. Wolfbeis, “Optical methods for sensing glucose,” Chem. Soc. Rev. 40(9), 4805–4839 (2011). [CrossRef]
10. H. Yuan, W. Ji, S. Chu, S. Qian, F. Wang, J. F. Masson, X. Han, and W. Peng, “Fiber-optic surface plasmon resonance glucose sensor enhanced with phenylboronic acid modified Au nanoparticles,” Biosens. Bioelectron. 117(June), 637–643 (2018). [CrossRef]
11. M. Lobry, D. Lahem, M. Loyez, M. Debliquy, K. Chah, M. David, and C. Caucheteur, “Non-enzymatic D-glucose plasmonic optical fiber grating biosensor,” Biosens. Bioelectron. 142(June), 111506 (2019). [CrossRef]
12. M. Li, S. K. Cushing, and N. Wu, “Plasmon-enhanced optical sensors: a review,” Analyst 140(2), 386–406 (2015). [CrossRef]
13. S. K. Arya, S. Saha, J. E. Ramirez-Vick, V. Gupta, S. Bhansali, and S. P. Singh, “Recent advances in ZnO nanostructures and thin films for biosensor applications: Review,” Anal. Chim. Acta 737, 1–21 (2012). [CrossRef]
14. A. Tereshchenko, M. Bechelany, R. Viter, V. Khranovskyy, V. Smyntyna, N. Starodub, and R. Yakimova, “Optical biosensors based on ZnO nanostructures: Advantages and perspectives. A review,” Sens. Actuators, B 229, 664–677 (2016). [CrossRef]
15. J. Zang, C. M. Li, X. Cui, J. Wang, X. Sun, H. Dong, and C. Q. Sun, “Tailoring Zinc Oxide Nanowires for High Performance Amperometric Glucose Sensor,” Electroanalysis 19(9), 1008–1014 (2007). [CrossRef]
16. Y. F. Li, Z. M. Liu, Y. L. Liu, Y. H. Yang, G.-L. Shen, and R.-Q. Yu, “A mediator-free phenol biosensor based on immobilizing tyrosinase to ZnO nanoparticles,” Anal. Biochem. 349(1), 33–40 (2006). [CrossRef]
17. R. Zhong, Q. Tang, S. Wang, H. Zhang, F. Zhang, M. Xiao, T. Man, X. Qu, L. Li, W. Zhang, and H. Pei, “Self-Assembly of Enzyme-Like Nanofibrous G-Molecular Hydrogel for Printed Flexible Electrochemical Sensors,” Adv. Mater. 30(12), 1706887 (2018). [CrossRef]
18. R. Atchudan, N. Muthuchamy, T. N. J. I. Edison, S. Perumal, R. Vinodh, K. H. Park, and Y. R. Lee, “An ultrasensitive photoelectrochemical biosensor for glucose based on bio-derived nitrogen-doped carbon sheets wrapped titanium dioxide nanoparticles,” Biosens. Bioelectron. 126, 160–169 (2019). [CrossRef]
19. H. Muguruma, A. Hiratsuka, and I. Karube, “Thin-Film Glucose Biosensor Based on Plasma-Polymerized Film: Simple Design for Mass Production,” Anal. Chem. 72(11), 2671–2675 (2000). [CrossRef]
20. L. Su, X. Qiu, L. Guo, F. Zhang, and C. Tung, “Amperometric glucose sensor based on enzyme-modified boron-doped diamond electrode by cross-linking method,” Sens. Actuators, B 99(2-3), 499–504 (2004). [CrossRef]
21. A. I. Minett, J. N. Barisci, and G. G. Wallace, “Immobilisation of anti-Listeria in a polypyrrole film,” React. Funct. Polym. 53(2-3), 217–227 (2002). [CrossRef]
22. S. K. Smith, L. Z. Lugo-Morales, C. Tang, S. P. Gosrani, C. A. Lee, J. G. Roberts, S. W. Morton, G. S. McCarty, S. A. Khan, and L. A. Sombers, “Quantitative Comparison of Enzyme Immobilization Strategies for Glucose Biosensing in Real-Time Using Fast-Scan Cyclic Voltammetry Coupled with Carbon-Fiber Microelectrodes,” ChemPhysChem 19(10), 1197–1204 (2018). [CrossRef]
23. M. Maniruzzaman, S.-D. Jang, and J. Kim, “Titanium dioxide–cellulose hybrid nanocomposite and its glucose biosensor application,” Mater. Sci. Eng., B 177(11), 844–848 (2012). [CrossRef]
24. Y. F. Li, Z. M. Liu, Y. L. Liu, Y. H. Yang, G.-L. Shen, and R.-Q. Yu, “A mediator-free phenol biosensor based on immobilizing tyrosinase to ZnO nanoparticles,” Anal. Biochem. 349(1), 33–40 (2006). [CrossRef]
25. X. Liu, Q. Hu, Q. Wu, W. Zhang, Z. Fang, and Q. Xie, “Aligned ZnO nanorods: A useful film to fabricate amperometric glucose biosensor,” Colloids Surf., B 74(1), 154–158 (2009). [CrossRef]
26. R. Kumar, A. S. Kushwaha, M. Srivastava, H. Mishra, and S. K. Srivastava, “Enhancement in sensitivity of graphene-based zinc oxide assisted bimetallic surface plasmon resonance (SPR) biosensor,” Appl. Phys. A 124(3), 235 (2018). [CrossRef]
27. K. Shah, S. Shukla, N. K. Sharma, and V. Sajal, “Theoretical study of surface plasmon resonance based fiber optic sensor utilizing an additional layer of zinc oxide,” Optik 127(14), 5743–5749 (2016). [CrossRef]
28. A. Paliwal, M. Tomar, and V. Gupta, “Table top surface plasmon resonance measurement system for efficient urea biosensing using ZnO thin film matrix,” J. Biomed. Opt. 21(8), 087006 (2016). [CrossRef]
29. R. Tabassum and B. D. Gupta, “Surface plasmon resonance based fiber optic detection of chlorine utilizing polyvinylpyrolidone supported zinc oxide thin films,” Analyst 140(6), 1863–1870 (2015). [CrossRef]
30. H. M. Kalisz, J. Hendle, and R. D. Schmid, “Structural and biochemical properties of glycosylated and deglycosylated glucose oxidase from Penicillium amagasakiense,” Appl. Microbiol. Biotechnol. 47(5), 502–507 (1997). [CrossRef]
31. V. G. Pahurkar, Y. S. Tamgadge, A. B. Gambhire, and G. G. Muley, “Glucose oxidase immobilized PANI cladding modified fiber optic intrinsic biosensor for detection of glucose,” Sens. Actuators, B 210, 362–368 (2015). [CrossRef]
32. X. Zhang, Z. Wu, F. Liu, Q. Fu, X. Chen, J. Xu, Z. Zhang, Y. Huang, Y. Tang, T. Guo, and J. Albert, “Hydrogen peroxide and glucose concentration measurement using optical fiber grating sensors with corrodible plasmonic nanocoatings,” Biomed. Opt. Express 9(4), 1735 (2018). [CrossRef]
33. S. Yu, L. Ding, H. Lin, W. Wu, and J. Huang, “A novel optical fiber glucose biosensor based on carbon quantum dots-glucose oxidase/cellulose acetate complex sensitive film,” Biosens. Bioelectron. 146(August), 111760 (2019). [CrossRef]
34. J. S. Velázquez-González, D. Monzón-Hernández, D. Moreno-Hernández, F. Martínez-Piñón, and I. Hernández-Romano, “Simultaneous measurement of refractive index and temperature using a SPR-based fiber optic sensor,” Sens. Actuators, B 242, 912–920 (2017). [CrossRef]
35. A. Lokman, H. Arof, and S. W. Harun, “Dumbbell shaped inline Mach–Zehnder interferometer for glucose detection,” Measurement 59, 167–170 (2015). [CrossRef]
36. L. Viet Nguyen, D. Hwang, S. Moon, D. Seung Moon, and Y. Chung, “High temperature fiber sensor with high sensitivity based on core diameter mismatch,” Opt. Express 16(15), 11369–11375 (2008). [CrossRef]
37. E. Reyes-Vera, C. M. B. Cordeiro, and P. Torres, “Highly sensitive temperature sensor using a Sagnac loop interferometer based on a side-hole photonic crystal fiber filled with metal,” Appl. Opt. 56(2), 156–162 (2017). [CrossRef]
38. Y. Peng, J. Hou, Z. Huang, and Q. Lu, “Temperature sensor based on surface plasmon resonance within selectively coated photonic crystal fiber,” Appl. Opt. 51(26), 6361–6367 (2012). [CrossRef]
39. R. C. Jorgenson and S. S. Yee, “A fiber-optic chemical sensor based on surface plasmon resonance,” Sens. Actuators, B 12(3), 213–220 (1993). [CrossRef]
40. J. Homola and M. Piliarik, “Surface Plasmon Resonance (SPR) Sensors BT - Surface Plasmon Resonance Based Sensors,” in J. Homola, ed. (Springer, Berlin Heidelberg, 2006), pp. 45–67.
41. S. B. Bankar, M. V. Bule, R. S. Singhal, and L. Ananthanarayan, “Glucose oxidase — An overview,” Biotechnol. Adv. 27(4), 489–501 (2009). [CrossRef]
42. Y. Zhao, Q. lu Wu, and Y. nan Zhang, “Theoretical analysis of high-sensitive seawater temperature and salinity measurement based on C-type micro-structured fiber,” Sens. Actuators, B 258, 822–828 (2018). [CrossRef]
43. J. Banerjee, M. Bera, and M. Ray, “Simultaneous excitation of multi-spectral surface plasmon resonance using multi-stepped-thickness metallic film,” J. Appl. Phys. 117(11), 113102 (2015). [CrossRef]
44. M. Mitsushio, K. Miyashita, and M. Higo, “Sensor properties and surface characterization of the metal-deposited SPR optical fiber sensors with Au, Ag, Cu, and Al,” Sens. Actuators, A 125(2), 296–303 (2006). [CrossRef]
45. X. Yang, Y. Yuan, Z. Dai, F. Liu, and J. Huang, “Optical property and adsorption isotherm models of glucose sensitive membrane based on prism SPR sensor,” Sens. Actuators, B 237, 150–158 (2016). [CrossRef]
46. S. P. Usha, A. M. Shrivastav, and B. D. Gupta, “FO-SPR based dextrose sensor using Ag/ZnO nanorods/GOx for insulinoma detection,” Biosens. Bioelectron. 85, 986–995 (2016). [CrossRef]
47. S. Wu, G. Yan, B. Zhou, E. H. Lee, and S. He, “Open-Cavity Fabry-Perot Interferometer Based on Etched Side-Hole Fiber for Microfluidic Sensing,” IEEE Photonics Technol. Lett. 27(17), 1813–1816 (2015). [CrossRef]
48. Z. Liu, Z. Zhu, L. Liu, Y. Zhang, Y. Wei, Y. Wang, Y. Zhang, E. Zhao, X. Yang, J. Yang, and L. Yuan, “Dual-truncated-cone structure for quasi-distributed multichannel fiber surface plasmon resonance sensor,” Opt. Lett. 41(18), 4320 (2016). [CrossRef]
49. G. P. Nic M, J. Jirat, B. Košata, S. A. Jenkins, and A. D. Mcnaught, “IUPAC, Compendium of Chemical Terminology,” Encycl. Dict. Polym. (2006).
50. B. Luo, Z. Yan, Z. Sun, Y. Liu, M. Zhao, and L. Zhang, “Biosensor based on excessively tilted fiber grating in thin-cladding optical fiber for sensitive and selective detection of low glucose concentration,” Opt. Express 23(25), 32429–32440 (2015). [CrossRef]
51. D. Li, D. Yang, J. Yang, Y. Lin, Y. Sun, H. Yu, and K. Xu, “Glucose affinity measurement by surface plasmon resonance with borate polymer binding,” Sens. Actuators, A 222, 58–66 (2015). [CrossRef]
52. H. M. Kalisz, H. J. Hecht, D. Schomburg, and R. D. Schmid, “Effects of carbohydrate depletion on the structure, stability and activity of glucose oxidase from Aspergillus niger,” Biochim. Biophys. Acta, Protein Struct. Mol. Enzymol. 1080(2), 138–142 (1991). [CrossRef]
53. T. Kong, Y. Chen, Y. Ye, K. Zhang, Z. Wang, and X. Wang, “An amperometric glucose biosensor based on the immobilization of glucose oxidase on the ZnO nanotubes,” Sens. Actuators, B 138(1), 344–350 (2009). [CrossRef]
54. C. Jiang, G. Wang, R. Hein, N. Liu, X. Luo, and J. J. Davis, “Antifouling Strategies for Selective in Vitro and in Vivo Sensing,” Chem. Rev. 120(8), 3852–3889 (2020). [CrossRef]
