Investigation for terminal reflection optical fiber SPR glucose sensor and glucose sensitive membrane with immobilized GODs

Yinquan Yuan; Xi Yang; Dejing Gong; Fang Liu; Wenbin Hu; Weiquan Cai; Jun Huang; Minghong Yang

doi:10.1364/OE.25.003884

1. Introduction

Glucose sensors have attracted continuous attentions in the past few decades due to considerable and urgent requirements in diabetes diagnosis, food industry, biotechnology, and environmental protection [1–3]. Glucose sensors can be classified into two types, i.e., non-enzymatic sensors and enzymatic biosensors [1–3]. The non-enzymatic glucose sensors have outstanding superiorities in stability, simplicity, reproducibility, and freedom from oxygen limitation, but bear the fatal drawback of low selectivity [4, 5]. The enzymatic biosensors can well meet the requirements of high selectivity and sensitivity, although they have the weaknesses from the thermal and chemical instabilities of enzymes [6, 7]. From the view of physics, glucose sensors can also be classified into acoustic [8], thermal [9], electrical [1] and optical [2, 3] sensors, wherein the major sensing methods are electrochemical sensors [1] and optical sensors [2, 3]. For the electrochemical glucose biosensors, a comprehensive review has summarized the history, principles, major developments, current status, key opportunities and challenges [1]. By improving various electrode materials and structures [10, 11], the electrochemical sensors have the unique advantages of high sensitivity, robustness, low fabrication cost and operational simplicity, but they are susceptible to electromagnetic interferences.

The optical glucose sensors are mainly based on fluorescence signals [12, 13], chemiluminescent signals [14], optical fiber gratings [15, 16], and surface plasmon resonance signals [17–22]. Therein, the optical glucose sensors based on the intensity measurement of fluorescence or chemiluminescent signal are susceptible to intensity fluctuation of excitation light, but the optical fiber grating and surface plasmon resonance sensors, by measuring their characteristic wavelength or resonance angle, may avoid the effect of intensity fluctuation of light source. Surface plasmon resonance (SPR) refers to the optical excitation of surface plasmon wave at the interface between a noble metal (gold or silver) and a dielectric. SPR sensors can be divided into two kinds: prism based SPR sensors and optical fiber SPR sensors. Till now, the most commonly used SPR sensors are the prism based SPR sensors [17, 18], but the optical fiber SPR sensors have also attracted wide attention and strong interest [19–22], especially the terminal reflection SPR sensors have the unique advantage of probe miniaturization compared to the transmission SPR sensors [23,24].

To achieve the high selectivity of optical glucose sensors, the utilizing of enzymes is necessary for glucose monitoring, and immobilized glucose oxidases (GODs) on a support have drawn significant attentions for the stability, reusability and high activity [25–27]. The technique of GODs immobilization and the support property are key factors to develop excellent glucose sensors. The common methods for GODs immobilization are physics adsorption [28], encapsulation [29], covalent binding [30], and chemical crosslinking [31]. Physical adsorption and encapsulation can achieve high activity, but chemical crosslinking can bring better stability [31], so we used the chemical crosslinking to immobilize the GODs on the surface of SiO₂ nanoparticles. In most cases, various inorganic and organic gels, such as silica gels [26, 29], hydrogels [32–34], chitosan [14], aminopropyltriethoxysilane [15], polyethylenimine and polyacrylicacid [16], polyacrylamide [19], and so on, are often used as host matrix in which enzyme molecules can diffuse and be entrapped. The polyacrylamide (PAM) gel is stable, nontoxic and biocompatible, moreover the thickness of gel film and the pore size in gel could be adjusted by changing the process parameters, the ratios of monomers and crosslinking agent, and these advantages allow the PAM gel to be a kind of excellent host matrix [19].

Various host matrixes with immobilized GODs may be considered as glucose sensitive membranes (GSMs), of course, their optical property and adsorption kinetics directly affect the performance of optical glucose sensors. But less papers discusses systematically the refractive index (RI) and adsorption law of GSM till now. In previous paper [35], based on the prism based SPR glucose sensor in angle interrogation, we investigated the RI and adsorption models of a kind of GSM which was made by spin-coating the PAM gel trapped immobilized GODs on the surface of flat gold film. In this paper, looking at high sensitivity, fine selectivity, good stability, intrinsic safety and miniaturized probe, we have constructed an terminal reflection optical fiber SPR glucose sensor coated with a GSM, where the GSM was made by dip-coating the PAM gel trapped immobilized GODs on the surface of cylindrical gold film, moreover investigated the optical property and adsorption isotherm model of the GSM. In Section 2, we have reported the fabrication and characterization of the GSM with immobilized GODs. The merit of immobilized GOD is to prevent enzyme leaking from the gel while allowing free movements of glucoses and its by-products. In Section 3, we have simulated the terminal reflection optical fiber SPR sensor with GSM, discussed the relationship between the resonance wavelength versus the RI of GSM, and the relationship between the RI of GSM versus the glucose concentration in the solution based on the isotherm adsorption model of GSM. In Section 4, we have investigated the effects of the film thickness, GOD content in GSM, the pH of solution and measuring temperature. Progressively, using the optimized sensors coated with immobilized GODs, we have obtained their resonance spectra of the sensors in sample solutions and the absorption isotherms of GSM, by which one can understand the absorption mechanism of GSM for glucose molecules.

2. Experimental

2.1. Materials

N-tetramethylethylenediamine (TEMED), ammonium persulphate (APS), acrylamide and bisacrylamide, SiO₂ nanoparticles, glucose reagent, potassium phosphate monobasic and sodium phosphate dibasic were purchased from Sinopharm Chemical Reagent. Glucose oxidase (GOD, EC 1.1.3.4, from Aspergillusniger, 100 U/mg) was obtained from Biosharp, 3-aminopropiltrietoxysilane (APTES) and glutaraldehyde (GA) were purchased from Alfa Aesar. Phosphate-buffered solution (0.1 M) was prepared by using potassium phosphate monobasic and sodium phosphate dibasic in deionized water. All other reagents were of analytical grade and used without further purification. Deionized water was used throughout the experiments [35].

2.2. Immobilization of GODs on SiO₂ nanoparticles

The immobilization of GODs in a suitable matrix is beneﬁcial to the retention and reuse of isolated enzymes and to build the sensors with high performance. A typical immobilization of GODs on the surface of SiO₂ nanoparticles was performed as following [35]. At first, 4 mL of freshly prepared APTES solution (2% (v/v)) was stirred for 30 min to complete its hydrolyzation. 40 mg of SiO₂ nanoparticles was added into the solution and continued to stir for 18-24 h at room temperature of 25 °C. The modified SiO₂ nanoparticles were collected by centrifugation and washed for several times in deionized water and redispersed in a certain volume of phosphate buffer solution (PBS, 0.1 M, pH = 7.4). 160 μL (25% v/v) of GA was added into the mixture and stirred slowly for 1.5 h at 25 °C. The mixture was washed several times with PBS buffer (0.1 M, pH = 7.4), and once with PBS buffer (0.1 M, pH = 6.5). The activated SiO₂ nanoparticles were obtained. Then, the activated SiO₂ nanoparticles were dispersed in PBS buffer, the GODs were added to the mixture with the content of 1 mg/mL, and put the mixture in a centrifugetube at 4 °C for 12 h with occasional shaking. The SiO₂ nanoparticles crosslinked with the GODs were washed in PBS buffer for several times to remove the residual reactants and free GODs, and were stored in PBS buffer (0.1 M, pH = 7.0) under 4 °C for use.

The FT-IR device (Thermoelectron scientific instruments, Nicolet6700, U.S.) was used to test if the GODs were immobilized on the surface of SiO₂ nanoparticles as shown in Fig. 1(a) [35]. The FT-IR spectra of SiO₂ nanoparticles showed that the features at 1109, 801 and 474 cm⁻¹ are related to asymmetric stretching mode, symmetric stretching mode and bending mode of Si-O bonds, respectively. The bands at 965 and 3441 cm⁻¹ are ascribed to Si-OH groups and stretching modes of -OH groups respectively [36, 37]. The FT-IR spectra of GODs showed four infrared bands with the center positions at 3300 cm⁻¹ (amide A), 1657 cm⁻¹ (amide I), 1538 cm⁻¹ (amide II) and 1246 cm⁻¹ (amide III). The peak at 3300 cm⁻¹ can be related to amide A with a shoulder at 3064 cm⁻¹. A peak at 1657 cm⁻¹ is attributed to C = O stretching vibrations of the peptide linkages in the backbone and the band at 1538 cm⁻¹ is associated with the combination of C-N stretching and N-H in-plane bending [38, 39]. From the spectra of GOD/SiO₂, the features of SiO₂ nanoparticles and GODs are reserved except for the band of Si-OH groups at 965 cm⁻¹, which means that the Si-OH groups is reacted with the groups of crosslinker.

Fig. 1 FT-IR spectra of GOD/SiO₂/PAM film.

Download Full Size | PDF

2.3. Fabrication of probes

The multimode optical fiber (MOF) with numerical aperture of 0.39 and core diameter of 600 μm was purchased from Thorlabs. Firstly, enough numbers of MOF segments (about 10 cm lengths) were cut off and their ends were grinded by optical fiber muller to be uniform and smooth. Then, the cladding of end portion (about 1 cm length) of every MOF segment was removed by a sharp blade and the uncladed portions of the MOF segments were cleaned with acetone and alcohol, followed by dry in vacuum. Thirdly, the cylindricalsurface of uncladed portions were deposited with a gold film of 50 nm thickness by using BESTECH magnetron sputtering machine from Berlin, Germany, this gold film coated portion acts as the sensing region of the fiber probe. Fourthly, the end surface was deposited with a gold film of more than 200 nm thickness which plays the role of optical reflector. The thickness of Au film was monitored by a quartz digital thickness monitor fixed inside magnetron sputtering machine and an equipped rotation device of MOF segments ensures the thickness uniform of gold film.

The acrylamide/bisacrylamide (29:1) mixed powder was added to the PBS solution including immobilized GODs, and then the immobilized GODs got trapped within the gel matrix in the PBS solution. As photochemical initiator, 10% ammonium persulphate was also added into the PBS solution. In order to accelerate the reaction, TEMED was added into the PBS solution, and then the polymerization gets initiated in the solution. To increase the hydrophilicity of the gold film surface of optical fiber tip, the optical fiber SPR tip was rinsed with ethanol and acetone several times, and soaked in the PBS solution for 30 min. Then, a thin GSM trapped the immobilized GODs was coated on the gold film of SPR probe by using the dip-coating instrument (ZR-4200) purchased from Qingdao Zhongrui Intelligent Instrument Co., Ltd., China. Subsequently, the SPR probes coated with GSM were allowed to dry for few minutes at 25 °C and stored at 4 °C to prevent the enzyme from losing its activity.

The FT-IR spectra of the GOD/SiO₂, PAM gel and GOD/SiO₂@PAM gel samples are shown in Fig. 1(b). The absorption band at 3406 cm⁻¹ is owing to N-H stretching of NH₂ group. The bands at 1670 cm⁻¹ (amide I) due to C = O stretching, at 1609 cm⁻¹ (amide II) due to the N-H bending and at 621 cm⁻¹ (amide IV) due to the OCN deformation in amide, have been observed and these bands imply the existence of PAM. The absorption band at 3199 cm⁻¹ is related to the symmetrical stretch vibration of –NH₂ groups. The characteristic peaks at 2948 and 2868 cm⁻¹ reflect the C-H stretching vibration within PAM polymers. The peaks at 1324, 1118 and 986 cm⁻¹ are associated with C-O and C-N stretching bands. In addition, the band at 1454 cm⁻¹ is ascribed to CH₂ scissoring. The FT-IR spectrum of GOD/SiO₂@PAM film contains the features of all components, of which the characteristic peaks at about 1670 and 1609 cm⁻¹ correspond to the overlapping absorption of GOD and PAM [40, 41].

2.4. Apparatus

Figure 2 shows the terminal reflection optical fiber SPR sensing system. The sensing system consists of (1) a light source (HL-2000 Tungsten Halogen Light Sources from Ocean Optics, wavelength range: 360-1200 nm), (2) a Y-type optical fiber coupler, (3) a spectrometer (Ocean Optics, USB4000, wavelength range: 320-1050 nm, wavelength resolution: 0.3 nm), (4) a terminal reflection optical fiber tip, and (5) a microcomputer. Through the Y-type optical coupler, the incident light emitted from the light source goes to the terminal reflection probe; the reflected light produced by the sensing probe contains the glucose information of GSM, through the optical coupler, goes back to the spectrometer; the output spectrum is recorded by the spectrometer and then transferred to a computer for data analysis.

Fig. 2 Terminal reflection optical fiber SPR sensing system. (1) light source, (2) Y-type optical fiber coupler, (3) spectrometer, (4) a terminal reflection optical fiber tip, (5) a microcomputer.

Download Full Size | PDF

3. Simulations

For the optical fiber SPR glucose sensor, the output quantity (resonance wavelength λ_res), by an intermediate variable, the RI of GSM, can be expressed as the function of glucose concentration.

λ_{res} = λ_{res} [n_{f} (C)]

and its sensitivity is

S = \frac{\partial λ_{res}}{\partial C} = \frac{\partial λ_{res}}{\partial n_{f}} \frac{\partial n_{f}}{\partial C}

The factor ∂λ_res/∂n_f is decided by the sensing structure; the factor ∂n_f/∂C depends on the material property and pore structure of GSM which decides the adsorptions of glucose molecule to GSM.

To understand the dependence relationship of resonance wavelength (λ_res) on the RI of GSM (n_f), we have made the simulations based on planar structure approximation. In Fig. 2, the optical fiber tip is cylindrical, whose surface is coated with Au film and the flat end is coated with totally reflected Au film. Let a₁ denotes the radius of fiber core, L denotes the tip length. The angle range of the rays is [θ_cr, π/2], where θ_cr = sin⁻¹(n_cl/n₁) is critical angle, n₁ is the RI of fiber core and n_cl is the RI of fiber cladding. As the excitation light from a collimated source is launched into one end of the fiber at the axial point, the normalized reflecting power of p-polarized light can be written as [24]

P_{refl} = \frac{\int_{θ_{cr}}^{π / 2} R_{p}^{2 N_{ref} (θ)} P (θ) d θ}{\int_{θ_{cr}}^{π / 2} P (θ) d θ}

where

P (θ) = \frac{n_{1}^{2} \sin θ \cos θ}{{(1 - n_{1}^{2} \cos^{2} θ)}^{2}}

N_{r e f} (θ) = \frac{L}{2 a_{1} \tan θ}

And R_p is the amplitude reflection coefficient for p-polarized incident wave [24], and can be calculated by using four-layer configuration (fiber core/Au layer/GSM/glucose solution). The first layer is the fiber core with the RI of n₁; the second layer is the Au film with thickness of d₂, whose dielectric constant is denoted by ε_m; the third layer is the sensing layer with thickness of d₃, whose RI and dielectric constant are denoted by n_f and ε_f, respectively, and they are related to the concentration of goal molecules in the sensing layer; the fourth layer is the sample, whose RI and dielectric constant are denoted by n_sam and ε_sam, respectively. Additionally, the sensing probe should be immersed in the measured solution, so that the surface plasmon resonance could be excited for the sensing.

Considering the experimental data, it is, the resonance wavelength decreases from 679.50 nm to about 656.9 nm as the RI of film decreases from 1.470 to an unknown quantity, we simulated the SPR sensor working in the wavelength interrogation mode, where n₁ = 1.610, n_cl = 1.600, a₁ = 300 μm, d₂ = 50 nm and d₃ = 1000 nm were chosen. Figure 3(a) shows the reflection spectra for the different RIs of GSM, and Fig. 3(b) gives the resonance wavelengths corresponding to different values of n_f by squares. n_f0 = 1.470 represents the initial RI of GSM, and the solid line was drawn according to the fitting equation:

λ_{res} (n_{f}) = λ_{res} (n_{f0}) + ξ_{1} (n_{f} - n_{f0}) + ξ_{2} {(n_{f} - n_{f0})}^{2}

where R² = 0.9997, λ_res(n_f0) = 679.50 nm, ξ₁ = 1719.794 nm/RIU and ξ₂ = 7066.017 nm/RIU². Since the experimental value of resonance wavelength shift is about 22 nm, we can estimate that the difference of RI is less than 0.01 RIU, so the third term is much smaller than the second term, and the resonance wavelength has approximate linear relationship with the film RI is in the RI range of 1.450-1.470 and the change rate equals to 1719.794 nm/RIU. For the case of optical fiber SPR sensor, due to the differences between the non-inclusion of skew rays and the two-dimensional simulations based on geometrical optics, there exists deviation between the simulated data and experimental results, so we have made necessary corrections for the model parameters. Based on the measured reflection spectra of the optimized sensor in the solution with no glucose (the resonance wavelength is 678.5 nm corresponding to the film RI of 1.470), the equivalent RIs of n₁ = 1.610 and n_cl = 1.600 can be obtained by the reconstitution method according to Eqs. (3)-(5).

Fig. 3 Simulated reflection spectra (a) and resonance wavelength versus RI of GSM (b).

Download Full Size | PDF

4. Results and discussion

We have finished a serial of glucose sensing experiments and investigated the effects of GSM thickness, GOD concentration in GSM, solution pH and ambient temperature on the reflection spectra, and then obtained the optimized parameters of the probe.

4.1. Effect of GSM thickness

The lifting instrument was purchased from Qingdao Zhongrui Intelligent Instrument Co., Ltd., China (ZR-4200). The PAM gel with immobilized GODs (1 mg/mL) was prepared in 0.01 M PBS of pH 7.0. The SPR probes were prepared in the PAM gel by using P-times dip-coating technique (elevation velocity of 50 mm/min, room temperature of 25 °C, relative humidity of about 60%, 30 minutes of interval time between two dip-coatings), where seven SPR probes with different pulling times (P = 2, 5, 7, 10, 12, 15, 17 times, respectively) in the PAM gel were prepared. The probes undergoing P-times pulling has Pl₁ of the thickness of GSM, where l₁ is the average thickness of one pulling. The SPR spectra from these probes in the solutions of 0 mg/dL and 80 mg/dL (pH 7.0) glucose concentrations were recorded. Figure 4 shows the resonance wavelength differences for different GSM thicknesses, which is represented by the pulling times. It may be observed that the shift in resonance wavelength arrives at a maximum of about 4.4 nm at 12 times pulling. The maximum sensitivity at 12 pulling times means the GSM thickness (d₃) arrives at the transmission depth of evanescent wave (d_t) and then the evanescent wave is used to the greatest extent. As the GSM thickness exceeds the best thickness, the glucose molecules in the solution are difficult to enter into the innermost sensitive region of the GSM closed to Au film, which leads to the decrease of sensitivity. So the selected probe with the optimal thickness was made by 12 times pulling.

Fig. 4 Effect of GSM thickness on the resonance wavelength difference.

Download Full Size | PDF

4.2. Effect of GOD content in the GSM

The GOD solutions with different concentrations such as 1, 2, 3, 4 and 5 mg/mL were prepared in the PBS solution of pH 7.0. Five SPR probes were made in these PAM gels by using 12 times pulling respectively. Figure 5 shows the wavelength differences for different GOD contents, it can be seen that the sensitivity increases with the increase of GOD content and arrives at the maximum of about 8.4 nm when the GOD concentration is at 4.0 mg/mL, but the sensitivity begins to decrease as the concentration of GOD is greater than 4 mg/mL. The increase of the sensitivity is easy to understand, since more GOD can catalyst more glucose to produce more gluconic acid and H₂O₂ and lead to greater change in the RI of the sensing film. As the GOD content exceeds its saturation value (4.0 mg/mL), further increase of GOD content will probably cause steric hindrance which declines the enzyme.

Fig. 5 Effect of GOD content in GSM on the resonance wavelength difference.

Download Full Size | PDF

4.3. Effect of solution pH

It is well known that high enzyme activity is desirable for a better glucose sensor and the enzyme activity is influenced by pH. In this subsection, we have investigated the effect of solution pH by using the optimal probe (GOD content of 4 mg/mL).The glucose solutions of 0 mg/dL and 80 mg/dL with the pH varied from 5.0 to 9.0 were prepared. Figure 6 gives the wavelength difference for different pHs (5, 6, 6.5, 7, 7.5, 8 and 9), it can be seen that the maximum shift in resonance wavelength can be obtained for pH 7.0. The existence of optimal pH (7.0) may be explained in terms of the point of zero charge on solid surface [42]. As pH is 7.0, the GOD enzyme has the maximum activity and can produce maximum products compared to other pH values, and then a maximum change in the RI of gel layer happened and caused a maximum shift in resonance wavelength in the solution of pH = 7.0.

Fig. 6 Effect of solution pH on the resonance wavelength difference.

Download Full Size | PDF

4.4. Effect of ambient temperature

In this subsection, we have investigated the effect of ambient temperature. The beaker containing the sample solution was placed the temperature adjustable water bath, and the water temperature changes from 20 to 60 °C. Figure 7 shows the wavelength difference for different temperatures, one can see that the maximum shift in resonance wavelength is obtained at the temperature of 50 °C. This change comes mainly from the activity of the GOD enzyme, the activity of the GOD enzyme increases as the temperature rises from 10 °C to the optimum temperature of GODs (50 ~55 °C) and decreases sharply as the temperature exceeds the optimum temperature.

Fig. 7 Effect of measuring temperature on the wavelength difference.

Download Full Size | PDF

4.5. Performance of optimized probes

The optimized optical fiber sensor consists of a cylindrical fiber tip with a gold film of 50 nm thickness and an optimized GSM, and the optimized GSM is prepared of a PAM film with immobilized GODs on the SiO₂ nanoparticles. In order to prevent the GODs from losing activity at the long time of high temperature, we keep the measuring temperature as 40 °C. To characterize the selected sensor, the glucose concentration in PBS solution of pH 7.0 was adjusted from 0 to 500 mg/dL. Figure 8(a) shows the reflection spectra of the sensor in the solutions with different glucose concentrations. As the glucose concentration of the solution increases, the reflection spectrum shifts toward small wavelength, and this blue shift is due to the smaller RIs of the by-products of the enzymatic reaction than the RI of the PAM. Progressively, Fig. 8(b) gives the resonance wavelength differences corresponding to different glucose concentrations. It can be obtained that the resonance wavelength decreases about 21.5 nm when the glucose concentration increases from 0 to 500 mg/dL. Moreover, the resonance wavelength has a good linear relationship with glucose concentration as the glucose concentration is in the range of 0−80 mg/dL, the sensitivity is about 0.14 nm/(mg/dL), the corresponding limit of detection (LOD) is 0.02/0.14 = 0.142 mg/mL for the 0.02 nm wavelength resolution of a spectrometer. At low concentration, the glucose can be absorbed into the PAM film in greater degree, which results in higher production rate of by-products. But as the concentration is larger than 80 mg/dL, the glucose being absorbed into the GSM tends to be saturation, so the plot shows a saturating shape at higher concentrations.

Fig. 8 Reflection spectra of optimized probe (a) and resonance wavelength shift versus glucose concentration (b).

Download Full Size | PDF

4.6. Selectivity and response time

In the process of measuring the glucose concentration in blood, ascorbic acid, sodium ion, potassium ion and chloride ion are the main interferences. Hence, we have investigated the influences of different ions (NaCl, KCl, CaCl₂, MgCl₂, NaHCO₃, Na₂HPO₄), ascorbic acid, fructose, urea and sucrose, their concentrations are 80 mg/dL, and Fig. 9(a) gives their differences of two resonance wavelengths of the optimized sensor. Having similar structure and property as glucose, fructose and sucrose give the shifts of 1.6 nm and 1.7 nm in resonance wavelength respectively. No distinct shifts are observed after the optimized probe was put into the solutions with other interferents, indicating the optimized glucose sensors have good selectivity.

Fig. 9 (a) Selectivity with the interferent concentrations of 80 mg/dL and (b) response time at 675 nm wavelength.

Download Full Size | PDF

To understand the response time of the sensor, the solution with the glucose concentration of 80 mg/dL was used to detect. When the selected probe was put into the solution, the glucose molecules react with dissolve oxygen under the catalysis of GODs and produce the by-products, the resonance wavelength changes and will be stable after about 22 s as plotted in Fig. 9(b).

In this paper, the GODs were immobilized on the activated SiO₂ nanoparticles by chemical crosslink, and then the immobilize GODs were trapped on the PAM gel. Comparing with the physics entrapping of PAM gel [19], no doubt, this immobilization method can improve the efficiency and stability of GODs. In fact, we also observed that the sensitivity of about 0.14 nm/(mg/dL) in the range of 0−80 mg/dL is less lower than the sensitivity given by Singh et al [19], a possible reason may be that the spatial positions of SiO₂ nanoparticles weakened the catalytic action of GODs. Even so, the sensitivity of about 0.14 nm/(mg/dL) in the range of 0−80 mg/dL is much greater than the sensitivity of fiber optic gratings [15, 16].

4.7. Optical property and adsorption models of GSM

GSM is a kind of composite membrane consisting of immobilized GODs on the surface of SiO₂ nanoparticles and PAM matrix, its RI is a measurement of average optical property in its aggregated state, and is also a key factor that affects the resonance spectrum of optical sensor. For composite membrane, its RI is decided by the RIs and volume fractions of all components. In practical applications, the simplest formula is Gladstone-Dale mixing rule [43, 44]:

n_{f} = \sum_{i = 1}^{n c} n_{i} ϕ_{i}

where n_f is the RI and volume of mixture, nc is the number of components, n_i is the RI of pure component i, ϕ_i is the volume fraction of pure component i.

As the GSM was put in the measured environment, due to the adsorption interactions of GSM, some glucose, water and dissolved oxygen molecules were absorbed into the membrane. Under the catalyst of GOD in the GSM, glucose, water and dissolved oxygen react and produce gluconic acid (GA) and H₂O₂ [45]:

C_{6} H_{12} O_{6} + O_{2} + H_{2} O \overset{GOD}{\to} C_{6} H_{10} O_{7} (GA) + H_{2} O_{2}

After enough time, the system will arrive at chemical equilibrium. Considering 1 mol of glucose, consuming 1 mol of oxygen and 1 mol of water, produce 1 mol of GA and 1 mol of hydrogen peroxide, according to Eq. (7), the average RI of the GSM adsorbed glucose molecules can be expressed as

n_{f} = n_{f0} - (n_{f0} - n_{GA}) (1 + \frac{v_{H2O2}}{v_{GA}}) ϕ_{g} (C)

Where ϕ_g(C) is the volume fraction of glucose into the GSM, v_GA = 111.2 cm³/mol and v_H2O2 = 23.5 cm³/mol are the molar volumes of glucose and GA. In the sensing membrane, the initial RI of the PAM gel layer is n_f0 = 1.470, the RIs of gluconic acid and H₂O₂ are n_GA = 1.416 and n_H2O2 = 1.414 respectively [46], they are lower than the RI of the PAM gel layer (n_f0>n_GA≈n_H2O2). When the enzymatic reaction occurs, the overall RI of the gel layer will decrease with the producing of gluconic acid and H₂O₂. Moreover, as the concentration of glucose increases, more glucose molecules and dissolved oxygen diffuse into the gel layer, and react under the action of GODs, which leads to the decrease in the RI of gel layer and the decrease of resonance wavelength.

When the optical probe coated with GSM is placed in the measured sample, the GSM (solid phase) and the sample (liquid phase) form a thermodynamic system. After a sufficient time, the whole system will arrive at a thermodynamic equilibrium. Till now, some adsorption isotherm models, such as Langmuir isotherm, Freundlich isotherm, Dubinin-Radushkevich isotherm, BET isotherm, Temkin isotherm, Flory-Huggins isotherm, Redlich-Peterson isotherm, et al, have been presented. The most commonly used adsorption model is Langmuir adsorption isotherm; it can be expressed as [47, 48]:

\frac{C}{ϕ_{g}} = \frac{1}{ϕ_{gm} K_{L}} + \frac{C}{ϕ_{gm}}

or

ϕ_{g} = ϕ_{gm} \frac{K_{L} C}{1 + K_{L} C}

Where ϕ_gm is the maximum adsorption capacity of glucose molecules in the GSM, K_L is the Langmuir adsorption equilibrium constant. The Langmuir equation assumes that there is no molecular interaction between the adsorbed molecules, the adsorption is localized in a monolayer, and once an adsorbed molecule occupies a site, no further adsorption can take place at that site.

In fact, Freundlich model is the earliest known adsorption isotherm model and can be expressed by the following equations:

ϕ_{g} = K_{F} C_{}^{1 / F r}

\lg ϕ_{g} = \lg K_{F} + \frac{1}{F r} \lg C

where the Freundlich constants K_F and 1/Fr can be determined by a plot of lgϕ_g versus lgC. This model describes reversible adsorption and can be applied to nonideal adsorption on heterogeneous surfaces as well as multilayer adsorption.

Combining Eqs. (6) and (9), the change of resonance wavelength can be rewritten as

Δ λ_{res} (n_{f}) = - ξ_{1} (n_{f0} - n_{GA}) (1 + \frac{v_{H2O2}}{v_{GA}}) ϕ_{g} (C)

and the volume fraction of glucose into the GSM can be obtained:

ϕ_{g} (C) = - \frac{1}{ξ_{1} (n_{f0} - n_{GA}) (1 + \frac{v_{H2O2}}{v_{GA}})} Δ λ_{res} (C)

Considering n_f0 = 1.470, n_GA = 1.416, v_GA = 111.2 cm³/mol, v_H2O2 = 23.5 cm³/mol, one can obtain $ϕ_{g} (C) = - 0.0089 {nm}^{- 1} \cdot Δ λ_{res} (C)$ .Using the data in Fig. 8(b), we obtained the relationship dependence of the volume fraction of glucose into the GSM upon the glucose concentration as shown in Fig. 10(a). It can be seen that, as the glucose concentration is in the region of [0, 80] mg/dL, the volume fraction of glucose into the GSM is good linear with the concentration of glucose, which can be explained by Freundlich isotherm model:

ϕ_{g} = K_{F} C

With Freundlich constant of K_F = 1.3 × 10⁻³ (mg/dL)⁻¹. In our previous paper [35], as the glucose concentration is in the region of [0, 80] mg/dL, Freundlich isotherm and constant K_F = 3.0 × 10⁻³ (mg/dL)⁻¹ were also obtained. The comparison for two values of K_F shows the adsorption constant of the GSM confined by cylindrical optical fiber is smaller than that of the GSM confined by flat gold-glass.

Fig. 10 Volume fraction of glucose into the GSM versus glucose concentration (a) and Langmuir isotherm (b).

Download Full Size | PDF

As the concentration of glucose is in the region of [120, 500] mg/dL, we calculated the values of C/ϕ_g corresponding to different values of C, the obtained values were shown in Fig. 10(b) by close circles, it can be seen that these circles are in a line when the glucose concentration C is in the region of [120, 500] mg/dL. The line in Fig. 10(b) was drawn by the fitting equation:

\frac{C}{ϕ_{g}} = 387.729 + 4.3846 C

Where R² = 0.9997. By which, the maximum adsorption capacity of glucose molecules in the GSM ϕ_gm = 0.2286 and the Langmuir adsorption constant K_L = 0.0113 can be obtained. In previous paper [35], as the glucose concentration is in the region of [80, 200] mg/dL, the model parameters of Langmuir isotherm and D-R isotherm were calculated, and the R² value of D-R model (0.9799) is slightly greater than that of Langmuir model (0.9158), by which we said that D-R model is more suitable to describe the adsorption process. Now, considering the good linear relationship in Fig. 10(b) and R² = 0.9997, we confirmed that the Langmuir isotherm model is more suitable to describe the adsorption process of GSM for glucose in the region of [120, 500] mg/dL.

5. Conclusions

In this paper, a terminal reflection optical fiber SPR glucose sensor has been investigated and a research method for the GSM has been proposed. (1) The fabrication and characterization of the GSM with immobilized GOD on SiO₂ nanoparticles and PAM gel have been reported. (2) In theory, the quantitative relationship between the resonance wavelength of SPR glucose sensor and the RI of GSM, between the RI of GSM and the glucose concentration in GSM, between the glucose concentrations in GSM and in solution, have been outlined and obtained by using the simulation, RI formula and adsorption isotherm models respectively. (3) The optimum parameters of the sensor, such as, GSM thickness of 12 times pulling, 4 mg/mL of GOD content in GSM,7.0 of solution pH and 40 °C of measuring temperature have been obtained. (4) By the optimized sensors and optimum measuring condition, the dependence relationship of the resonance wavelength upon the glucose concentration in sample solution has obtained. (5) The adsorption isotherms of the GSM have been obtained. As the glucose concentration is in the region of [0, 80] mg/dL, the volume fraction of glucose into the GSM is good linear with the concentration of glucose, which can be explained by Freundlich isotherm model with Freundlich constant of K_F = 1.3 × 10⁻³ (mg/dL)⁻¹; As the glucose concentration is in the region of [120, 500] mg/dL, the Langmuir isotherm model is more suitable to describe the adsorption process of GSM for glucose, the maximum adsorption capacity of glucose molecules in the GSM ϕ_gm = 0.2286 and the Langmuir constant K_L = 0.0113 can be obtained.

Acknowledgments

Supported by the Natural Science Foundation of Hubei Province, China (No. 20141j015), the National Nature Science Foundation of China (61402345, 21476179) and New Century Excellent Talents in University of the Ministry of Education (NCET-13-0942).

References and links

1. J. Wang, “Electrochemical glucose biosensors,” Chem. Rev. 108(2), 814–825 (2008). [CrossRef] [PubMed]

2. S. M. Borisov and O. S. Wolfbeis, “Optical biosensors,” Chem. Rev. 108(2), 423–461 (2008). [CrossRef] [PubMed]

3. M. S. Steiner, A. Duerkop, and O. S. Wolfbeis, “Optical methods for sensing glucose,” Chem. Soc. Rev. 40(9), 4805–4839 (2011). [CrossRef] [PubMed]

4. H. Li, C. Y. Guo, and C. L. Xu, “A highly sensitive non-enzymatic glucose sensor based on bimetallic Cu-Ag superstructures,” Biosens. Bioelectron. 63, 339–346 (2015). [CrossRef] [PubMed]

5. X. Niu, X. Li, J. Pan, Y. He, F. Qiu, and Y. Yan, “Recent advances in non-enzymatic electrochemical glucose sensors based on non-precious transition metal materials: opportunities and challenges,” RSC Advances 6(88), 84893–84905 (2016). [CrossRef]

6. A. Kaushik, R. Khan, P. R. Solanki, P. Pandey, J. Alam, S. Ahmad, and B. D. Malhotra, “Iron oxide nanoparticles-chitosan composite based glucose biosensor,” Biosens. Bioelectron. 24(4), 676–683 (2008). [CrossRef] [PubMed]

7. A. Umar, M. M. Rahman, A. Al-Hajry, and Y. B. Hahn, “Enzymatic glucose biosensor based on flower-shaped copper oxide nanostructures composed of thin nanosheets,” Electrochem. Commun. 11(2), 278–281 (2009). [CrossRef]

8. J. Luo, P. Luo, M. Xie, K. Du, B. Zhao, F. Pan, P. Fan, F. Zeng, D. Zhang, Z. Zheng, and G. Liang, “A new type of glucose biosensor based on surface acoustic wave resonator using Mn-doped ZnO multilayer structure,” Biosens. Bioelectron. 49, 512–518 (2013). [CrossRef] [PubMed]

9. C. D. Malchoff, K. Shoukri, J. I. Landau, and J. M. Buchert, “A novel noninvasive blood glucose monitor,” Diabetes Care 25(12), 2268–2275 (2002). [CrossRef] [PubMed]

10. B. Liang, L. Fang, G. Yang, Y. Hu, X. Guo, and X. Ye, “Direct electron transfer glucose biosensor based on glucose oxidase self-assembled on electrochemically reduced carboxyl graphene,” Biosens. Bioelectron. 43, 131–136 (2013). [CrossRef] [PubMed]

11. Y. J. Lee, S. J. Park, K. S. Yun, J. Y. Kang, and S. H. Lee, “Enzymeless glucose sensor integrated with chronically implantable nerve cuff electrode for in-situ inflammation monitoring,” Sens. Actuat. B 222, 425–432 (2016). [CrossRef]

12. G. Chang, Y. Tatsu, T. Goto, H. Imaishi, and K. Morigaki, “Glucose concentration determination based on silica sol-gel encapsulated glucose oxidase optical biosensor arrays,” Talanta 83(1), 61–65 (2010). [CrossRef] [PubMed]

13. J. Peng, Y. Wang, J. Wang, X. Zhou, and Z. Liu, “A new biosensor for glucose determination in serum based on up-converting fluorescence resonance energy transfer,” Biosens. Bioelectron. 28(1), 414–420 (2011). [CrossRef] [PubMed]

14. M. J. Chaichi and M. Ehsani, “A novel glucose sensor based on immobilization of glucose oxidase on the chitosan-coated Fe3O4 nanoparticles and the luminol-H2O2-gold nanoparticle chemiluminesence detection system,” Sens. Actuat. B 223, 713–722 (2016). [CrossRef]

15. 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] [PubMed]

16. A. Deep, U. Tiwari, P. Kumar, V. Mishra, S. C. Jain, N. Singh, P. Kapur, and L. M. Bharadwaj, “Immobilization of enzyme on long period grating fibers for sensitive glucose detection,” Biosens. Bioelectron. 33(1), 190–195 (2012). [CrossRef] [PubMed]

17. W. W. Lam, L. H. Chu, C. L. Wong, and Y. T. Zhang, “A surface plasmon resonance system for the measurement of glucose in aqueous solution,” Sens. Actuat. B 105(2), 138–143 (2005). [CrossRef]

18. 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. Actuat. A 222, 58–66 (2015). [CrossRef]

19. S. Singh and B. D. Gupta, “Fabrication and characterization of a surface plasmon resonance based fiber optic sensor using gel entrapment technique for the detection of low glucose concentration,” Sens. Actuat. B 177, 589–595 (2013). [CrossRef]

20. H. Suzuki, M. Sugimoto, Y. Matsui, and J. Kondoh, “Effects of gold film thickness on spectrum profile and sensitivity of a multimode-optical-fiber SPR sensor,” Sens. Actuat. B 132(1), 26–33 (2008). [CrossRef]

21. Y. Yuan, L. Ding, and Z. Guo, “Numerical investigation for SPR-based optical fiber sensor,” Sens. Actuat. B 157(1), 240–245 (2011). [CrossRef]

22. D. Li, J. Wu, P. Wu, Y. Lin, Y. Sun, R. Zhu, J. Yang, and K. Xu, “Affinity based glucose measurement using fiber optic surface plasmon resonance sensor with surface modification by borate polymer,” Sens. Actuat. B 213, 295–304 (2015). [CrossRef]

23. Y. Zhao, Z. Deng, and Q. Wang, “Fiber optic SPR sensor for liquid concentration measurement,” Sens. Actuat. B 192, 229–233 (2014). [CrossRef]

24. Y. Yuan, D. Hu, L. Hua, and M. Li, “Theoretical investigations for surface plasmon resonance based optical fiber tip sensor,” Sens. Actuat. B 188, 757–760 (2013). [CrossRef]

25. M. Hartmann and D. Jung, “Biocatalysis with enzymes immobilized on mesoporous hosts: the status quo and future trends,” J. Mater. Chem. 20(5), 844–857 (2010). [CrossRef]

26. A. Y. Khan, S. B. Noronha, and R. Bandyopadhyaya, “Glucose oxidase enzyme immobilized porous silica for improved performance of a glucose biosensor,” Biochem. Eng. J. 91, 78–85 (2014). [CrossRef]

27. H. Ikemoto, Q. Chi, and J. Ulstrup, “Stability and catalytic kinetics of horse radish peroxidase confined in nanoporous SBA-15,” J. Phys. Chem. C 114(39), 1840–1846 (2010). [CrossRef]

28. M. Ferreira, P. A. Fiorito, O. N. Oliveira Jr, and S. I. Córdoba de Torresi, “Enzyme-mediated amperometric biosensors prepared with the Layer-by-Layer (LbL) adsorption technique,” Biosens. Bioelectron. 19(12), 1611–1615 (2004). [CrossRef] [PubMed]

29. J. Livage, T. Coradin, and C. Roux, “Encapsulation of biomolecules in silica gels,” J. Phys. Condens. Matter 13(33), R673–R691 (2001). [CrossRef]

30. N. Balistreri, D. Gaboriaua, C. Jolivalt, and F. Launay, “Covalent immobilization of glucose oxidase on mesocellular silica foams: Characterization and stability towards temperature andorganic solvents,” J. Mol. Catal., B Enzym. 127, 26–33 (2016). [CrossRef]

31. J. Huang, H. Wang, D. Li, W. Zhao, L. Ding, and Y. Han, “A new immobilized glucose oxidase using SiO2 nanoparticles as carrier,” Mater. Sci. Eng. C 31(7), 1374–1378 (2011). [CrossRef]

32. D. S. Bagal, A. Vijayan, R. C. Aiyer, R. N. Karekar, and M. S. Karve, “Fabrication of sucrose biosensor based on single mode planar optical waveguide using co-immobilized plant invertase and GOD,” Biosens. Bioelectron. 22(12), 3072–3079 (2007). [CrossRef] [PubMed]

33. Q. Wang, Z. Yang, Y. Gao, W. Ge, L. Wang, and B. Xu, “Enzymatic hydrogelation to immobilize an enzyme for high activity and stability,” Soft Matter 4(3), 550–553 (2008). [CrossRef]

34. S. Tierney, S. Volden, and B. T. Stokke, “Glucose sensors based on a responsive gel incorporated as a Fabry-Perot cavity on a fiber-optic readout platform,” Biosens. Bioelectron. 24(7), 2034–2039 (2009). [CrossRef] [PubMed]

35. 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. Actuat. B 237, 150–158 (2016). [CrossRef]

36. H. Y. Jung, R. K. Gupta, E. O. Oh, Y. H. Kim, and C. M. Whang, “Vibrational spectroscopic studies of sol-gel derived physical and chemical bonded ORMOSILs,” J. Non-Cryst. Solids 351(5), 372–379 (2005). [CrossRef]

37. R. F. S. Lenza, E. H. M. Nunes, D. C. L. Vasconcelos, and W. L. Vasconcelos, “Preparation of sol-gel silica samples modified with drying control chemical additives,” J. Non-Cryst. Solids 423, 35–40 (2015). [CrossRef]

38. I. Delfino, M. Portaccio, B. Della Ventura, D. G. Mita, and M. Lepore, “Enzyme distribution and secondary structure of sol-gel immobilized glucose oxidase by micro-attenuated total reflection FT-IR spectroscopy,” Mater. Sci. Eng. C 33(1), 304–310 (2013). [CrossRef] [PubMed]

39. 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. Actuat. B 138(1), 344–350 (2009). [CrossRef]

40. Y. L. Luo, L. H. Fan, F. Xu, Y. S. Chen, C. H. Zhang, and Q. B. Wei, “Synthesis and characterization of Fe3O4/PPy/P(MAA-co-AAm) trilayered composite microspheres with electric, magnetic and pH response characteristics,” Mater. Chem. Phys. 120(2-3), 590–597 (2010). [CrossRef]

41. B. Singh and L. Pal, “Sterculia crosslinked PVA and PVA-poly(AAm) hydrogel wound dressings for slow drug delivery: mechanical, mucoadhesive, biocompatible and permeability properties,” J. Mech. Behav. Biomed. Mater. 9, 9–21 (2012). [CrossRef] [PubMed]

42. T. Sheela and Y. A. Nayaka, “Kinetics and thermodynamics of cadmium and lead ions adsorptionon NiO nanoparticles,” Chem. Eng. J. 191, 123–131 (2012). [CrossRef]

43. Y. Liu and P. H. Daum, “Relationship of refractive index to mass density and self-consistency of mixing rules for multicomponent mixtures like ambient aerosols,” J. Aerosol Sci. 39(11), 974–986 (2008). [CrossRef]

44. V. V. Sechenyh, J. C. Legros, and V. Shevtsova, “Experimental and predicted refractive index properties in ternary mixtures of associated liquids,” J. Chem. Thermodyn. 43(11), 1700–1707 (2011). [CrossRef]

45. H. Znad, J. Markos, and V. Bales, “Production of gluconic acid from glucose by Aspergillus niger: growth and non-growth conditions,” Process Biochem. 39(11), 1341–1345 (2004). [CrossRef]

46. C. Y. Lin, H. M. Huang, and H. M. Chen, “Use of backlit light plate to enhance visualization of imidazole-zinc reverse stained gels,” Biotechniques 41(5), 560–564 (2006). [CrossRef] [PubMed]

47. C. Sarici-Ozdemir and Y. Onal, “Equilibrium, kinetic and thermodynamic adsorptions of the environmental pollutant tannic acid onto activated carbon,” Desalination 251(1-3), 146–152 (2010). [CrossRef]

48. A. O. Babatunde and Y. Q. Zhao, “Equilibrium and kinetic analysis of phosphorus adsorption from aqueous solution using waste alum sludge,” J. Hazard. Mater. 184(1-3), 746–752 (2010). [CrossRef] [PubMed]

Investigation for terminal reflection optical fiber SPR glucose sensor and glucose sensitive membrane with immobilized GODs

Abstract

1. Introduction