PART I
Fabrication of Polypyrrole-based Enzyme Electrode
Conducting polymer fiber has been proposed to have a potential to reach the activity site of the enzyme, transferring electron produced from the oxidase directly to the electrode (Gregg et al., 1990; Heller et al., 1990).
This function could raise the efficiency of the biosensor. In the electrochemical cell, 50 mM pyrrole and 1 mg/mL glucose oxidase mixed well. When electropolymerization took place, we took advantage of pyrrole polymerization to entrap glucose oxidase near the electrode surface. As shown in Figure 1.1, the pyrrole polymerization occurred near the 700 mV as suggested by sharp increasing in the current. The current increased probably because of the polymerization of pyrrole.
Response of the Ppy/GOx Biosensor
The generated polypyrrole-encapsulated glucose oxidase electrode (Ppy/GOx) was further characterized. Using the Ppy/GOx electrode as the working electrode to measure 20 mM glucose at the potential of 500 mV.
The product from the oxidation of glucose was gluconic acid and hydrogen peroxide. Under the potential, hydrogen peroxide was future oxidized and caused an electron transfer to the electrode. The currents that produced from the bioelectrochemical reactions were 67 nA, 25 nA and 47 nA for three successive additions (Figure 1.2). Thus, this result suggested an inconsistence read out and hence is probably the low efficiency of entrapping the glucose oxidase by polypyrrole. Glucose
oxidase may leak out of the polypyrrole matrix during successive reactions, meanwhile free radicals produced during the process of polymerization may attack the glucose oxidase and result in inactivation.
Photopolymerization
One of mild gel entrapments is based on the photo-crosslinking of polyvinyl alcohol substituted with light-sensitive styrylpyridinium groups (PVA-SbQ). PVA-SbQ, photopolymer, was used in this work due to its excellent mechanical property and providing a direct coating process without other redundant steps. The entrapment of bimolecular may influence its structure as a well-organized network structure. Hence, an optimal ratio of bimolecular and PVA-SbQ should be carefully evaluated.
Thus, we prepared different ratio of glucose oxidase to PVA-SbQ in the entrapment process in order to obtain an optimal protocol for the preparation of PVA-SbQ entrapped glucose oxidase electrode. Although, the best response was obtained when the ratio of GOx to PVA-SbQ was 1:15 ratio, the reproducibility was low. The Gox to PVA-SbQ was finally set at 1:12 due to (Ⅰ) the process was reproducible (Ⅱ) the entrapped enzyme can be stored for a longer time and (Ⅲ) a good response can be obtained (Fig 1.3). Notably, the response was much more sensitive than the enzyme electrode fabricated by polypyrrole polymerization.
Electrode Modification by Polypyrrole
Although polypyrrole may not be suitable for the fabrication of the biosensor, a higher response was obtained from the oxidative response of
hydrogen peroxide on the Ppy-modified electrode. It suggested that the electro-signal could be amplified by the polypyrrole. Generally, the eletropolymerization of pyrrole was performed using cyclic voltammetry between 0.0 V and 1.0 V. Theoretically, the current represents the amount of polypyrrole deposited on the electrode, as indicated why the increasing on the cyclic voltammetry of Ppy/GOx electrodes produced by electropolymerization at different potential. The responses of electrodes were 1277 nA, 474 nA and 219 nA for 0-1 V, 0-1.5 V and 0-2 V, respectively (Fig 1.4). The responses become lower when the electropolymerization of pyrrole was profound under wider potential range. This is probably overoxidization of polypyrrole at high potential.
The overoxidation of polypyrrole may incorporate hydroxyl groups at the 4’ position or carbonyl groups at the 3’ position on the polypyrrole chain (Beck et al., 1987). A conformational change may occur and cause the attenuation of the conductivity of polypyrrole. The overoxidation can bring a lot of positive charge on the pyrrole ring so that more time was needed to yield a stable background current (Fig 1.4).
Electropolymerization under Different Ionic Strengths
The initial step of pyrrole electropolymerization is to induce the formation of free radicals, which repetitively couples to touch other and form network like structure. Hence, the presence of electrolyte can promote the release of the electron from pyrrole monomer and make the polymerization easier. The character of the ionic species was important for the property of polypyrrole, such as film morphology, conductivity and mechanical behaviors (David et al., 1995; Mikhail et al., 1997).
Chloride ion has high electron negativity, an ability for an atom in a molecule to attract shared electrons to itself that can bring out more free radicals on the pyrrole monomer. When polypyrrole takes the chloride molecule as the dopant, its conducting ability could be greatly enhanced (Heeger et al., 1986). Thus, different concentrations of potassium chloride were used as the electrolytes in the electropolymerization of pyrrole. As expected, electropolymerization of pyrrole was more efficient, if the ionic strength was higher as indicated by the higher oxidative current, a representative of the pyrrole polymerization, of the resulting polypyrrole-modified CPE (Fig. 1.5). The maximum enhancing effect of KCl was seen at 750 mM. Oxidative potential of hydrogen peroxide was not changed by polypyrroles generated under different conditions. The GOX was then immobilized on the polypyrrole-modified carbon paste electrode (CPE) by PVA-SbQ with a ratio of GOX:PVA-SbQ fixed at 1:12. The fabricated glucose sensors were used for subsequent electrochemical analyses. The result shows that polypyrrole generated in the high ionic strength (>750 mM KCl) could not only improve its conducting property (Fig. 1.5), but also enhance the oxidative response to the hydrogen peroxide (Fig. 1.6).
The glucose biosensors fabricated on the plan CPE or polypyrrole-modified CPE are not quite stable. The reactivity of these types of biosensors to glucose decreased rapidly within 3~5 days (Figs.
1.3 and 1.6). Polypyrrole may not be the proper microenvironment for the glucose oxidase. Alternatively, polypyrrole-modified electrode may lose its intrinsic electroactivity after being oxidized by oxygen or hydrogen peroxidase under the atmosphere (Takeoka et al., 1998).
SEM of the Electrode Surface
The morphology of polypyrrole-modified electrode was monitored under the scanning electronic microscope (SEM). We hope that the result can help us to find out why this electrode has the superior response than bare carbon paste electrode. The result from SEM shows that the bare carbon paste electrode showed a lot of colloidal materials and carbon slices (Fig 1.7A). After modified with polypyrrole, a thick, smooth polypyrrole layer covered the surface of electrode could be seen (Fig 1.7B). The surface area of carbon-paste electrode can be enlarged by these polymers and hence increase the conductivity of modified electrode. It is consistent with the previous observation (Heeger et al., 1980 and references therein) that ionic concentration would increase the porosity of the polypyrrole on the electrode so that surface area could increase (Fig 1.8).
Reusability of the Polypyrrole-Modified Electrode
The glucose biosensor fabricated on the polypyrrole-modified electrodes was used as the working electrode to perform their reusability.
As shown in Fig. 1.9, the response of the constructed glucose biosensor attenuated gradually along with the numbers of operation. It has been suggested that polypyrrole could be damaged by the oxidation of hydrogen peroxide as well as oxygen species. Hence, it is possible that the backbone of the polypyrrole was deteriorated by the hydrogen peroxide produced from the reaction of glucose oxidase leading to the reduction of the conductivity and the oxidative responses of pyrrole polymers (Belanger et al., 1989). The adhered polypyrrole might ware off the surface of electrode, so that the surface area of polypyrrole and the
conductivity of the polypyrrole-modified electrode was reduced compared with the untreated one (Fig. 1.10).
In conclusion, we have fabricated a glucose biosensor using a polypyrrole-modified CPE. Several properties of this electrode, such as the morphology, conducting ability and sensitivity for the substrate, were studied. It is demonstrated that electropolymerization of pyrrole can be enhanced by high ionic strength. The modification of carbon-paste electrode by polypyrrole can increase the surface area and hence the conductivity of the electrode. Notably, the integrity of the architecture of polypyrrole may be deteriorated by the oxygen and hydrogen peroxide.
Generally, oxidase enzymes incorporate oxygen to take place oxidative reaction and then subsequent product was H2O2. Using the pyrrole derivatives to improve its application in the fabrication of biosensors may solve this problem. Alternatively, mediators can be used to replace oxygen. Mediators can react with oxidases so that subsequent product of the oxidases would be the reductive mediator instead of H2O2. The advantage of using mediators is to avoid the deterioration of the polypyrrole layer. Another interesting strategy is using dehydrogenases instead of oxidases for the fabrication of biosensors. Dehydrogenases can carry out the reoxidative reaction without the formation of H2O2.
Part II
Palladium (Pd), a silver-white transition metal, is commonly used in the fabrication of electrochemical type electrode due to its ability to catalyze the oxidation of H2O2. In this work we fabricated glucose biosensors by immobilizing glucose oxidase on the Pd electrodes. Various thicknesses of Pd thin films (50, 100, and 150 nm) were deposited on the carbon-paste electrode (CPE) by sputtering and were designated as PdS50-CPE, PdS100-CPE and PdS150-CPE, respectively.
Characterization of Pd-modified Carbon Paste Electrode
The redox potential of PdS-CPE was first characterized by cyclic voltammetry (CV). In order to avoid over-oxidation at high anodic applied voltages, the potential employed in this experiment in the pH 7.0 phosphorus buffer was set below 800 mV. The CV analysis was carried out with a scanning range between –0.8 and 0.8 V for PdS-CPEs and bare CPE (Fig 2.1). The cyclic voltammogram of PdS50-CPEs was similar to that of bare CPE, suggesting that the Pd layer on the CPE did not have an integral property (Chang et al., 2003). During sputtering, Pd layer deposited on the solid substrate changes gradually from many isolated Pd islands to a discontinuous thin film and finally into a continuous thin film.
It is well established that the resistivity of metallic films increases with reducing thickness (Kawamura et al., 2000). As shown in Fig. 1.7, the surface of CPE is rough and consists of a lot of cavities. Hence, deposition of 50 nm thick Pd thin films might not be enough to cover all the surface of the CPE and produce the discontinuous films like isolated islands on the electrode.
The cyclic voltammograms of PdS100-CPE and PdS150-CPE exhibit distinct patterns, which were similar to that of Pt electrode having pronounced cathodic arms associated with the reduction of oxides on the surface (Hall et al., 1998). It has been shown that an oxidative current of palladium (0) to palladium (II) could be observed with potential above 450 mV (Lubert et al., 2001). PdS150-CPE showed lower current density than PdS100-CPE may be owing to its smaller surface area (Fig. 2.1). Most of the cavities on the carbon paste electrode may be filled by the Pd layer of 150 nm thick and results in a smooth surface and small surface area.
Roughness of the electrode surface can provide more active area for the interaction of substrates than the one with smooth surface. Among three PdS-CPEs, PdS100-CPE exhibits a highest conductivity and catalytic ability to H2O2. As a result, PdS100-CPE was used for the fabrication of glucose biosensor.
Catalytic Property of Palladium for Hydrogen Peroxide
Hydrogen peroxide can be oxidized on the carbon paste electrode with a potential above 700 mV (Fig. 2.2). This result is consistent with the normal oxidative potential that catalyzes the conversion of hydrogen peroxide to O2 on a electrode without any catalyzing activity. Under such a high oxidative potential, however, several interference may be observed from the oxidative responses of ascorbic acid, urea and acetaminophen in the serum. Several strategies were used to solve this problem, such as using mediators to decrease working potential or covering by a selective membrane to hinder the interfering substances outside the reaction area (Karyakin et al., 1999; Vaidya et al., 1995). The usage of catalytic
metallic film, such as Pd and Pt, as the electrode is another way to reduce the working potential.
As mentioned in Section IV-2-1, the anodic potential was applied to facilitate reaction (4) (Section IV-2-1) to take place rapidly so as to produce oxidative current. Our studies suggest that a potential as low as 200 mV is sufficient to trigger the electron transfer from H2O2 to the electrode. The oxidative current markedly increased at 200 mV, while that at 100 mV was about 10 times less. The phenomenon might result from the change of the composition of complex on the electrode, that is, Pd(OH)2·H2O2 increased drastically and the change of equilibrium constants, that is, K3 was larger than K2 (Hall et al., 1998). With sufficient anodic potential, Pd(0) would be oxidized to Pd(II) and subsequently turned to Pd(OH)2 with H2O. Pd(OH)2 could offer bonding site for H2O2 to allow subsequent reaction to occur. When K3 larger than K2, it means reaction (3) become rate-limiting step so that reaction (4) carried out as soon as reaction (3) took place. As the result, the oxidative current could raise drastically. Therefore, palladinized electrode allows a substantial decrease in applied working potential and amplified oxidative current from the oxidation of H2O2. In order to increase detecting sensitivity and lower interference noise, we chose the 500 mV as working potential. To accord with our data, all operating potential was set at 500 mV.
Fabrication and Characterization of PVA-SbQ/Gox/Pd Electrode Glucose oxidase was encapsulated on various PdS-CPEs using PVA-SbQ, PVA-SbQ/GOx/PdS50-CPE, PVA-SbQ/GOx/PdS100-CPE and PVA-SbQ/GOx/PdS150-CPE, as described previously. Fig. 2.2 shows the
responses of fabricated glucose biosensors in the presence of 1 mM glucose. The oxidative currents were 28 nA, 3100 nA and 234 nA for
PVA-SbQ/GOx/PdS50-CPE, PVA-SbQ/GOx/PdS100-CPE and PVA-SbQ/GOx/PdS150-CPE, respectively. As expected, PVA-SbQ/GOx/PdS100-CPE showed the best response among three glucose biosensors. Notably, the balanced background current of the glucose biosensor decreased from the scale of 10-7 A to 10-9 A after several runs of performance (data not shown). Background current acts like the threshold. The response of biosensor must be at least equivalent or even superior to the background current or the signal would not be clear enough to be detected. In this aspect, Pd electrode was more suitable than polypyrrole electrode for the fabricating the biosensors.
Electrodeposition of Palladium on the Carbon-paste Electrode
Besides the sputtering, we tried another method, electrodeposition by cyclic voltammetry, to modify the surface of the CPEs (Lubert et al., 2001; Chang et al., 2003). Fig. 2.4 indicates the Pd (II) started to reduce under the – 0.1 V and reductive potential was located at – 0.3 V. In a pH 7.0 solution, chloropalladate can be adsorbed on the carbon paste electrode surface in a slow reaction during potentiostatic treatment. It means that chloropalladate species could also be accumulated on the electrode surface during cyclic voltammetry. Therefore, the increase of the peak current at – 0.3 V could be attributed to the accumulation of the Pd (0) deposition during cycling. Three Pd-electrodeposited CPEs, Pde10-CPE, Pde15-CPE, and Pde20-CPE, were generated by cyclic scanning in the presence of 5 mM palladium chloride at pH 7.0 between -0.4 V and
1.0 V for 10, 15, and 20 cycles, respectively.
Subsequently, glucose oxidase was immobilized on the palladinized CPE using PVA-SbQ, termed PVA-SbQ/GOx/Pde10-CPE, PVA-SbQ/GOx/Pde15-CPE and PVA-SbQ/GOx/Pde20-CPE as described previously. The sensing responses of these glucose biosensors were characterized by subjecting to electrochemical analysis. The result showed that the sensing sensitivity of all three PVA-SbQ/GOx/Pde-CPEs was hundreds times higher than that of glucose biosensor fabricated on the unmodified CPE (Fig. 2.5). Notable, all three PVA-SbQ/GOx/Pde-CPEs exhibit a similar sensing response to that of PVA-SbQ/GOx/PdS100-CPE, although PVA-SbQ/GOx/Pde15-CPE is the best one among three similar types of glucose biosensors. It is postulated that the thickness of Pd during electrodeposition was affected by the scanning cycles. Although modify CPE by electrodeposition is easy, quick and less expensive, this method is suffered from the inconsistency due to the unknown reason. We could not obtain the consistent response for each preparation of Pd-electrodeposited CPE. It seems that several experiment parameters, including potentiostatic treatment, the type of palladate and electrolyte, need to be improved in the future studies (Lubert et al., 2001). Consequently, we still took advantage of sputtering technology to perform subsequent experiments.
Sensing Dynamic Range of PVA-SbQ/GOx/PdS100-CPE biosensor
Fig. 2.6 showed the current-time response curve of PVA-SbQ/GOx/PdS100-CPE biosensor by step-wise adding 10 µM glucose.
A fast, sensitive and uniform response to each addition was observed in
an additive manner was observed (Fig. 2.6). The result suggests that PVA-SbQ/GOx/PdS100-CPE biosensor can be used for a continuous performance in the medical application. Furthermore, this result also demonstrated that PVA-SbQ could allow glucose to pass through and reach glucose oxidase. The electron transfer to the electrode was also not obstructed by this photosensitive emulsion. The PVA-SbQ/GOx/PdS100-CPE biosensor exhibits a linear dynamic range for glucose from 0.5 µM to 1000 µM, with a slope of 4.8 nA/µM glucose and a correlation coefficient of 0.997 (Fig. 2.7). The detection limit of PVA-SbQ/GOx/PdS100-CPE biosensor for glucose can be as low as 0.5 μ M.
Reproducibility and Stability of PVA-SbQ/GOx/PdS100-CPE biosensor The availability of PVA-SbQ/GOx/PdS100-CPE biosensor for repeating measurements were performed using 10μM and 1 mM glucose as working conditions. The PVA-SbQ/GOx/PdS100-CPE biosensor exhibited a stable response to 10 µM or 1 mM glucose even after 10-20 consecutive reactions. The relative standard derivation for reproducibility of PVA-SbQ/GOx/PdS100-CPE biosensor in response to 10 µM and 1 mM glucoses were 4.21 % and 8.62 %, respectively. This result demonstrates that the fabricated glucose biosensor can be repeatedly used at least 10-20 times without losing the encapsulated enzyme as well as the integrity of the Pd-modified electrode and the PVA-SbQ membrane. This observation indicates that PVA-SbQ/GOx/PdS100-CPE biosensor has a good operational stability for multiple-use or continuous analysis. Furthermore,
this result revealed that PVA-SbQ/GOx/PdS100-CPE biosensor was quite reliable when detecting low concentration of substrate, but with moderate reliability when detecting high concentration of substrate. According to Fick’s law, the redox current (at any time) is proportional to the concentration gradient of the electroactive species from bulk solution to the electrode. The reaction chamber used in this study is relative large (10 mL). With the addition of high concentration of substrate into the reaction solution with stir, the concentration gradient between bulk and reaction layer becomes complicated. So, the unreliable result could occur when detecting high concentration of substrates. This problem can be improved by designing a small volume chamber, such as microfabrication, or using continuous reaction system, such as FIA.
The storage stability of PVA-SbQ/GOx/PdS100-CPE biosensor was also investigated. The PVA-SbQ/GOx/PdS100-CPE biosensor was stored at 4 ℃ in the intervals between the consecutive reactions. The activity assay of PVA-SbQ/GOx/PdS100-CPE biosensor in the presence of 100 µM glucose after long-term storage was shown in Fig. 2.9. The response of the PVA-SbQ/GOx/PdS100-CPE biosensor to glucose increased gradually during the initial days and elevated abruptly after 21 days storage. This may be due to the swelling of PVA-SbQ membrane during storage and reached to the optimal structure after 21 days (Fig. 2.9). Similar result was also observed elsewhere (Chang et al. 2003). The response of fabricated glucose biosensor remained stable for at least 3 months. After
The storage stability of PVA-SbQ/GOx/PdS100-CPE biosensor was also investigated. The PVA-SbQ/GOx/PdS100-CPE biosensor was stored at 4 ℃ in the intervals between the consecutive reactions. The activity assay of PVA-SbQ/GOx/PdS100-CPE biosensor in the presence of 100 µM glucose after long-term storage was shown in Fig. 2.9. The response of the PVA-SbQ/GOx/PdS100-CPE biosensor to glucose increased gradually during the initial days and elevated abruptly after 21 days storage. This may be due to the swelling of PVA-SbQ membrane during storage and reached to the optimal structure after 21 days (Fig. 2.9). Similar result was also observed elsewhere (Chang et al. 2003). The response of fabricated glucose biosensor remained stable for at least 3 months. After