Figure 2 shows three SEM images of electrodes Nano A, B and C. In all specimens, the nanopillars had a diameter of about 150 nm. From the side views (see insets in Fig. 2) of the specimens, the height of the nanopillars is estimated to be 1 μm, 2.5 μm and 6 μm for specimen A, B and C, respectively. From these images, we noted that in electrodes with taller nanopillars (e.g., Nano B and Nano C) there are slight bunching deformations in nanopillars. This kind of deformation is caused by the capillary interaction (during the wetting of the electrodes) compounded by the reduced flexure rigidity of these taller nanopillars [12]. Also shown here is the cyclic voltammograms for three NAEs and a flat electrode measured in 0.3 M sulphuric acid. In all these voltammograms, a reduction peak is seen in between 0.70 V and 1.1 V. To quantify the difference in the height of the nanopillars in these NAEs, we defined a roughness ratio as the area under the reduction peak (calculated by integrating the voltammogram from 0.70 V to 1.1 V) of a NAEs electrode divided by that of the flat electrode. The roughness ratio was found to be about 20, 38.8 and 63.4 for specimens A, B and C, respectively (see Table 2).
Figure
3A shows the amperometric current response for the bare electrodes (NAEs and flat) at various K
4Fe(CN)
6 concentrations. In general, all the NAEs exhibited a higher current than the flat electrode at each K
4Fe(CN)
6 concentration. To further quantify the sensing performance of these bare electrodes, we analyzed the relationship between the current response and K
4Fe(CN)
6 concentration by a linear regression analysis. Figure
3B shows the variation of the steady-state amperometric current with the concentration of K
4Fe(CN)
6 (from 4 mM to 24 mM) along with the corresponding regression lines. By taking the slope of the regression lines and normalizing it with respect to the geometrical area of the electrodes (3.2 mm × 3.2 mm), we obtained sensitivity values for the electrodes and these values were listed in Table
2. Clearly, for all the bare electrodes, the NAEs showed sensitivity about two times higher than that of the flat electrode.
One surprising observation, however, was that the sensitivity of these bare NAEs did not increase with the increase of the roughness ratio. This implies that the benefit of the increased surface area due to nanopillars has not been fully realized. It seems that only the top part of the nanopillars has contributed to the increase of active electrode surface for electron transfer, which may explain why there is only a two-fold increase in the current responses of all the NAEs as compared with the flat electrode. We speculate that the electroactive species K4Fe(CN)6 may encounter certain difficulties in its transport to the small spaces between the bare nanopillars as the result of either a low diffusivity or a fast electron transfer rate constant. With a low diffusivity, it would be difficult for K4Fe(CN)6 to diffuse deep into the small spaces between the nanopillars, while with a fast electron transfer rate constant, most of the species K4Fe(CN)6 would get oxidized near the top ends of the nanopillars before it diffuses down the gaps. Under such a circumstance, it is conceivable that only the top regions of the nanopillars are serving their active duty in transferring electrons.
Figure 4A shows the amperometric currents for the functionalized NAEs and flat electrode at various glucose concentrations. Again, all the NAEs exhibited a higher current response than the flat electrode at each glucose concentration. Note that in each incremental step, the current response of Nano C is still rising indicating that it has not reached its steady state. We believe this phenomenon is due to the increased response times for electrodes with taller nanopillars. However, for a quick comparison between these nano electrodes, we took a more conservative approach to get the current readings for Nano C at the same time as for Nano B and Nano C.
Figure 4B shows the variations of the steady-state amperometric current with glucose concentration (from 2.5 mM to 15 mM) along with the corresponding linear regression lines. By taking the slope of the regression lines and normalizing it with respect to the geometric area of the electrode in each case, we obtained the sensitivity measurement for the functionalized electrodes (NAEs and flat). From the obtained sensitivity values listed in Table 2, we observed that unlike in the bare electrode cases, the sensitivity of NAEs increases as the roughness ratio increases. The highest sensitivity value (Nano C) is about 3.13 μA·mM-1·cm-2 (about 12 times higher than that for a flat electrode) which is significantly higher than the value reported for a gold nanotube electrode (0.4 μA·mM-1·cm-2) [8]. So for the functionalized NAEs, increasing the surface roughness of the NAEs does contribute to an increase in detection sensitivity.
Figure 5 shows the variations of the steady-state amperometric current with the glucose concentration over a wider concentration range (2.5 mM to 30 mM). By performing nonlinear curve fitting to the data using Eq.4, we obtained values for Km and Imax in each case as listed in Table 2. Clearly, both Imax and Km values are higher for the NAEs than for the flat electrode and they increase as the roughness ratio increases. Furthermore, the Km values for all the NAEs are larger than the reported intrinsic Km value of 25 mM for dissolved glucose oxidase [17]. This indicates that the activity of the enzyme immobilized on these NAEs has actually been lowered as compared with the freely dissolved enzyme, which further suggests that the increase in sensitivity in the functionalized NAEs is due to factors other than enzyme activity.
In comparing the bare with the functionalized electrodes, we found that the highest nanostructure-induced sensitivity increase for the functionalized electrodes (12 times) is higher than that for the bare electrodes (2 times). This could be due to the difference in electrochemical species involved (i.e., glucose versus K4Fe(CN)6). These two electroactive species, however, have a similar diffusivity value (8 × 10-10m2/s for K4Fe(CN)6 and 7.6 × 10-10m2/s for glucose) [15]. This fact suggests that the difference in the reaction rate constant at the bare and functionalized electrodes may play a more dominate role in affecting the current response. It is also possible that such an increase in the sensitivity of functionalized NAEs is the result of heightened retention of the mediator during glucose detection [11].
To see the influence of the reaction rate constant on the current response of the NAEs, we now turn to the simulation results. Figure 6A shows the simulated amperometric current obtained for a functionalized NAEs and a flat electrode in response to glucose at two different reaction rate constants: 1.5 × 10-5 and 1.5 × 10-7 (m/s). As expected, a higher current response was found for the nanopillar electrode than for the flat electrode (see Table 3). But the nanostructure-induced increase in the current response was affected significantly by the reaction rate constant of glucose. At a rate constant of 1.5 × 10-5 m/s the increase in current due to nanopillars was only 3.26 fold, whereas at a rate constant of 1.5 × 10-7 m/s the increase was 22.26 fold. This fact suggests that at a lower reaction rate constant more glucose will be able to diffuse into the deep space between the nanopillars to get oxidized, thus leading to a higher current response. By contrast, K4Fe(CN)6 has a rate constant of 1.5 × 10-4, and at this rate constant the nanostructure-induced increase in current response is found to be only 1.28 fold (see Table 3). This is so because at such a high reaction rate constant, K4Fe(CN)6 will get oxidized quickly at the top regions of the nanopillars before it can diffuse down to the space between the nanopillars. These arguments were supported by the fact that a higher glucose concentration was found at the bottom of the spaces between nanopillars in the case with a lower reaction rate constant: a concentration of 0.497 mol/m3 and 13.583 mol/m3 was found at the bottom of the spaces between nanopillars when the rate constant is 1.5 × 10-5 m/s and 1.5 × 10-7 m/s, respectively. Figure 6B shows a contour plot for glucose concentration at a rate constant of 1.5 × 10-7m/s, where it is seen that a significant amount of glucose reached to the bottom of the spaces between nanopillars. In the case of K4Fe(CN)6 its concentration is found to be zero at the bottom of the spaces between nanopillars (see Figure 6C).
The results obtained from simulation are consistent with the prediction for the catalytic reactions in porous media based on the effectiveness factor (
η) and thiele modulus (
φ) concepts. The value of
η can be determined by using the following formula
[
18]:
(10)
Here 
and 
are the thiele moduli calculated based on the transverse and longitudinal diffusion times, and they are defined by = 2krp/DG, = 2kL2/rpDG, where rp is the pore radius, L is the pore length, DG is the glucose diffusivity and k is the surface rate constant. For a surface rate constant of 5 × 10-4 m/s, we calculated = 0.131 and = 328.8. Under this condition, η is found to be η ≈ 1/
= 0.055. This low effectiveness factor will hinder the transport of the target species to the spaces between the nanopillars. For a lower surface reaction rate constant of 5 × 10-7 m/s, we calculated = 131.5 × 10-6 and = 0.328, and under this condition η will be close to unity (η ≈ 1). This high effectiveness factor will surely enable more efficient transport of glucose to the functionalized surfaces in between the nanopillars.
The above results clearly indicate that the enhanced current response in glucose sensing with functionalized NAEs can be attributed to the effective mass transport facilitated by the relatively lower reaction rate constant of glucose. This fact suggests that to reap the true benefit of using nanostructured electrodes for enhancing the performance of biosensors, it is necessary to optimize the geometry of the nanopillars (their diameter, spacing and height) in order to accommodate the specific analyte species in terms of its reaction kinetics and mass transport.
Answers to all your Biology Questions
