The present work was carried out in aqueous solutions. Purified water obtained with a Milli-Q (Millipore) water purification system was used as solvent. Dopamine, epinephrine, norepinephrine and ascorbic acid, were obtained from Sigma (St. Louis, MO, USA). All the chemicals from commercial sources were of analytical grade. The solutions were prepared using 0.1 M phosphate buffer (pH 7.4). Before each electrochemical experiment, purified nitrogen gas (N₂, Rivoira Italy) was used to deoxygenate the solutions. Wafers of Si, p-Si (100) of different shape and size (i.e., minimum diameter of 1 × 1 cm²; max. diameter of 5 × 5 cm²) were purchased from Merck.

TiO₂ nanostructured films were deposited by Metal Organic Chemical Vapour Deposition (MOCVD) on Silicon substrates), according to the literature method [13]. After the deposition of nanostructured films, the TiO₂-modified Si plates (1 cm × 1 cm) were used as working electrode, an Ag/AgCl was used as reference electrode, and a Pt wire as a counter electrode, in a three-electrode electrochemical cell [15]. In order to assure that the working electrodes have the same geometrical area (0.008 ± 0.003 cm²), the Si plates coated by the TiO₂ deposit, were fixed by a Teflon ring with a disk diameter of 1 mm. This Teflon ring equipped with the TiO₂-modified Si electrode, was terminated with an electrical contact consisting of a copper wire. This was directly connected to an AUTOLAB PGstat/12 potentiostat/galvanostat (Eco Chemie BV, Utrecht, Netherlands) with a metallic crocodile (banana) clip. In addition, a conventional bare Glassy Carbon (GC) electrode (Model Amel CG/492/2, 2 mm diameter, Milan, Italy) was used as working electrode for comparison. All the electrochemical measurements were carried out by AUTOLAB PGstat/12 potentiostat/galvanostat (Eco Chemie BV, Utrecht, Netherlands). All experiments were carried out at room temperature.

A complete material morphological and electrochemical characterisation, using several probes, and their corresponding apparent kinetic constants, K_app, were described in detail in our previous publications [13–15].

Because the main objective of this investigation was to selectively determine the content of DA in the presence of AA, the electrochemical response of DA and AA binary mixtures at TiO₂-modified Si electrodes has been investigated by the DPV method with an applied potential E(V) ranging from + 100 mV to + 400 mV vs. Ag/AgCl reference electrode. The Differential Pulsed Voltammograms of a binary mixture of 1 μM DA and 1 mM AA in 0.1 phosphate buffer solution at pH 7.4 indicate that the interactions between nanostructured TiO₂ surfaces and AA and DA lead to the resolution of the overlapped voltammetric wave, observed in the case of conventional GC electrode (Figure 1, dashed line), into well defined peaks at + 280 mV and + 380 mV, corresponding to the oxidation of AA and DA, respectively (Figure 1, continuous line).

The shift of the anodic peak potentials (E_pa) and the formation of the described well resolved peaks is due to the different interaction of DA and AA with TiO₂ electrodes, even in the presence of a large amount of AA interferent (i.e., AA concentration ≈ one order of magnitude higher than that of DA in the binary mixture, as in the real biological samples) [12]. The separation between the two peak potentials is 100 mV which is large enough for the selective determination of DA in the presence of AA, and also for the simultaneous determination of DA and AA in their mixtures. An important issue that needs to be taken into consideration is the electrocatalytic oxidation of AA by DA [16–19]. The electrooxidation of DA in the presence of AA results in a homogeneous catalytic oxidation of AA. The oxidized DA, dopamine-o-quinone, is chemically reduced by AA to DA, which can be reoxidized at the electrode surface, thus, the oxidation of AA is affected by DA (and “viceversa”). Therefore, eliminating this catalytic reaction should be necessary for the accurate determination of DA. For this purpose, considering the predominantly electrostatic nature of the size selectivity in porous TiO₂ nanostructured films [13–15], one can take advantage of the difference in the ionization state and in the characteristic charge of the permeating ions (e.g., dopamine and ascorbic acid). Provided a sufficient space charge exists inside the nanostructured channels, a Si electrode coated with a TiO₂ nanostructured layer may behave as if it could be considered simultaneously in the “open” state for dopamine and in the “closed” state for ascorbic acid. This makes possible the selective enhancement of the permeation of positively charged dopamine and retardation of the transport of negatively charged ascorbic acid. All measurements were made at pH 7.4 which represents the physiological value of pH parameter in biological systems. At this pH value, TiO₂ functionalized nanostructured films are negatively charged and, therefore, the diffusion part of the double layer is made up primarily by cations, as required for the separation of positively charged dopamine and negatively charged ascorbic acid [13–15]. The ratio between the dopamine and ascorbic acid signals changes from 1:3 for a native GC bare surface electrodes [12] to 10:1 for a TiO₂ nanostructured-modified Si electrode, which give an overall 20-fold enhancement of the selectivity between these substances.

A more detailed evaluation was performed with respect to the analytical performance of the TiO₂-modified Si electrodes for measuring dopamine, epinephrine, and norepinephine neurotransmitters, which have particular relevance for clinical analysis and biosensor development. The calibrations of these three neurotransmitters at TiO₂-modified Si electrodes were obtained by Differential Pulse Voltammetry (DPV) with a pulse amplitude of 50 mV; a pulse width of 60 ms; a scan rate of 50 mV/s; a pulse interval of 200 ms; and a sampling time of 20 ms; E_i = + 100 mV; and E_f = + 400 mV, with an applied potential E(V) ranging from + 300 mV to + 600 mV (see Figure 2, showing the dopamine detection)

Analytical parameters, such as the linearity range, detection limit, reproducibility, and linear regression equations found for the TiO₂-modified Si electrodes, are reported in Table 1. Overall the results demonstrated the superior behavior of the TiO₂ nanostructured-modified Si electrodes relative to that of bare GCEs used to measure the same neurotransmitters. In particular, the Si electrodes assembled with TiO₂ nanostructured films, were able to detect dopamine, epinephrine, and norepinephrine with detection limits an order of magnitude lower, relative to that found for bare GCEs tested under the same conditions.

The detection limit (LOD) was defined as a signal-to-noise ratio of 3 (S/N) = 3. The sensitivity was determined as the slope of calibration curve (μA μM⁻¹cm⁻²). The response time of these chemical sensors was evaluated as the time necessary to reach the peak current value, from the current background (I, μA) referred level.

Experimental conditions of DPV measurements were described in Experimental section in the text. For comparison with the TiO₂-modified Si electrode data given in Table 1, the parameters obtained for dopamine using the reference system, were as given in Table 2.

A bare Si electrode was not used as reference systems because no electrochemistry was observed for DA, epinephrine, and norepinephrine investigated here under the same experimental conditions of TiO₂-modified Si electrodes, by cyclic voltammetry and DPV methods, consequently a bare GC electrode has been used as reference system.

At a bare GCE, the corresponding values were a detection limit of 3.0 × 10⁻⁷ M, a linearity range of 3.7 × 10⁻⁷ – 1.0 × 10⁻⁴ M, and a linear regression equation Ipa/μA = 1.522 + 0.154 C/μM (correlation coefficient r = 0.99). From these data, the detection limit ranging from to 1.0 × 10⁻⁸ M, reported here for dopamine at these TiO₂-modified Si electrodes is very interesting (one order of magnitude compared to GC bare electrodes, the reference system) compared other electrochemical probes described in literature [11,12,16–19]. The electroanalytical performances observed also for epinephrine, and norepinephrine neurotransmitters resulted better than GC bare electrodes, as reported in Table 2, with a detection limit one order of magnitude lower than the corresponding values calculated for a GC bare reference system.

The electrodes show high stability toward neurotransmitters in the presence of AA, in terms of operational, and storage (Long-Term) stability. The operational stability of the chemical sensors towards all three neurotransmitters was evaluated keeping these sensors continuously working in a solution of dopamine, epinephrine, and norepinephrine (0.01 μM) solutions, containing 0.01 μM AA, in 0.1 M phosphate buffer solution at pH 7.4, and renewing these solutions every morning. Also in this case the signal, continuously recorded by DPV, showed that during the first 6 hours the analytical signal displayed a 1% decrease, and in the next 24 hours only 2%. The recorded oxidation current, at 72 h (corresponding to 3 days) showed only a 3% decrease vs. the original oxidation current values and over 3 days, 17%.

The storage stability of the chemical sensors in solution was evaluated at + 4 °C and a RT. The response to dopamine, epinephrine, and norepinephrine 0.01 μM solutions containing 0.01 μM AA of 3 sensors, stored at + 4 °C when not in use, and tested every 2 weeks for 10 weeks (over 2 months, 70 days) showed an average decrease of 40% of the initial current value. Another three electrodes were stored at room temperature during 4 weeks and tested every 2 days. The response to dopamine, epinephrine, and norepinephrine 0.01 μM solutions containing 0.01 μM AA of the three sensors was found to be 55 ± 10% of the original one at the end of the period.

Figure 3 shows the occasional response obtained for dopamine, epinephrine and norepinephrine (0.01 μM solutions containing 0.01 μmol L⁻¹ AA in 0.1 M phosphate buffer, pH 7.0) of 3 different chemical sensors which, during a period over 2 months, have been routinely tested for dopamine, epinephrine and norepinephrine. After a large number of measurements (300) for each sensor, an average response decrease of 40% was observed (see Figure 3). Again several examples of neutransmitters sensors present in literature could be considered [11,12,16–19], but, to our knowledge, the sensor described in this paper was more stable (over three months vs. 15 days maximum).