The nano-gold layer on platinum rotating disk electrode (nano-Au/Pt-RDE) is prepared by the potentiostat operating in DC mode and the typical three-electrode cell. The Pt-RDE was polished with the water slurry of 0.05 µm alumina particles. The residual polishing material was removed using ethanol and then water in an ultrasonic bath. After conditioning in 0.25 M H₂SO₄ solution at 400 mV for 10 min, the Pt-RDE was immersed into 3 mM HAuCl₄ solution and rotated at 400 rpm. Deposition was conducted at the applied potential of −200 mV (vs. Ag/AgCl, 3 M KCl) for 3 min. Finally, the nano‑Au/Pt-RDE was washed with water and the surface of working electrode was activated in 0.15 M NaF buffer solution pH 7.8 by the linear potential sweep with potential range from 400 to −200 mV and scan rate of 10 mV∙s⁻¹ until the voltammetric response was reproducible.

The nano-Au/Pt-RDE was rotated at 600 rpm into the 30 mM PET solution for 10 min. The PET-modified nano- Au/RDE (PET/nano-Au/Pt-RDE) were conditioned in 0.1 M HClO₄ solution by electrocycling with potential range from 400 to −200 mV and scan rate of 10 mV∙s⁻¹ until the voltammetric response was reproducible.

The PET/nano-Au/Pt-RDE was first put in 10 mL of sample or sample spiked Cr(VI) pH 4.5 consisting of 0.15 M NaF and rotated at 400 rpm under open-circuit potential to accumulate Cr(VI) for 2 min. After the accumulation, the determination of Cr(VI) was carried out by transferring the electrodes to 0.15 M NaF buffer solution pH 7.8 without Cr(VI) and performing differential pulse adsorptive cathodic stripping voltammetry (DPAdCSV) from 400 to −100 mV with scan rate of 50 mV∙s⁻¹ (potential step of 5 mV and time step of 0.1 s) and amplitude of 50 mV under non-stirring. The electrode surface was regenerated between measurements by electrocycling in 0.1 M HClO₄ for eliminating Cr(VI) from the electrode surface.

Coastal water sample was filtered by 0.45 µm membrane filter and diluted at least 5000 times by 0.15 M NaF buffer solution pH 4.5.

The electrodeposition of gold nano-flakes (nano-Au) on platinum surface from HAuCl₄ solution depends on the HAuCl₄ concentration(C_Au(III)), the rotating speed of electrode (V_R), the deposition potential (E_P), the deposition time (t_P) and temperature (T_P). Therefore, C_Au(III), V_R, E_P and t_P were investigated while the T_P was kept constant at 25˚C.

The activity of the nano-Au/Pt-RDE deposited was evaluated by the peak current (I_P) obtained when electrode potential was swept from 400 to −100 mV with scan rate of 5 mV∙s⁻¹ in 0.15 M NaF buffer solution pH 7.8 consisting of 100 mg∙L⁻¹ Cr(VI). Each electrodeposition condition of nano-Au was repeated three times to evaluate the reproducibility of electrodeposition through standard deviation (SD) of I_P.

The E_P conducted the activity of the Pt-RDE for the reduction of HAuCl₄ into nano-Au. The highest I_P was obtained with E_P of −200 mV (Figure 1). When E_P was more negative, the reduction of hydrogen ion affected the deposition of the nano-Au, which caused rapid diminution of I_P.

The V_R governed the transferring rate of HAuCl₄ to electrode surface. Thus, the thickness of deposited layer increased with V_R. Increasing the thickness made nano-Au layer rougher and the electrochemically active surface of the obtained nano-Au/Pt-RDE larger [27] . Consequently, the I_P increased. The highest I_P was obtained with V_R of 400 rpm (Figure 2). When the thickness was too high, nano-Au layer could lose the catalytic property inheriting from platinum surface [28] , which caused rapid diminution of I_P.

The thickness of nano-gold layer increased with the increase of C_Au(III) and t_P. Figure 3 and Figure 4 showed that the SD values of I_P rapidly increased at high values of the C_Au(III) and t_P. This might be ascribed to the rapid changing from platinum to gold property when the thickness of nano-Au layer was high enough. Figure 3 indicated that an optimal C_Au(III) was 3 mM. Similarly, an optimal t_P was 3 min (Figure 4).

Figure 5 depicts the voltammetric responses of Au-RDE (a), nano-Au/GC-RDE (b), nano-Au/Pt-RDE (c) and Pt-RDE (d) obtained for 100 mg∙L⁻¹ Cr(VI) in 0.15 M NaF buffer pH 7.8. The NaF buffer is used as electrolyte because it is suitable for all of the electrodes studied [3] [16] - [19] . The Au-RDE showed the lowest peak current while the Pt‑RDE exhibited the highest peak current for Cr(VI) reduction. However, the Pt-RDE also gave a high wave of hydrogen ion reduction near the wave of Cr(VI) reduction. The hydrogen wave could interfere to Cr(VI) determination at trace level.

The peak current of the nano‑Au/GC‑RDE is greatly higher than that of the Au‑RDE. This phenomenon was also observed in the study of Chen et al. [25] . The considerable enhancement of the peak current was usually attributed to the high surface area of the nano-Au layer [25] [29] and the advantageous diffusion of Cr(VI) to the nano-Au layer [25] [30] because the surface of the nano-Au/GC-RDE is rougher than the surface of Au-RDE (Figure 6(A)). In addition, the peak potential for Cr(VI) reduction on the nano-Au/GC-RDE was found at 80 mV, approximating to the potential observed on the gold nanoparticle-modified electrode of the previous studies [3] [25] .

It is interesting that the peak potential for Cr(VI) reduction on the nano-Au/Pt-RDE and the Pt-RDE was found at 160 mV and it was 80 mV more positive than that of the nano-Au/GC-RDE. The positive shifting of Cr(VI) reductive potential showed that Cr(VI) reduction on the nano-Au/Pt-RDE and the Pt-RDE was easier than that on the nano-Au/GC-RDE. On the other hand, the peak current of the nano-Au/Pt-RDE was significantly higher than the one of the nano-Au/GC‑RDE while the surface of the nano-Au/Pt-RDE (Figure 6(B)) was less rugged than the nano-Au/GC-RDE (Figure 6(A)). Therefore, the activity for Cr(VI) reduction of the nano-Au/Pt-RDE was higher than the activity of the nano-Au/GC-RDE.

The peak current of the nano-Au/Pt-RDE was similar to the one of the Pt-RDE while the hydrogen wave did not appear on the nano-Au/Pt-RDE. This indicated that the nano-Au layer on the nano-Au/Pt-RDE covered nearly whole of platinum surface of Pt-RDE and copied the catalysis of platinum surface for Cr(VI) reduction. Thus, the nano-Au/Pt-RDE possessed the catalysis of platinum surface and the advantages of gold nanoparticle such as insensitiveness with hydrogen ion and convenience for assembling with PET to accumulate Cr(VI). Therefore, the nano-Au/Pt-RDE was better than Pt-RDE and nano-Au/GC-RDE.

According to Ouyang et al. [3] , the more the number of the pyridium groups on working electrode were, the more quickly and thoroughly the accumulation of Cr(VI) took place. The nano-Au/Pt-RDE possessed a higher surface area than the Au-RDE did. Thus, the PET concentration used was 30 mM, which was 6-fold higher than that of the previously published reports [3] [19] . For optimizing nano-Au layer assembling with the PET molecules, immersion time of electrode (t_I) in PET solution was surveyed with the PET concentration of 30 mM and the electrode rotation speed of 600 rpm. The quality of the PET/nano-Au/Pt-RDE was evaluated by the amount of Cr(VI) was accumulated into the obtained electrode from 0.15 NaF buffer solution pH 4.5 consisting of 10 ng∙L⁻¹ Cr(VI). The amount of Cr(VI) accumulated gave the Cr(VI) reductive current peak when the electrode potential was swept from 400 mV to ?100 mV in NaF buffer pH 7.8 by DPAdCSV. Figure 7 showed that the optimum of t_I was 10 min.

For direct electroreduction of Cr(VI) in 0.15 M NaF buffer pH 7.8, the PET/nano-Au/Pt-RDE showed the peak current that is 30% lower than that of the nano-Au/Pt-RDE. This result indicated that the PET molecules have been assembled on the nano-Au/Pt-RDE and they could make the electrochemically active surface of the PET/ nano-Au/Pt-RDR smaller than the surface of nano-Au/Pt-RDE.

Figure 8(A) showed the differential pulse adsorptive cathodic stripping voltammogram of the PET/nano-Au/ Pt-RDE in 0.15 M NaF buffer pH 7.8 after preconcentrating Cr(VI) from 0.15 M NaF buffer pH 4.5 consisting of Cr(VI) in the concentration range from 2.5 to 40 ng∙L⁻¹. The PET/nano-Au/Pt-RDE showed well-defined peaks at ~240 mV vs Ag/AgCl, 3 M KCl and the peak current proportionally increased with the concentration of Cr(VI) in the NaF buffer pH 4.5. Thus, this peak was attributed to the electrochemical reduction of Cr(VI) to Cr(III). This indicated that the PET/nano-Au/Pt-RDE was able to preconcentrate Cr(VI) from the NaF buffer pH 4.5 into it.

Using the accumulation time of 2 min, the linear response for Cr(VI) was observed over the concentration range from 2.5 to 40 ng∙L⁻¹ with a correlation coefficient of 0.999 (Figure 8(B)) when the peak current was used as a function of the Cr(VI) concentration in the NaF buffer pH 4.5. The detection limit was of 1.09 ppt based on three times the standard deviation of the baseline [24] . Although the accumulation time was shorter, the PET/nano-Au/Pt-RDE still showed the LOD lower than the LOD of previous studies [2] - [6] [10] - [20] .

The reproducibility of the PET/nano-Au/Pt-RDE was estimated from the DPAdCSV peak current of 25 ng∙L⁻¹ Cr(VI) in NaF buffer solution pH 4.5. Results in Figure 9 showed that the peak currents were in the range from 95% to 105% of the average value with relative standard deviation (RSD) of 3.4% when the accumulating and stripping Cr(VI) were repeated 30 times on the same PET/nano-Au/Pt-RDE. This reproducibility showed that the PET/nano-Au/Pt-RDE is suitable for standard addition method.

Beside good sensitivity and reproducibility, the PET/nano-Au/Pt-RDE also showed a high selectivity for the accumulating and stripping Cr(VI). The NaF buffer solutions (pH 4.5) consisting of 20 ng∙L⁻¹ Cr(VI) and 20 mg∙L⁻¹ investigated ions exhibited the DPAdCSV peak currents which were in the range from 95% to 105% of the peak current of the NaF buffer solution (pH 4.5) consisting of 20 ng∙L⁻¹ Cr(VI) (Figure 10). Therefore, the ions such as Cr³⁺, Fe³⁺, Cu²⁺, Mg²⁺, Ca²⁺, K⁺, Na⁺, , , and Cl⁻ could not interfere to DPAdCSV of Cr(VI) on PET/nano-Au/Pt-RDE even though their contents were million-fold higher than that of Cr(VI).

The analytical reliability and potential practical applications of new electrode (PET/nano-Au/Pt-RDE) were confirmed by measuring Cr(VI) in real samples. Coastal water samples containing low content of Cr(VI) and high content of other ions are suitable to evaluate the performance of new electrode. After eliminating suspended impurities, the coastal water samples were diluted 5000 times by 0.15 M NaF buffer pH 4.5. The data of Cr(VI) analysis on the PET/nano-Au/Pt-RDE was shown in Figure 11. The slope of the standard addition curve of Cr(VI) on the coastal water sample (Figure 11(B)) was very close to the slope of the standard curve of Cr(VI) (Figure 8(B)). This showed that the sample background did not interfere to Cr(VI) determination on the PET/ nano-Au/Pt-RDE. The relative standard deviation (RSD) of the result was of 6.24%.