Photovoltammetry of p-Phenylenediamine Mediated by Hexacyanoferrate Immobilized on CdS-Graphene Nanocomposites

Cd(ClO₄)₂·6H₂O was purchased from Alfa Aesar Chemical Co., Ltd. (Tianjin, China). Poly(diallyldimethylammonium chloride) (PDDA) and Na₂S·9H₂O were obtained from Aladdin Reagent Co., Ltd. (Shanghai, China). Other reagents of analytical grade were obtained from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China) and used without further purification. Distilled water was used throughout the investigation.

CdS quantum dots (QDs) were synthesized by a hydrothermal method.²⁵ Briefly, 1.5 mL of mercaptoacetic acid was added into 100 mL of 0.2 mol·L⁻¹ Cd(ClO₄)₂ solution, and the pH of this solution was adjusted to 10 with 2.0 mol·L⁻¹ NaOH. The solution was transferred to a 250 mL three-necked flask and refluxed under constant passage of high purity nitrogen gas for thirty minutes. Then, 90 mL of 0.2 mol·L⁻¹ Na₂S was rapidly added into this solution. After 4-h reaction, the product was collected by centrifugation, washed several times with ethanol, and dried at 60°C.

Graphene was prepared with graphene oxide according to a previous report.²⁶ Briefly, 50 mg of graphene oxide was dispersed in 50 mL of water with the aid of sonication for 2 hours. Subsequently, 2 mL of 20% PDDA was added into the graphene oxide suspension under vigorous stirring. After 30 min, 2 mL of hydrazine hydride was added to the mixture. The resultant mixture was stirred for another 10 min, and then heated under reflux for 24 h. The product was washed with distilled water for several times by centrifugation and dried at 60°C.

The CdS and graphene (CdS-G) composites were prepared by directly mixing the aqueous solutions of CdS QDs (3 g·L⁻¹) and graphene (3 g·L⁻¹) under ultrasonic agitation. Different weight contents of graphene in CdS-G composites were obtained by varying the mixture ratio of two suspensions. The prepared composites were marked as CdS-x%G, where x represented the weight percentage of graphene in the whole suspension. Generally, CdS-1%G was used except where otherwise indicated.

Prior to modification, the surface of glassy carbon electrode (GCE) was polished to a mirror-like smoothness with 0.05 μm alumina slurry on a damp silk cloth, and then sonicated in ethanol and water. The GCE surface was dried with nitrogen gas and coated with 10 μL 2% PDDA solution containing 0.5 mol·L⁻¹ NaCl. The electrode was dried at 60°C and then rinsed with distilled water to remove redundant PDDA. After being dried with nitrogen gas, the electrode surface was coated with 10 μL of CdS-G suspension. The obtained CdS-G electrode was dried at 60°C for further use.

The immobilization of HCF on electrode was performed by scanning the CdS-G electrode in 0.1 mol·L⁻¹ KCl solution containing 0.3 mmol·L⁻¹ K₄[Fe(CN)₆] with multiple CV cycles, referring to a previously described procedure²⁷ with a little modification. The CV scanning process was carried out between −0.1 V and 0.45 V with a scan rate of 30 mV·s⁻¹ for 20 cycles. Afterwards, the electrode was thoroughly rinsed with water, and dried at room temperature. The obtained HCF/CdS-G/GCE was stored in a refrigerator at 4°C before use.

Scanning electron microscopy (SEM) and energy dispersive spectroscopy (EDS) were obtained on a Nova NanoSEM 450 instrument (FEI, Netherlands) equipped with IE250X-Max50 EDS (Oxford, UK). Transmission electron microscopic (TEM) observation was carried out on a Tecnai G2 20 TEM instrument (FEI, The Netherlands). The UV-visible absorption spectra were measured with a TU-1900 spectrometer (Beijing Purkinje General Instrument Company, China). All electrochemical and PEC experiments were performed on a CHI830C electrochemical working station (Shanghai Chenhua Instrument Company, China) in a conventional three-electrode system. A modified electrode and a platinum wire were employed as the working and auxiliary electrodes, respectively. All potentials were referred to a saturated calomel electrode (SCE). In PEC measurement, a CEL-S500/350 xenon lamp (CEAULIGHT Co., China) with an optical filter (λ > 420 nm) was used as the irradiation source, and the distance between the light source and working electrode surface was 10 cm.

We immobilized HCF on CdS-G electrode by scanning CV for multiple cycles in 0.1 mol·L⁻¹ KCl solution containing 0.3 mmol·L⁻¹ K₄Fe(CN)₆ from −0.1 to 0.45 V at a scan rate of 30 mV·s⁻¹. Fig. 1 shows the CV curves recorded during this immobilization process. As can be seen, a pair of well-defined redox peaks appear at 0.16 V and 0.08 V, which are assigned to the oxidation of [Fe(CN)₆]⁴⁻ in the positive scan and the reduction of [Fe(CN)₆]³⁻ in the reverse scan. Moreover, both the anodic and cathodic peak currents shift with the proceeding of multiple CV scans. Further observation shows that the anodic peak quickly becomes stable while the cathodic peak tends to be stable after 20 CV scans. Accordingly, the HCF/CdS-G electrode prepared by 20 CV scans was used as the photoelectrode in this work.

Zoom In Zoom Out Reset image size

The fabricated HCF/CdS-G electrode was characterized by SEM, and compared with graphene and CdS-G modified electrodes. Fig. 2A shows that graphene displays a typical flake-like structure with wrinkled surface. The morphology of the CdS-G electrode reveals that a great number of CdS QDs are distributed overall electrode surface to form a uniform film with some cracks (Fig. 2B). The size of CdS estimated by TEM is ca. 6 nm (inset of Fig. 2B), consistent with our previous work.¹² Due to the low content of graphene, no wrinkled structure is seen on such a CdS-G hybrid film.²⁵ By contrast, the morphology of the HCF/CdS-G electrode indicates that some additional aggregated particles are grown on the cracks of the CdS-G film (Fig. 2C). Further, we selected different SEM observation regions on HCF/CdS-G to carry out EDS mapping analysis. Interestingly, the EDS mapping results show that the elements of Cd, S and C exist in any regions but Fe can only be found in the region including these aggregated particles (Fig. 2D). Accordingly, we conclude that these aggregated particles are the products of HCF immobilized on CdS-G, which should be the precipitate of Cd₂[Fe(CN)₆] formed between HCF and anodically stripped Cd(II) from CdS during the multiple CV scanning immobilization process.

Zoom In Zoom Out Reset image size

The CV curves of CdS-G electrode before and after immobilization of HCF were recorded in 0.1 mol·L⁻¹ PBS (pH 6.0) in the dark. As can be seen, there is no any peak in the CV curve of CdS-G electrode (curve a of Fig. 3A); whereas a pair of redox peaks appear around 0.2 V in the CV curve of HCF/CdS-G electrode (curve b of Fig. 3A), confirming the successful immobilization of HCF on CdS-G electrode. While the CdS-G electrode is exposed under visible light (curve c of Fig. 3A), the CV shape is not changed but the current is enhanced as compared with the result of the same electrode in the dark. Interestingly, when the HCF/CdS-G electrode is irradiated with light, the CV curve becomes sigmoidal in shape, accompanied by dramatic enhancement in both anodic and cathodic currents (curve d of Figure 3A). This is in accordance with the previous observation for phovoltammetric behavior of HCF solution on CdS-G electrode,¹³ indicating the electron transfer between the immobilized HCF and CdS-G electrode. The appearance of limiting current can be attributed to the participation of photogenerated electrons in the redox reaction of HCF,¹³ demonstrating the high photoelectrocatalytic activity of CdS-G electrode toward Fe^III/Fe^II couple in HCF.

Zoom In Zoom Out Reset image size

The voltammetric responses of PPD on CdS-G and HCF/CdS-G electrodes in the dark were compared. As shown in curve a of Fig. 3B, the CV curve of PPD on CdS-G electrode exhibits a pair of redox peaks, due to the reversible electro-oxidation process of PPD to diamine.^28,29 While a HCF/CdS-G electrode is employed instead of CdS-G electrode, both the oxidation and reduction peak currents are enhanced (curve b of Fig. 3B), attributed to the electrochemical-chemical redox cycling occurred between the HCF and PPD. That is, PPD reacts with the electro-oxidation products of HCF to generate diamine and reduced species of HCF, leading to the realization of redox cycling involving PPD and HCF. In order to reveal the interaction between PPD and oxidized species of HCF, we recorded the UV-Vis absorption spectrum of PPD. As can be seen in curve a of Fig. 4, PPD exhibits two strong absorption peaks around 202.5 nm and 238.5 nm, and a weak absorption peak around 302 nm. The UV-Vis absorption spectrum of K₃[Fe(CN)₆] (namely HCF(III)) has a strong absorption peak around 206 nm and two weak absorption peaks in the UV region, as well as an absorption peak around 424 nm in the visible region (curve b of Fig. 4). After PPD solution is mixed with K₃[Fe(CN)₆], the spectrum displays a new strong absorption peak around 217.5 nm accompanied with a shoulder peak around 203 nm (curve c of Fig. 4), which is obviously different from the addition of the absorbance values (curve d of Fig. 4) of curves a and b. This result confirms the oxidation of PPD by K₃[Fe(CN)₆], similar to that observed for chemical reaction between PPD and β-MnO₂.²⁹ Additionally, the mixed solution of PPD and K₃[Fe(CN)₆] looks dark (inset of Fig. 4) and shows noticeable absorption from 300 to 700 nm. The absorption band at long wavelength is attributed to the absorption of polymers formed by the reaction between two oxidation intermediates of PPD (namely PPD·⁺ and diimine).²⁹

Zoom In Zoom Out Reset image size

Further, we compared the photovoltammetric responses of PPD on CdS-G and HCF/CdS-G electrodes. As shown in curve c of Fig. 3B, the anodic peak current for PPD on CdS-G electrode is dramatically enhanced when irradiated under light, due to the photoelectrocatalytic reaction of PPD. However, the oxidation current increases continuously with potential and does not reach the limiting state because the concentration of PPD is too low to consume the photogenerated carriers.¹² By contrast, the photovoltammetric response of PPD on HCF/CdS-G electrode increases to the limiting state around 0.3 V (curve d of Fig. 3B). The photovoltammetric limiting current (I_PL) on HCF/CdS-G electrode is obviously higher than the current on CdS-G electrode. Moreover, it is also higher than the photovoltammetric limiting current on HCF/CdS-G electrode in the absence of PPD. This can be attributed to the reaction of PPD mediated by the redox active HCF on CdS-G. As illustrated in Scheme 1, electrons and holes are generated on CdS under visible light irradiation. When HCF(II) is electroxidized to HCF(III), photogenerated electrons are shuttled by graphene to reduce HCF(III), and thus forming an electrochemical-photochemical cycle of HCF reaction. The appearance of limiting current means that the amount of immobilized HCF on the photoelectrode surface is large enough to consume photogenerated carriers.¹³ On the other hand, in presence of PPD, electro-oxidized PPD can be reduced to PPD by photogenerated electrons and forms an electrochemical-photochemical cycle of PPD reaction.¹² Moreover, some PPD molecules can be oxidized by HCF(III) to form diimine, which contributes to enhanced anodic limiting current. That is, the immobilized HCF can efficiently consume photogenerated carriers and mediate the reaction of PPD at low concentrations to produce detectible photovoltammetric limiting current.

Scheme 1. Schematic illustration of mediated PEC reaction of PPD on HCF/CdS-G electrode.

In order to achieve desirable photovoltammetric responses of PPD, the conditions for the preparation of HCF/CdS-G electrode such as the content of graphene and the concentration of K₄[Fe(CN)₆] were optimized by comparing the photovoltammetric limiting current difference (ΔI_PL) before and after the addition of 0.01 mmol·L⁻¹ PPD. Fig. 5 shows the photovoltammetric responses of PPD on HCF/CdS-G electrodes fabricated with different contents of graphene. As can be seen, with increasing the content of graphene from 0.5% to 1% in CdS-G composites, the anodic current enhances until reaching the platform. When the content of graphene increases to 3%, the ΔI_PL value representing the photovoltammetric response toward PPD slightly decreases. However, the photovoltammetric curve of HCF/CdS-G electrode is obviously changed and no limiting current is observable when the content of graphene reaches 5%. It is likely that introduction of excessive graphene in CdS-G leads to shielding of CdS to absorb the visible light.^30,31 Therefore, 1% graphene was selected to fabricate the photoelectrode and used throughout this research, except indicated otherwise.

Zoom In Zoom Out Reset image size

On the other hand, the photoelectrode was also optimized by varying the concentration of K₄[Fe(CN)₆] during the electrochemical immobilization process. As shown in Fig. 6, the photovoltammetric response toward PPD represented by ΔI_PL increases with increasing the concentration of K₄[Fe(CN)₆] from 0.1 to 0.3 mmol·L⁻¹. Nevertheless, when the concentration of K₄[Fe(CN)₆] reaches 0.5 or 1.0 mmol·L⁻¹, the ΔI_PL value decreases. This is consistent with the requirement that the amount of HCF immobilized on photocatalyst should be suitable so that a substantial portion of photocatalyst surface is exposed to the electrolyte to generate favorable PEC response.³² Thus, 0.3 mmol·L⁻¹ K₄[Fe(CN)₆] was used as the optimal condition for immobilizing suitable amount of HCF on CdS-G electrode.

Zoom In Zoom Out Reset image size

The linear sweep voltammetric (LSV) curves of PPD at various concentrations were recorded on HCF/CdS-G electrode under visible light irradiation. As illustrated in Fig. 7A, the photovoltammetric response of HCF/CdS-G electrode, which is represented by the photovoltammetric limiting current (I_PV) at 0.3 V after subtracting the background current, increases linearly with increasing the concentration of PPD from 0.1 to 10 μmol·L⁻¹. The linear regression equation can be expressed as I_PV (μA) = 0.129C (μmol·L⁻¹) + 1.316 (linear coefficient R² = 0.998), as depicted in curve a in Fig. 8A. The detection limit (3S/N) is estimated to be 0.047 μmol·L⁻¹. The prepared HCF/CdS-G based sensor shows favorable detection limit and linear range in comparison with many previously reported methods for PPD determination (Table I). Meanwhile, we also recorded the photovoltammetric responses of CdS-G electrode toward PPD at different concentrations from 1 to 10 μmol·L⁻¹ (Fig. 7B). Consistent with our previous observation, the oxidation current increases continuously with potential and are difficult to reach the limiting state because the concentration of PPD is too low to consume the photogenerated carriers.¹² For quantitative comparison, the current at 0.3 V after subtracting the background current is still adopted to evaluate the photovoltammetric response of CdS-G electrode. The result shown in curve b in Fig. 8A indicates that the linear regression equation for PPD on CdS-G electrode can be expressed as I_PV (μA) = 0.111C (μmol·L⁻¹) + 0.320 (R² = 0.997). The detection limit (3S/N) is 0.11 μmol·L⁻¹, more than two times higher than that obtained on HCF/CdS-G electrode, demonstrating that the presence of HCF on the electrode surface provides a better sensing performance.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Moreover, the voltammetric responses of HCF/CdS-G and CdS-G electrodes toward PPD at different concentrations were also recorded in the dark, as illustrated in Figs. 7C and 7D. The oxidation peak currents (I_pa) on the two electrodes are found to be linearly proportional to the concentration of PPD. On HCF/CdS-G electrode, the linear regression equation for PPD is I_pa (μA) = 0.0737C (μmol·L⁻¹) + 0.113 (R² = 0.996) (curve c in Fig. 8A), with a detection limit (3S/N) of 0.12 μmol·L⁻¹. On CdS-G electrode, the linear regression equation for PPD is I_pa (μA) = 0.0638C (μmol·L⁻¹) + 0.0849 (R² = 0.996) (curve d in Fig. 8A), with a detection limit (3S/N) of 0.56 μmol·L⁻¹. These results indicate: (1) HCF can promote the sensing of PPD on CdS-G electrode either in the dark or under photoirradiation; (2) photovoltammetry can provide better sensing performance for PPD on photoelectrodes than ordinary voltammetry in the dark.

The reproducibility of the proposed HCF/CdS-G electrode was estimated by determining the photovoltammetric responses of five independently prepared HCF/CdS-G electrode toward 10 μmol·L⁻¹ PPD. The relative standard deviation (RSD) is 3.3%, showing a good reproducibility of the proposed method. Moreover, the stability of the as-prepared HCF/CdS-G electrode was assessed by the recording the photovoltammetric limiting current after storing at 4°C for four days. The result indicates that the current response maintains 97.2% of the initial value, demonstrating a good stability. Moreover, the selectivity of the HCF/CdS-G electrode was evaluated by measuring the responses of 0.01 mmol·L⁻¹ PPD in the presence of other organic compounds such as 4-chlorophenol (4-CP), 4-chlorophenylamine (4-CA), p-nitrophenol (PNP) and tetrabromobisphenol A (TBBPA) at the same concentration level. As depicted in Fig. 8B, all these compounds do not show obvious interference in the response of PPD, demonstrating that the proposed photovoltammetric method has acceptable selectivity for PPD determination.

The applicability of the proposed HCF/CdS-G electrode was evaluated by photovoltammetric determination of PPD concentration in environmental water samples which were collected from two pools in the campus of Huazhong University of Science and Technology. Prior to analysis, both water samples were filtered through a 0.22 μm membrane. Then, 5.0 mL of water sample was added into the cell containing 5.0 mL of 0.2 mol·L⁻¹ PBS (pH 6.0) to carry out the photovoltammetric determination. However, no PPD was detected in both samples. Then, standard PPD solutions were spiked into the water samples. As shown in Table II, the recoveries of the proposed method for these water samples are in the range from 96.6% to 104.3%, demonstrating the feasibility of the proposed photovoltammetric method for PPD determination in environmental water samples.