Visible-Light Driven Conversion of Pollutants into Hydrogen and Electricity Based on a Polyaniline Dynamic Electrode

In this work, PANI was synthesized through the simultaneous reaction of the electropolymerization and the doping with protonic acids. The obtained PANI was green in color (Fig. 2 inset), which pointed out the conducting ES form. The surface structures of PANI characterized by FESEM are shown in Figs. 2a and 2b. The low-magnification image shows a layer of smooth film, which is incorporated with granular precipitates. Further increasing the magnification reveals that the granular precipitates are with a rough surface. The electrochemical polymerization of aniline mainly involved two stages.^31,32 Firstly, aniline cation radicals adsorb primarily on the bare plate and give rise to smooth film composed of PANI macromolecules anchored to the substrates. Subsequently, the secondary nucleation of PANI occurs on the already grown PANI film, resulting in granular surface morphology with the incorporation of PANI precipitates into the continuous layer. The diameter of small granules composing the continuous layer is about 1 μm, leading to high surface areas and favoring good illumination utilization.

Zoom In Zoom Out Reset image size

The FT-IR analysis was carried out to understand the chemical constitution of PANI. As seen in Fig. 3a, PANI has the characteristic peaks at 1560 cm⁻¹ and 1478 cm⁻¹ ascribed to the C=C stretching deformation of quinoid and benzene rings, respectively, 1298 cm⁻¹ and 1240 cm⁻¹ to the C-N and C=N stretching band of an aromatic amine, and 1110 cm⁻¹ and 788 cm⁻¹ to the in-plane and out of plane bending of C-H, respectively.^15,33 The presence of the benzenoid and quinoid units is evidence of the emeraldine form of PANI. Fig. 3b illustrates that the amorphous nature of the sample is confirmed by the XRD spectrum, which shows broad and low intense peaks related to a poorly crystallized PANI. It has a primary characteristic peak at 25.2° attributed to the scattering from the periodicity perpendicular to PANI chains and the one at 20.3° to the alternating distance between layers of polymer chains.^33,34 The UV−vis absorption spectra (Fig. 3c) indicates that PANI can absorb most light with a wavelength ranging from 200 to 800 nm, and the characteristic absorption peaks for PANI appears at 318 nm and 628 nm. The former is due to the π−π^∗ transition between the benzoid units, and the latter is attributed to a charge-transfer excitation-like transition from the highest occupied energy level to the lowest unoccupied energy level.^35,36 The gap is determined by extrapolating the linear portion of the curve (αhν)² to hν axis and the transition is directly allowed, α being the optical absorption coefficient and hν the energy of the incident photon.³⁴ The experimental absorption spectrum for PANI shows a bandgap of 1.63 eV. This was consistent with the previous reports,^34,37,38 which is interpreted as excitations to the polaron band. The above results indicate that the as-prepared PANI can be excited by visible light due to its narrow bandgap energy.

Zoom In Zoom Out Reset image size

The photo-generated electrons during the reaction under UV or visible light irradiation were detected by the chopped linear sweep voltammograms (LSV). The results are shown in Fig. 3d, the current with no illumination (dark current) was relatively low. When illuminated by UV or visible light, the photocurrent was significantly increased, and the current obtained by UV light was higher than that of visible-light irradiation. The photocurrent densities were all further increased with the increasing of applied voltage. This demonstrates that the PANI on Ti plate plays a role of photocatalyst and can be excited by visible light. The stability of the polymer electrode for durable application was investigated by recording the current-time curves for 60 min under UV and visible light, respectively (Fig. 3e). The total current generated was stable and gradually decreased with time, which resulted from the consumption of RBR, or the "fuel". It indicates that the PANI covered Ti plate electrode possess potential to be employed as an photoanode for the degradation of organic pollutants with simultaneous generation of electric power.

The incident-photon-to-current conversion efficiency (IPCE) experiment was used to measure the effectiveness in converting photon incident to photocurrent flowing, when using the PANI covered Ti plate as photoelectrode. The IPCE was defined as a function of wavelength by the equation (IPCE  = 1240 I/λ J_light ), where I, J_light, λ represent the photocurrent density, the incident light power density and the wavelength of the incident light, respectively.^39,40 The wavelength range was 430 nm∼720 nm (with a resolution, Δλ = 5 nm). Fig. 3f shows the IPCE of PANI covered Ti electrode with respect to the wavelength, in which the photocurrent response tendency is consistent with the UV−vis absorption spectra (Fig. 3c). The IPCE curve exhibited the lowest values at the wavelength of about 500 nm and the maximum value obtained at 630 nm was ∼2.75%, which is in line with a previous report.⁴⁰ From the graph, it is confirmed that the absorption of photons takes place in the visible region, and the photocurrent action spectrum resembled well with the absorption spectrum of PANI indicating that photocurrent is generated by the electron transition in the excited PANI molecules.

Parameters optimization was performed according to the conditions described in Experimental section during 6 different current densities, i.e. 0.4, 0.6, 0.7, 0.8, 1.0 and 1.4 mA cm⁻² electrode area (vs. SCE), and 6 different polymerization times, i.e. 5, 10, 20, 30, 40 and 50 min. The optimum current density and duration time were judged by the hydrogen-evolution performance of the constructed photo-driving energy conversion system under UV irradiation with above prepared PANI dynamic electrode as photoanode (Fig. 4). With the increase of current density and the prolongation of polymerization duration, the hydrogen-evolution accumulations all shown a trend of "increase first and then decrease" and the maximum amount was about 300 μmol during the experimental period. Due to the two-stage polymer formation process in the galvanostatic polymerization of PANI, reaction conditions (polymerization time and current density) affect the polymerization velocity and film formation, thus influencing the thickness, homogeneity and morphology of the polymer film on the electrode surface.^36,40 Especially when the polymerization current density increased to 1.4 mA cm⁻² (vs. SCE), the hydrogen-evolution amounts significantly decreased to 165 μmol, which may be caused by the formation of the completely oxidized state. In brief, the polymer electrodes synthesized under different conditions possess different photoactivities, which further manifest as the difference of hydrogen-evolution performance obtained in the energy conversion system. In this sense, the optimal polymerization parameters were 0.7 mA cm⁻² electrode area (vs. SCE) and 40 min.

Zoom In Zoom Out Reset image size

The photoactivity of the as-prepared polymer electrode was assessed under UV and visible light illumination by applying it to degrade model contaminated water of RBR aqueous solution whose initial concentration was 80 mg L⁻¹. Figs. 5a and 5b were the real pictures of the photo-driving energy conversion system equipped with dynamic electrode under UV and visible light irradiation, respectively. In addition, a TiO₂/Ti electrode prepared by a sol-gel and dip-coating method was used to compare and evaluate the photocatalytic performance of the PANI dynamic electrode. As shown in Fig. 5c, when UV light irradiated, the RBR dye was removed rapidly at the early stage of the reaction and about 90% RBR dye was degraded in 60 min for both the PANI electrode and TiO₂ electrode. Besides, for the first 40 min, the RBR dye removal efficiency of PANI dynamic electrode was more than 10% better than that of TiO₂/Ti electrode. Under visible light irradiation with λ>400 nm, the degradation efficiency increased gradually with time and about 40% of RBR was degraded by PANI dynamic electrode during the 180 min reaction. For comparison, the degradation by TiO₂ electrode could be ignored due to that TiO₂ can hardly be excited by visible light.

Zoom In Zoom Out Reset image size

The photodegradation and adsorption of model pollutants can be described by the Langmuir-Hinshelwood kinetic model.^41,42 When initial concentration of the reactant is very small, of around millimolar level, the reactions kinetic can be expressed as: −ln(C/C₀) = k_appt, where k_app is apparent rate constant (min⁻¹), C₀ is the initial concentration of dye and C is the concentration of dye after degradation for t min. The linear correlation of the plots of −ln(C/C₀) versus time suggested that photocatalytic degradation of RBR by PANI dynamic electrode and TiO₂/Ti electrode under different irradiation conditions basically obeyed the first-order kinetic model (Fig. 5c, inset). The apparent rate constants were 0.04076 and 0.04328 min⁻¹ for PANI and TiO₂ under UV irradiation, 0.00328 and 0.00005 min⁻¹ for PANI and TiO₂ under visible irradiation, respectively. These results suggest that compared with TiO₂ electrode, thus prepared PANI dynamic electrode all represent a certain photocatalytic activity under UV and visible irradiation. On one side, under illumination, PANI absorb photons and generate electrons, while some positive carbon radicals in polymer chains are formed. These positive carbon radicals might be reduced by an electron from organic pollutants or water, or react with oxygen to generated active species (hydrogen radical, singlet/triplet oxygen) which are capable of degrading organic pollutants.^43,44 On the other side, the prepared PANI with a narrow bandgap could be excited by visible light, and its granular surface structure not only enhanced the surface area, but also utilized the irradiation light more efficiently due to the multi-reflections on the surface.

In this system, the electrons generated by the photo-excitation of anode were transferred to the cathode surface for the hydrogen evolution through external circuit. Figs. 5d and 5e shown the hydrogen evolution and the current flow as a function of time, respectively. When using PANI dynamic photoanode, the hydrogen accumulation increased progressively with the reaction time but the hydrogen evolution rate decreased gradually. Correspondingly, the current flow decreased with time and ultimately became stable. Under visible light, the hydrogen accumulation quantity and current flow detected at 180 min were about 163.5 μmol and 2.7 mA, respectively, which were about half of that produced under UV irradiation (300.2 μmol and 5.2 mA). These results show that the photo-driving energy conversion system equipped with PANI dynamic electrode possess dual environmental benefits: pollutant can be cleaned and light energy, especially visible-light energy, can be converted into hydrogen and/or electricity. For the TiO₂ electrode, the accumulation of hydrogen increased linearly with time (360.1 μmol at 180 min) and the current flow detected was 7.9 mA and relatively stable irradiated by UV light, while almost no hydrogen and current flow was detected under visible light. As is known, the combination of TiO₂ and UV irradiation possess a strong ability to completely and non-specifically mineralize these organic compounds, and thus the electric flow and the hydrogen generation rate remain relatively stable during the experimental period. Moreover, at the early stage of the reaction under UV illumination, the hydrogen evolution and current flow obtained by PANI dynamic electrode were all greater than that of TiO₂ electrode, which were in accord with the result of degradation. When the removal efficiency reached above 90%, the generation of hydrogen and electric power were still detected. Because of its unique structural features, RBR dye shows different UV-visible absorption spectrograms affected by the acidity-alkalinity of solution (Fig. S1) and then will appear different mechanisms of photo-degradation accordingly. According to previous reports,^29,45 RBR dye is characterized by the presence of one azo group (−N=N−), which is a lower energy bond and easily broken, i.e., the conjugation system is damaged and the RBR dye loses its color. Subsequently, the colorless intermediates contained the aromatic structures were generated and would be further degraded. As shown in Fig. S2, the characteristic peaks at 400–600 nm ascribed to the chromophoric group (−N=N−) of RBR were proved to be degraded firstly. With the experimental time prolonged, the peaks at 250–400 nm were decreased gradually. These results provided the preliminary evidence for the degradation mechanism of RBR dye, which was proved to be in accordance with literatures.^29,45 Even if the initial concentration of RBR dye increased, the dual performance of system were not significantly improved when using PANI dynamic electrode (Fig. S3). In addition, we found that the performance of hydrogen evolution and electricity generation cannot be maintained with time, whether under UV or visible-light irradiation. The following experiments were performed to explore the possible reason.

To evaluate the photocatalytic stability of the prepared electrode, PANI dynamic photoanode was used in this system for several photocatalytic runs under UV irradiation, and each run lasted 180 min. After each run, the system was evacuated and the polymer electrode was rinsed with deionized water, dried in air and used for the subsequent run. The results in Fig. 6a shown that the performance of hydrogen evolution markedly decreased with the increasing number of cycles, from 300.2 μmol to 79.3 μmol. After use, the change of the color on the PANI dynamic electrode was observed from green to purple-brown (Fig. 6c). The possible interpretation is that the anode electrolyte contains NaOH to establish a chemical bias and the PANI dynamic electrode is remained in contact with this alkaline electrolyte, which may caused the structural change during the reaction process because of its simple nonredox doping/dedoping chemistry based on acid/base reactions.^9,10,46 While as for PANI, different oxidation and protonation states present different structures, properties, and colors, etc.³¹ According to the previous reports, re-doping with an acid could convert the PANI chains to the conductive emeraldine salt (ES) form.^47,48 In order to verify this, we immersed the purple-brown polymer electrode in a 1 mol L⁻¹ H₂SO₄ solution again for at least 1 h. After the re-doping process, the color of PANI became green again (Fig. 6c) and the performance of PANI dynamic electrode was regenerated. Nine runs of photocatalytic degradation process were repeated to generate hydrogen and current flow. After each use, the PANI dynamic electrode was doped with H₂SO₄ solution. The results were shown in Fig. 6b, the hydrogen-evolution amounts in 180 min fluctuated between 300 μmol and 325 μmol. Similar hydrogen accumulation efficiency at the same time interval during the repeated process preliminarily confirmed the above inference, that was the dedoping of PANI in the photocatalytic reaction process caused the hydrogen-evolution performance decline. While the re-doping of PANI dynamic electrode after reaction could make the photocatalytic performance be easily regenerated and almost recovered to the initial.

Zoom In Zoom Out Reset image size

Whether the constructed energy conversion system or the photoelectrochemical reactor, electrons generated in the anode compartment were transferred by potential difference to the cathode surface and reacted with protons to generate hydrogen. The required potential difference between anode and cathode could be supplied by chemical bias or a small electrical bias. As for photoelectrochemical reactor, if the applied electrical bias is too high, the electrolysis of water will occur to release hydrogen rapidly, which may inhibit the photocatalytic degradation of organic wastes in anode compartment. At the same time, in order to examine the PANI dynamic electrode life under different voltage conditions, a series of applied forward biases were added to the anodic potential with 0.1 mol L⁻¹ Na₂SO₄ as electrolyte solution. As shown in Fig. 7a, under UV illumination, the accumulation amounts of hydrogen were nearly linear increasing with time and the hydrogen evolution rates increased gradually from 0.262 μmol min⁻¹ to 0.771 μmol min⁻¹ with the increase of applied biases from 0.6 V to 1.4 V. These results shown that the increase of voltage did not have adverse effects on the life and performance of PANI dynamic electrode. Without UV irradiation, there was no hydrogen detected in cathode region even at 1.4 V, demonstrating that when the external potential was lower than 1.4 V, the potential only promoted the transfer of electrons and did not electrolyze water. As a consequence, the applied bias was set at 1.4 V for the subsequent experiment. Considering that in real wastewater, the electrolyte type is different and the electrolyte concentration is relatively low. We investigated the effect of three types of electrolyte with low concentration and the results were shown in Fig. 7b. According to the principle of photoelectrochemical hydrogen evolution,^30,49 there are two factors at work: one is the concentration of H⁺ in the catholyte, and the other is the number of electrons released through photocatalytic oxidation and transferred from anode region. In this sense, it was understandable that the hydrogen evolution rate obtained in H₂SO₄ solution was 1.206 μmol min⁻¹ and much higher than that of Na₂SO₄ solution (0.771 μmol min⁻¹). Nevertheless, when using NaOH as electrolyte, the hydrogen-evolution performance was further enhanced to 1.530 μmol min⁻¹. This is owing to that the alkaline condition is propitious to the degradation of RBR and the formation of hydroxyl radicals, which may more effectively oxidize and degrade the aimed substrates.^29,30,50 It is worth noting that regardless of what electrolyte type, the hydrogen-evolution performance could be maintained at a stable level with duration time and there was no variation observed in the color of PANI dynamic photoanode after use. Above results shown that alkaline solutions with a lower concentration would not cause the dedoping of PANI and the performance decrease of dynamic electrode. As for the PANI dynamic electrode, the existence of hydroxyl ions was all-important for photocatalytic degradation and hydrogen evolution.

Zoom In Zoom Out Reset image size