Preparation and Characterization of New Modified System: Polyaniline/1.5-Naphtalene Disulfonic Acid as Novel Photocatalyst in the H 2 Production

The semiconducting properties of the system Polyaniline / 1.5-naphtalene disulfonic acid are investigated for the first time to assess its photocatalytic performance for the hydrogen evolution under visible light irradiation. PANI is thermally stable up to ~ 300 °C, above which a weight loss of ~ 1.2% occurs. The X-ray diffraction pattern shows broad peaks with a particle size of ~ 7 nm, leading to an active surface area of ∼ 400 m 2 g −1 . A direct optical transition at 1.96 eV, is determined from the diffuse reflectance spectrum. The electrical conductivity of PANI-NDSA follows an exponential law with activation energy of 0.24 eV. The p -type conduction of PANI-NDSA is evidenced from the (capacitance -2 – potential) characteristic plot; a flat band potential (E fb ) of 0.82 V SCE and a holes density (N A ) of 8.43× 10 24 m -3 are determined in neutral solution (Na 2 SO 4 0.1 M). The electrochemical impedance spectroscopy, measured over an extended frequency domain (1 mHz - 10 10 Hz), indicates the contribution of both the bulk and grain boundaries with a constant phase element (CPE). As application, PANI-NDSA is successfully tested for the hydrogen production upon visible light owing to the potential of its conduction band (-0.75 V SCE ), less cathodic than that of H 2 O/H 2 (~ -0.30 V SCE ). H 2 liberation rate of 3840 h -1 (g catalyst) -1 and a quantum efficiency of 0.34% under full light (29 mW cm -2 ) are obtained using Fe(CN) 64- as reducing agent. The photoactivity is completely restored during the second cycle.


Introduction
Fossil energy resources start to run out and the solar energy can partially take over. It becomes more and more attractive and meets an increasing demand in many fields like the photovoltaic [1,2] thermal conversion [3,4] photocatalysis [5,6] and environmental protection [7][8][9]. With a mean insulation of 1200 W m -2 and a duration exceeding 3000 h year -1 (South Algeria), the sun is an inexhaustible energetic source that represents a promising alternative.
However, the intermittency of this energy requires long-term storage and appropriate strategies are developed in this direction. In this regard, hydrogen is a clean fuel with no emission of the greenhouse gas CO2, the first responsible of the global warming [10,11]. Accordingly, it remains attractive for the chemical storage [12][13][14] with a high energy density per unit volume; it can be stored during the off hours and produced on demand. It can also been used in fuel cells that convert the mixture (H2/O2) into electricity more efficiently and cleanly than in combustion engines.
The water splitting into hydrogen and oxygen can occur thermally [15][16][17], by thermo chemical cycles [18,19], steam reforming [20], zinc acid [21][22][23], aluminium upon treatment with bases and by photocatalysis [24,25]. In the first case, only a small amount of water gas can be decomposed at high temperature (> 2200 °C) using solar concentrators. Additionally, the separation of gases (O2 and H2) renders the process complex and the gases mixture must be quenched to prevent their recombination. By contrast, the water photo-splitting occurs under soft conditions with no special equipment. However, the semiconductor (SC) oxides like SnO2 [26,27], NaNbO3 [28,29] and TiO2 used in photocatalysis [30,31], are stable against photocorrosion but have a large forbidden band (Eg> 3 eV) and are consequently less attractive for the solar energy conversion where less than 5% of UV light accounts for the solar radiation [32,33]. Therefore, some recent studies were directed toward organic semiconductors (SCs) [34,35]. In this respect, the conducting polymers [36] start to arouse a growing interest from an academic point of view [37][38][39][40][41]. Such materials not only work in a similar way to mineral semiconductors but also have positive features such as environmental friendliness characteristics, easy preparation and light weight. Using a polymer as photoelectrode is original and modified PANI is attractive owing to its better stability and low cost [42,43]. The potential of the conduction band (CB) is more cathodic than the H2 level making PANI attractive for the reduction of water into hydrogen [44,45]. It has a small band gap (Eg) allowing the exploitation of a large part of sunlight [46]. Unlike physical characterization, the photo-electrochemical (PEC) properties of modified PANI and hydrogen evolution have been little studied.
Conducting PANI is sensitive to the pH and the chain of polyaniline undergoes a deprotonation 3 in basic solution (pH ~ 9 -11). Conversely, in acid medium (pH ~ 2 -4) it converts into insulator quinoïd accompanied by the deintercalation of counter anions where the chain becomes protonated with a small increase of the conductivity. The present work is devoted to the PEC characterization of the modified polyaniline/1.5-naphtalene disulfonic acid (PANI-NDSA) prepared by oxidation of aniline using persulfate as oxidant and 1.5-naphtalene disulfonic acid (NDSA) as dopant. The photo-electrochemical characterization, reported for the first time, is a preamble of the application for the hydrogen formation under visible light illumination. During the interfacial polymerization of polyaniline, 1.5-Naphtalene Disulfonic Acid (NDSA, powder) was used as dopant. PANI is an intrinsically conductive polymer; insulator in the neutral state. Therefore, to modify the transport properties with an insulating-4 semiconductor transition), a direct chemical dopant is required and NDSA is suitable in such a case. The latter is a stable diacid up to 290 ° C, thus providing function-structure stability (scheme 1).

Experimental
Thermal analysis (TG) was carried out using TGA Q 500 TA Instrument at a heating rate of 10 °C min -1 .The phases were identified by X-ray diffraction using Cu Kanticathode (= 0.15418 nm) at a scanning rate of 2° (2) min -1 . The UV-Visible spectra were plotted with a spectrophotometer (Specord Plus 2000) attached to an integrating sphere, PTFE was used as standard. The electrical conductivity (σ) was measured in the interval (300-450 K) by the two probe method using home-made equipment. For the thermo-power, the pellet was sandwiched between two stainless steels; the thermal gradient (ΔT) was obtained by Cr-Al thermocouples whereas the electro-motrice force (e.m.f.) was measured thanks to a LCR Agilent 4263 B LCR equipment.
The permittivity ( ~ 5920) of PANI-NDSA was determined from dielectric measurements at 10 kHz on dense pellet with a GW Instek GDM-8261A LCR-meter.
The powder was compressed into discs (= 1.3 cm, thickness 0.6 mm) under 510 2 MPa. To minimize the contact resistance, Ag paint was contacted to the back surface of the discs, which are enrobed in Teflon tubes leaving a contact surface area of 1.32 cm 2 with the electrolyte. The photoelectrochemical (PEC) characterization was done in aqueous solutions in potentiostatic mode using a standard Pyrex cell with a Pt emergency electrode. The potential of the working electrode was monitored by a computer controlled potentiostat (Solartron, SI: 1287) and given with respect to a saturated calomel electrode (SCE).The capacitance measurements are plotted at a frequency of 10 KHz. The point of charge zero (pzc) is measured from the equilibrium pH of an aqueous powder solution with a digital pH meter. The complex impedance spectroscopy (EIS) was undertaken with small amplitude wave signals in a wide region (10 -2 -10 5 Hz) [47][48][49][50][51].
The photocatalysis was done in a double walled Pyrex reactor connected to a cooling system, the temperature was regulated at 50  1 °C;125 mg of PANI-NDSA powder were dispersed in 250 mL of electrolytic solution Na2SO4 (10 -2 M) containing K4Fe(CN)6 (10 -3 M) under magnetic agitation (210 rpm). Prior each experiment, N2 gas was bubbled in the solution for 20mn at a constant rate of 5 mL mL -1 . Visible light illumination was provided by three tungsten lamps (3×200 W), giving a total intensity of 29 mW cm -2 (2.0910 19 photons s -1 ). The liberated gas was identified as hydrogen by gas chromatography (Shimadzu IGC 121 ML). The volume was quantified with a water manometer in a graduated burette (5 mm inner diameter).
Blank experiments were realized and no H2 evolution was observed in the dark. The solutions were prepared from reagents grade chemicals in demineralized water.

Results and discussion
The TG plot of PANI-NDSA (Fig. 2) is carried out in order to delimit the thermal stability of PANI-NDSA. The decomposition, accounting for a first loss at ~ 50 ° C to 120 ° C attributed to the elimination of physisorbed water, a small weight loss (~ 1 %), occurs over a large temperature range; it starts at ~ 120 o Cand ends at ~ 290 °C with an inflexion point at ~300 °C. The XRD pattern illustrated in

Optic and transport properties
The optical properties of PANI-NDSA, important in photocatalysis, are determined from the diffuse reflectance spectrum. The relation between the absorption coefficient (, cm -1 ) and the incident photon energy (h, eV) is given by the Pankov relation: 6 The exponent m depends on the transition type, it is equal to 2 and 1/2 respectively for direct and indirect transitions. The extrapolation of the straight line (αh) 2 with the h-axis (Fig. 5) give a direct transition at 1.96 eV. The band structure of PANI-NDSA consists of a lowest unoccupied molecular orbital (LUMO) which constitutes the valence band (VB) separated by a forbidden band (Eg) from the highest occupied molecular orbital (HUMO: conduction band, CB), thus referring to   * transition [36]. The conjugated poly-heterocycles have two states, not equivalent namely the aromatic and quinonic states. The transition from the aromatic to the quinonic state augments the energy of HOMO and decreases that of LUMO, which explains the narrowing of the gap at 1.03 eV [57].
The thermal dependence of the AC conductivity (σ) of PANI-NDSA shows a non degenerate behavior (Fig. 6) with a conduction mechanism by electrons jump among localized sites; an activation energy Ea(= 0.24 eV) of was deduced from the slope log/dT. The knowledge of the energetic position of the bands LUMO and HOMO is a prerequisite for the photocatalytic study and the photo-electrochemistry is widely used for providing the semiconducting properties. Such value indicates that the conduction mechanism occurs by small polaron hopping, behaving like a heavy particle with a limited free path and a low mobility at 300 K (µh= 8.78 cm 2 V -1 s -1) as it could be anticipated from the moderate conductivity.
However, the thermo-power (S) defined as the proportionality constant between the gradients of potential (V) and temperature (T); is insensitive to defect states and grains boundaries and should highlight the conduction mechanism in PANI-NDSA. The positive sign of S over the studied temperature domain (Fig. 7), confirms that the majority charge carriers are holes, involved conduction mechanism by small polaron hopping among mixed states. S300K averages 0.80 mV K -1 in agreement with a non-degenerate conductivity and should yield a density of ~ 10 18 cm -3 where the lattice phonon scattering governs the conduction mechanism; in such a case S is expressed by:

ES (0.24 eV) is the thermal activation for the charge carriers and A an entropic constant which
can be ignored for small polaron jump. The large difference between the energy values (E and ES) is an evidence of a conduction mechanism by small polarons in PANI-NDSA. 7

Photo-electrochemical characterization
PANI-NDSA is a potential photocatalyst for the energy supply and the photoelectrochemistry is a powerful tool to establish the energy band diagram (Fig. 11). The Intensity-Potential J(E) plots were recorded to check the electrochemical stability and to elucidate the interfacial redox reactions; the measurements were carried out once the open circuit potential (OCP) stabilized. Na2SO4 (0.1 M) was used as support electrolyte because it high conductivity.
The cyclic J(E) characteristic of PANI-NDSA is plotted to elucidate the electrochemical behavior, the plot (Fig. 8) shows an hysteresis loop where the current is in conformity with a classical. Below ~ -0.30 V, the current shoots up due to H2 liberation (gas bubbles were noticeable on the electrode) and the corresponding potential is found to be pH independent. One can mention in passage that the over-voltages are minimized and PANI-NDSA can be attractive in the water splitting.
The flat band potential (Efb) of PANI-NDSA (= +0.82 V) is reliably provided from the intercept of the linear part to infinite capacity C -2 = 0 of the plot (Fig. 9):  Fig. 10. (a) is slightly depressed indicating a deviation from pure capacitor, Q(i) -k is related to the phase angle  (= n/2),  the angular frequency and k is the homogeneity factor (0 < k  1). The semicircle is due to the bulk material (R2= 437  cm 2 ) whose center is localized below the abscissa axis with a depletion angle and the impedance of the phase constant element (CPE1-P). The line at low frequencies, inclined at 62.3° indicates the disparity from a pure Warburg diffusion RW (= 1.076 KΩ cm 2 ). The small shift from the origin (~ 42 cm 2 ) is attributed to the electrolyte resistance (R1, Na2SO4 0.1 M). The data are modeled to an electrical equivalent circuit comprising the resistance R1 connected in series with the resistance RW in parallel with the CPE2-P. The EIS data are suitably adjusted to the proposed equivalent electrical circuit by using the software Zview (Fig. 10. (a), Inset).
The Bode representation i.e. │Z│and phase angle  as a function of log (frequency) is illustrated in Fig. 10. (b) Lets' recall that for a capacitive behavior,  approaches -90° while a tendency to 0° is reached for pure resistance. The phase  increases up to 37° with raising the frequency in the range (10 -2 -10 5 Hz), confirming the circuit Rct-Cdl where Rct is the charge transfer resistance, Cdl the capacity of the double layer ( Fig. 10. (b) A good activity is obtained at pH 6.8 in presence of Fe(CN)6 2which favours the separation of charge charges (reaction 11) and protect the catalyst against the photocorrosion. The volume of evolved hydrogen increases monotonically over illumination time with an average rate of 3840 μmol h -1 (g catalyst) -1 . The end product Fe(CN)6 2competes with water for the photoelectrons (reaction 10); its redox potential ( 0.12 V) is close to that of H2O/O2 level and should maintain a more or less linear behaviour. Indeed; when the reaction product, Fe(CN)6 3in the present case, is not withdrawn from the electrolytic solution by a chemical or physical process, it react at the opposite pole of the crystallite leading to a decline in the photocatalytic activity. This occurs with Fe(CN)6 2-, where the rate is high at the beginning and gradually slows down over time (Fig. 12). PANI-NDSA is black and absorbs over the whole solar spectrum, thus permitting the determination of the quantum efficiency () of photons into hydrogen: 10 Where N is the Avogadro number, V the volume of evolved H2, Vm the molar volume and P the incident photons (2.0910 19 photons s -1 ).The factor 2 in the relation 12 enters because the H2 molecule requires two electrons; a value of 0.34% is determined. Interestingly, the initial activity of the catalyst during the second cycle is practically restored and very weak deactivation ( 2%) is observed (Fig. 12). The hetero-system PANI-NDSA/ZnO will be applied for the reduction of heavy metals like nickel upon visible light. The preliminary results are conclusive and the detailed study will be reported very soon.
It is instructive at this level to compare the performance of PANI-NDSA with those already reported in the literature (Table 1). One can see that our catalyst shows an activity far superior than most modified PANIs. Moreover, we worked under visible light irradiation with a catalyst dose less than 0.5 g L -1 .

Conclusion
PANI-NDSA is chemically stable, inexpensiveness and non-toxic. Its semiconducting properties have been studied for the first time to assess its photocatalytic hydrogen evolution upon visible irradiation. It is thermally stable up to ~ 290 °C and the X-ray diffraction indicated broad peaks, characteristic of nano-morphology and a crystallite size of ~ 7 nm. An optical transition at 1.96 eV, directly permitted is obtained from the diffuse reflectance. The electrical conductivity obeys to an exponential law with activation energy of 0.24 eV. The p-type behavior of PANI-NDSA is highlighted from the (capacitance -2potential) plot. A flat band potential (Efb) of 0.82 VSCE was determined in neutral electrolyte. The contributions of the bulk and grain boundaries predominate in the electric conductivity. PANI-NDSA is successfully employed for the hydrogen formation when exposed to visible irradiation because the potential of its conduction band is more cathodic than that of H2O/H2.A quantum yield of 0.34%was obtained using Fe(CN)6 4as hole scavenger. The photoactivity was completely restored during the second cycle.