Investigation of nanostructured iron oxides as anodic material for water splitting

1. Introduction

To reduce our dependence on fossil fuels and reduce the emission of carbon dioxide, a large scale transition toward sustainable energy sources is necessary [1]. In this regards Hydrogen could play an important role in our modern life. It is a promising energy carrier, which could have a low impact on the environment and its energy content is 10 times greater than fossil fuels. Hydrogen production through the water splitting is a cheapest and clean source of energy [2]. On large scale production of hydrogen from the water splitting, largely depend upon the catalysts that are require to overcome the challenging requirement thermodynamically and kinetics of this half reaction [3, 4]. From an electrochemical point of view, it can be split into the two half reactions:

The oxygen evolution reaction (OER) is a key reaction and often the rate-determining step in electrochemical and Photoelectrochemical water splitting, which is a promising route to carbon-free hydrogen production. For this process we require the OER catalysts which are close to the state- of art catalysts. Iron oxide (\(\alpha\)-hematite) because it has been identified as a suitable water oxidation catalyst [5, 6, 7, 8, 9, 10], conjugating an overall good performance in both electrolysis and solar water splitting schemes with the advantage of being earth-abundant, non-toxic and environmentally safe.

Pulsed-laser deposition (PLD) of thin films is a technique that employs high-energy-density laser pulses to generate, in the regime of phase explosion, ablated material from a solid target, consisting of a mixture of vapor/liquid nanodroplets. PLD could present some significant advantages over the methods listed above: precise control of the quantity of the deposited material, enhanced adhesion due to the energetic nature of the process, and, most importantly, the possibility of nanostructuring the surface by the deposition of nanoparticles (NPs) [11, 12]. Additionally, being essentially a physical deposition method, it is suitable to all kind of substrates. The main drawback of the PLD technique is the need of specialized equipment, although this is already employed in industrial applications.

2. Materials and Methods

The target for the Pulse laser deposition consisted of the cold pressed with a pressure of about \(450-500Kg/cm^2\) for two times consecutively powders of Fe and Boric acid. Two sets of electrodes, fabricated on the FTO substrates with different thickness by changing the number of pulses: 5000 pluses for the thinner samples and \(10000\) pulses for the thicker samples. The fluency of the laser was fixed at 3\(J/cm^2\) and the deposition was carried out in a reactive oxygen atmosphere, at a pressure of \(1.5\times10^{-2}\) mbar. The number of pulses set at \(5000\) and the repetition rate at \(20Hz\). The coatings were deposited on FTO substrates, with a target - substrate distance of \(4.5 cm\). All the sample were subjected to a post-deposition annealing treatment, carried out at different Temperature i.e. \(500^oC, 600^oC\) and \(700^oC\) for 4 hours with heating rate of \(5^oC/min\) and one sample subjected at \(600^oC\) for 4 hours and again at \(800^oC\) for 1hours with heating rate of \(5^oC/min\). We represent the differently annealed sample with the following symbols:

Sample	Temperature
Thin sample annealed at \(500^oC\) for 4 hours	ANN\(@500^oC\)
Thin sample annealed at \(600^oC\) for 4 hours	ANN\(@600^oC\)
Thin sample annealed at \(700^oC\) for 4 hours	ANN\(@700^oC\)
Thin sample annealed at \(600^oC\) and \(800^oC\) for 4 hours	ANN\(@600^oC+800^oC\)
Thin sample annealed at \(500^oC\) for 4 hours	ANN\(@500^oC\) thick
Thin sample annealed at \(600^oC\) for 4 hours	ANN\(@600^oC\) thick
Thin sample annealed at \(700^oC\) for 4 hours	ANN\(@700^oC\) thick
Thin sample annealed at \(600^oC\) and \(800^oC\) for 4 hours	ANN\(@600^oC+800^oC\) thick

The following SEM images show the morphology of the thin films. Samples ANN\(@500^oC\) Figure 1 and ANN\(@600^oC\) Figure 2 show the presence of urchin-like structures, increasing the surface of the catalyst and thus the amount of active sites for the water oxidation.

The lengths of needles are not homogeneous, dense and aggregated nanoparticles are found at \(600^oC\) annealing temperature. Furthermore, SEM images show the nano needle with a non-uniform morphology. Samples show spherical nanoparticles covered with sparse nano flowers. The flower consists of needles petals grown radially from the surface. Probably the roughness present on the core surface is the starting point for the formation of these features. Overall SEM images confirm the urchin like structure on samples ANN\(@500^oC\) and ANN\(@600^oC\).

Sample ANN\(@600^oC + 800^oC\) Figure 3, show that there is no developed urchin like structures. At \(700^oC\) Figure 4 samples show spherical nanoparticles covered with sparse nano flowers. The flower consists of needles petals grown radially from the surface. Probably the roughness present on the core surface is the starting point for the formation of these features. As will be shown in the following sections, the electrochemical performance of these samples is not good. Here I discuss the thin sample SEM images because these are similar to the thick sample.

Figure 4 shows the Raman spectra of the \(Fe_2O_3\) annealed at different temperature. From the Raman spectra it is confirmed that the hematite modes are present. The peak present at 1070 might be an artifact or an impurity present on the surface of the sample after annealing at 500C for 4h. With increasing temperature the peaks became more sharp and visible; clearly indicating the crystallization of the samples.

2.1. Result and Discussion

Electrochemical water oxidation experiments were performed in a three electrode configuration with a working electrode (Iron oxide thin film on FTO), a platinum mesh as counter electrode and saturated calomel electrode (SCE) as a reference electrode, with Gamry Interface 1000 potentiostate. All the electrodes are immersed in a KOH electrolyte solution with pH=13. At this pH, oxygen evolution is expected at potential EO2/H2O = 1.23 - (0.059pH) = 0.463 V (463mV) versus reversible hydrogen electrode (RHE). The Figure 5 compares the anodic response of various thin electrodes annealed at various temperatures.

The oxygen evolution catalysts are required to speed up the reaction rate and reduce the overpotential. Thin Samples ANN@500oC and ANN@600oC shows the response toward the water oxidation at about 0.80 V vs RHE, with overpotentials of 343.9 mV and 370 mV respectively (taken at 0.2 mA/cm2 current density). Moreover, the Tafel slope of the electrode ANN\(@500^oC\) (81.6+0.6 mV/dec-1) is better than that of the electrode ANN\(@600^oC\) (109.8+0.2 mV/dec-1), showing good kinetic towards water oxidation. The electrode ANN\(@700^oC\) shows the response toward water oxidation at 0.90 V vs RHE. Moreover, it's provides the 455.6 V overpotential at 0.2 mA current density, with Tafel slope 135.8+0.5. So the efficiency toward the water oxidation is low. Electrode ANN\(@600^{o}C\)+\(800^{o}C\) shows the bad catalyst behavior towards the water oxidation (Figure 6).

The samples ANN\(@500^{o}C\) and ANN\(@600^{o}C\) show good catalytic performance, in term of overpotential. Large overpotential is associated with slower kinetics for the OER on hematite and with the electronic and structural characteristic of the hematite/electrolyte and hematite/substrate interfaces. After increasing the thickness of the film, then again film ANN\(@500^oC\) thick and ANN\(@600^oC\) thick showed good catalytic performance, both in term of overpotential and tafel slope as compared to the thin films (Table 1). Other samples ANN\(@700^oC\) and ANN\(@600^oC\)+\(800^oC\) display the bad catalytic enactment. The detrimental in water oxidation activity for the samples annealed more than \(600^oC\) could be due to the losing of the conducting properties of FTO. Since, it is clear known that FTO losses its conductivity above \(550^oC\). Thereby, it is clear that the decreased activity is mainly due to the losing FTO conductivity at high temperature annealing.
The overpotential is calculated as $$Overpotential=V_{Applied}+0.241-E^oO_2/H_2O$$ Where \(E^oO_2/H_2O = 463\) mV at pH = 13 and the 240 mV term is required to convert the applied potential Vapplied from SCE to RHE (Reversible Hydrogen Electrode). Samples overpotential and Table slopes are shown in Table 1.

Table 1. Overpotentials at various current densities and Table slopes of the thin and thick films annealed at various temperatures.

Sample	Overpotenital	Overpotenital	Overpotenital	Tafel slope
	at 0.2(mA)	at 1(mA)	at 10(mA)	mV/dec\(^{-1}\)
Ann\(@500^o C\) thin	343.9	412.4	769.8	81.6+0.6
Ann\(@600^o C\) thin	370.0	457.0	631.1	109.8+0.2
Ann\(@700^o C\) thin	455.6	534.5	699.9	135.8+0.5
Ann\(@600^o C+800^o C\) thin	708.9	103.9	---	373.7+2.9
Ann\(@500^o C\) thin	329.1	364.5	416.5	50.2+0.09
Ann\(@600^o C\) thin	398.8	436.6	495.6	56.0+0.1
Ann\(@700^o C\) thin	463.5	526.4	414.4	83.2+0.7
Ann\(@600^o C+800^o C\) thin	615.2	865.0	---	258.8+2.8

The significant difference in table slopes between thin and thick samples ANN@5000C and ANN@600oC could either depend on a different composition of the material (e.g. quantity of magnetite present) or simply be an effect of the different morphology, more porous in the case of thick samples.

3. Tafel analysis

The Tafel slope is also an important measure of electrode performance, because it accounts for changes in mechanism at different overpotential.

We calculate Tafel slope according to the Tafel equation Figure 7, which applies in the voltage region where the forward reaction (OER) is predominant compared to the backward reaction (HER). Figure 7 show the Tafel slope of best performing ANN\(@500^oC\) thick electrode.

4. Chronoamperometries

A good OER catalyst that use in the water splitting device must satisfy the two basic requirement. First it must be highly active towards its respective reaction. Secondly it must be stable in the oxidation condition and maintain the efficiency over the time scale for commercial use.

The 12 hours chronoamperometries were performed for checking the long term stability of the best performing samples: ANN@500oC thick and ANN\(@600^oC\) thick (Figure 8). In the first 2.8 hours the current density of the ANN@600oC thick electrode is increased, then it remains stable at 6mA/cm2 for next 9.2 hours. In the case of the ANN\(@500^oC\) thick electrode, in the first 6 hours the current density fluctuates between7 mA/cm2 and 5 mA/cm2 and then gradually decreases from 5.5 mA/cm2 to 5mA/cm2. It could be due to the bubbles formation on the surface of the electrode.The same pattern is detected when repeating the experiment a second time on the same electrode.

I am raw html block.
Click edit bBefore and after water oxidation the miro-Raman spectra for all the samples are represented in the Figure 9. It is indicating that after water oxidation the peak intensities were reduced, the probable reason could be due to the surface coverge of the electrolyte. However, the phase and crystallinity of the all the samples were well preserved ever after water oxidation, showing that the prepared thin film samples are highly stable to the water oxidation experiments.

5. Conclusion

We investigated that the Iron oxide (\(\alpha\)-hematite) was showing promising catalytic activity towards the oxygen evolution reaction and can be employed as anode in electrocatalytic water splitting techniques for hydrogen production. The best performing electrodes showed an overpotential@0.2 mA/cm2 of about 330 mV and a Tafel slope of 50mV/dec. These electrode metrics are in line with those of current state-of-the-art materials like IrO\(_2\) or CoPi. We observe that the deposition of more material, e.g 10000 instead of 5000 pulses, leads to better overpotentials and Tafel slopes, indicating that the performance can be enhanced simply by increasing the catalyst loading. That is normally not possible for bulk-like thick films or even thin-films where the surface coverage is already complete and in the present case it is likely resulting from the hierarchical 3D structure.

Author Contributions

All authors contributed equally to the writing of this paper. All authors read and approved the final manuscript.

Conflicts of Interest:

The authors declare no conflict of interest.utton to change this html