Chapter 2

Basics of PEC Water Splitting

2.1 Photoelectrochemical Splitting of Water PEC solar hydrogen production, a well established method of hydrogen generation by the splitting of water is considered to be a superior technology as hydrogen could be obtained directly from abundantly available renewables i.e. water and solar light [Fujishima et al, 1972; Gratzel, 2001; Dresselhaus et al, 2001; Lewis, 2001; Tseng et al, 2010]. Photoelectrochemical water splitting incorporates conversion of solar energy into electrical energy by using semiconductor/electrolyte junction. The efficient photoelectrochemical cell converts water into hydrogen and oxygen using sunlight in a two electrode system; semiconductor (possessing either p or n type conductivity) as working electrode and platinum (Pt) as counter electrode (as shown in Figure 2.1). However in most measurements reference electrode is also employed to investigate half reactions in the PEC cell.

2.1.1 Mechanism of Photoelectrochemical Water Splitting In PEC cell, upon illumination of semiconductor photoelectrode with photons having energy equal or larger than the band gap (Eg) of the semiconductor, electron-hole pairs are generated (Figure 2.1). For semiconductor material, the reaction is expressed as: SC + hυ

e-SC + h+SC

(2.1)

These electrons-holes get separated and migrate to the surface of the semiconductor and to the counter-electrode without recombining. In case of n-type semiconductor electrode, the holes react with water molecules at semiconductor surface (eq. 2.2) resulting into O2 formation whereas electrons travel through external circuit and are transported to the counter electrode where they reduce H+ into H2 (eq. 2.3). 2h+ + H2O (liquid)

½ O2 (gas) + 2H+

2H+ + 2e-

H2 (gas)

31

(2.2) (2.3)

Chapter 2

Basics of PEC Water Splitting

On the contrary, p-type semiconductor electrode produces H2 and O2 at semiconductor electrode and at counter electrode respectively [Gratzel, 2001]. In other words, fast transfer of electrons towards the electrolyte than that of holes leads to flow of cathodic photocurrent (p-type semiconductor), while fast hole transfer towards the electrolyte resulting into anodic photocurrent (n-type semiconductor) in photoelectrochemical water splitting [Walter et al, 2010].

Accordingly, the overall reaction of the photoelectrochemical splitting of water may be expressed as: 2hυ +H2O (liquid)

½ O2 (gas) +H2 (gas)

(2.4)

Splitting of water to generate hydrogen and oxygen is an endothermic reaction and is accompanied by a large increase in Gibbs free energy ΔG° that can be provided by energetic photons and is expressed as: H2O

H2 + ½ O2

ΔG° = 237.14 kJ/mol (2.5)

(equivalent to 2.46 eV/molecule)

This water splitting reaction takes place when the energy of the photons absorbed by the photo-anode is equal to or larger than Et, the threshold energy: (2.6)

=

Where ΔG°H2O is the standard free enthalpy per mole of equation 2.4, which is equal to 237.14 kJ/mol and NA is Avogadro’s number (= 6.022 × 1023 mol-1). This yield: = hυ = 1.2289 eV

(2.7)

To summarize, hydrogen generation from water is a two-electron process. However, for oxygen generation each atom releases two electrons, so that generation of one molecule of oxygen is a four-electron process. The average energy of the electron process is 118.5 kJ/mol equivalent to ΔE0H2O = 1.23 eV/electron (lower hydrogen 32

Chapter 2

Basics of PEC Water Splitting

heating value reaction energy) based on liquid water conversion at 25˚C. Though, in practice ~1.8 eV energy (wavelength around 685 nm) is required for splitting of water [Bak et al, 2002]. The extra energy is to account for unavoidable loss mechanisms such as electrode overpotentials (as discussed earlier in section 1.6.1).

Figure 2.1: Energy level schemes for a Photoelectrochemical cell: (a) At equilibrium; (b) Under illumination.

In a PEC cell, the electrolyte can be described by two redox potentials, Uo(H2O/H2) and Uo(H2O/O2), which differ by 1.23 eV. At equilibrium (Figure 2.1 (a)), the electrochemical potential (or Fermi level) is constant in the whole system and occurs somewhere between the two standard energies Uo(H2O/H2) and Uo(H2O/O2). Fermi energy level is an electronically conducting phase of a metal/semiconductor, measured with respect to reference energy level and defines that the probability of occupancy of an electronic energy level (free electrons) is one-half. The relative Fermi energy (Ef) position of the semiconductor depends on the electron and hole concentration i.e. on n- or p-type conductivity of the semiconductor. For p-type semiconductor, the Fermi energy level lie above the valence band, while Fermi energy level of n-type semiconductor is located below the conduction band. The two reactions, O2 formation at the n-type semiconductor and H2 formation at the counter electrode, can only occur if the bandgap of the semiconductor is >1.23 eV, and the 33

Chapter 2

Basics of PEC Water Splitting

conduction band edge is negative of Uo(H2O/H2) while the valence band edge is positive of Uo(H2O/O2). At equilibrium, the Fermi level within the metal is below Uo(H2O/H2) so that H2O cannot be reduced in the dark. The energy bands are flattened on light excitation until the Fermi level in the metal is moved above Uo(H2O/H2) so that electron transfer to electrolyte becomes possible [Fujishima et al, 1972; Gratzel, 2001; Bak et al, 2002].

2.2 Semiconductor - Electrolyte Interface Semiconductors possess intermediate conductivity i.e. between a conductor and an insulator. In semiconductor, highest and lowest energy levels of a band are referred as the band edges. As with molecular orbitals, the highest occupied orbitals, called the valence band (VB) and the lowest unoccupied orbitals called the conduction band (CB). The bandgap (Eg) is the energy gap between these bands i.e. the difference in energy between the upper edge of the valence band (E v) and the lower edge of the conduction band (Ec) that determines the properties of the material. The bandgap energy (Eg) of a semiconductor is usually determined by means of optical absorption (as discussed earlier in section 1.6.1). Thus, a semiconductor is characterised by its energy bands, i.e. by the conduction and valence band and its Fermi level. A semiconductor - electrolyte junction is formed when a semiconductor is immersed in an appropriate electrolyte. The chemical potential of a semiconductor is defined by Fermi level, whereas the chemical potential of a liquid electrolyte is determined by its redox potential.

2.2.1 Depletion Layer and Band Bending When the semiconductor electrode is immersed in a redox electrolyte, charge carriers are transferred at semiconductor/electrolyte interface until the equilibrium is attained i.e. the condition when Fermi energy of the semiconductor and the redox potential are at the same energy level on both sides of the interface. The charge transfer leaves behind a depletion region in the semiconductor and a potential barrier at the interface. The charge in the depletion layer is compensated by 34

Chapter 2

Basics of PEC Water Splitting

opposite charge i.e. induced in the electrolyte within a localized ~1 nm thick layer, known as Helmholtz layer. It is composed of oriented water molecule dipoles and electrolyte ions adsorbed at the electrode surface. The potential drop, VH, across the Helmholtz layer (also known as the Helmholtz barrier) is determined by the nature of the aqueous environment of the electrolyte and the properties of the photoelectrode surface. Broadly, at semiconductor - liquid interface, there is existence of three double layers as shown in Figure 2.2. First is the “Space Charge layer” or the “Depletion region” in the semiconductor. Second is the charged layer between two planes - one plane that is the surface of the semiconductor distinguished because the adsorbate and the surface state are bonded by the same charged bond and the second plane called “outer Helmholtz plane (ohp)” that is the position of the closest approach of mobile ions in the solution. This layer is called Helmholtz layer.

Figure 2.2: Band diagram of n-type semiconductor/electrolyte interface under illumination.

Typically, Helmholtz layers are on the order of a few nm, compared with several microns for the semiconductor space-charge region. The third layer is called “Gouy 35

Chapter 2

Basics of PEC Water Splitting

Chapman layer” that is an extended region of space charge layer that compensates the charge in the semiconductor uncompensated by the Helmholtz layer [Bak et al, 2002; Morrison et al, 1980]. It is worthwhile to summarize the three main situations to be found in the double layer of a semiconductor: (a) accumulation layer, where the excess charge has the same sign as the majority carrier in the bulk crystal and is constituted mainly by majority carriers, not ionized donors or acceptors; the width of this layer is in the range of 100 Å. (b) depletion layer, where the excess charge has the opposite sign of the majority carrier and consists mainly of ionized donors or acceptors for n- or p-type materials, respectively; the width of the depletion layer is usually in the range of micrometers. (c) inversion layer, where the excess charge carrier adjacent to the surface has the opposite sign of the majority carrier in the bulk crystal and is mainly constituted by minority carriers; the width of the inversion layer is in the range of 100 Å. The performance characteristics of photoelectrochemical (PEC) cell depend, to a large extent, on the potential drop, VH, across the Helmholtz layer. The junction has characteristics similar to those of a Schottky barrier formed between a semiconductor and a metal. The semiconductor has a net space charge of immobile, positively charged donor ions uncompensated by the mobile electron cloud. Figure 2.2 shows the formation of the depletion and Helmholtz layers, field, and potential across a semiconductor–electrolyte interface. The formation of the double layer leads to conduction and valence band bending and development of a field that opposes further charge transfer. Finally, equilibrium is reached when there is no net transfer of charge across the junction. The equilibrium potential barrier is formed with a built-in voltage VB as a result of the band bending. Under illumination, a photovoltage is generated in photoelectrode due to the separation of photogenerated e- - h+ pairs in the band bending region. The charge separation persists until the bands are flattened, afterwards the photovoltage or the charge 36

Chapter 2

Basics of PEC Water Splitting

separation cannot be increased by further intense illumination. To summarize, charge transfer processes at the interface results in generation of potential gradient (i. e. space charge region) across the interface that leads to bending of energy bands either upward (n-type semiconductor) or downward (p-type semiconductor) as shown in Figure 2.3. If the initial Fermi level of an n-type semiconductor is above (more negative with respect to) the initial electrolyte redox potential, electrons will flow from the n-type semiconductor to the electrolyte [Choudhary et al, 2012].

Figure 2.3: Band Diagram and Charge Carrier Profiles of n-type and p-type Semiconductor/Liquid Junctions.

Nozik [Nozik, 1977; Nozik et al, 1996] proposed that electron - hole separation can be enhanced by band bending. He proposed the concept of photochemical diodes that consist of two electrodes of a photoelectrochemical cell fused together producing either p - n type or Schottky - type devices. The upward band bending represents a negative surface that repels the electrons but attracts the holes and the opposite of this holds true for downward band bending. As a result, the semiconductor with bent bands should react more efficiently as compared to the flat

37

Chapter 2

Basics of PEC Water Splitting

band situation. Also, more band bending and/or wider space charge layers should make the reaction more efficient. Additionally, the presence of light leads to lowering of the H+/H2 potential while application of external bias results in elevation of Fermi energy level of cathode above the H+/H2 energy level, thereby facilitating electron transfer to H+ ions of the electrolyte for generation of hydrogen [Choudhary et al, 2012]. In PEC system the electrons are mainly responsible for generation of molecular hydrogen either on the counter (Pt) electrode in n-type semiconductor or on the semiconductor electrode in the p-type semiconductor [Mohapatra et al, 2008]. The average distance travelled by electron/hole before trapping or recombination is referred as e-/h+ diffusion length (L). It has been reported that the light absorbed within a distance of space charge width and e-/h+ diffusion length contribute to the photo-response in water splitting process [Morrison, 1980; Leng et al, 2010]. For efficient PEC systems, the diffusion length of charge carriers (L) should be comparable or larger than the thickness of film (d) i.e. L ~ d or L > d, so that photogenerated e-/h+ are collected before their recombination i.e. enhancement in the lifetime of charge carriers [Leng et al, 2010]. For efficient water splitting, photogenerated charge carriers needs to be separated quickly and transferred at the counter electrode through external circuit without recombining with holes. The recombination losses in the bulk as well as at the surface reduce efficiency if the charge transfer kinetics within the electrolyte is slow [Bisquert et al, 2004].

2.2.2 Flatband Potential The flat band potential, VFB, is the potential that has to be imposed over the electrode-electrolyte interface in order to make the bands flat [Morrison, 1980]. Flatband potential is the potential of the semiconductor bulk for which the band bending is completely removed so the band edges are represented by a flat line when approaching the interface from the bulk. PEC splitting of water may take place

38

Chapter 2

Basics of PEC Water Splitting

when the flat band potential is higher than the redox potential of the H+/H2 couple [Morrison, 1980]. The flat-band potential may be modified to the desired level through surface chemistry [Zhao et al, 1999; Kozuka et al, 2000]. When a photoelectrode is illuminated by photons with energy greater than its bandgap, a photovoltage is generated due to the separation of electron-hole pairs in the band bending region. On increasing the intensity of light, the height of band bending decreases. Once the bands are flattened, charge separation no longer occurs, so more intense illumination cannot further increase the photovoltage. Thus, the flatband potential is the most extreme potential (negative or positive depending upon the doping of the semiconductor) that can be generated in the semiconductor bulk by illumination. The flatband potential VFB is an important parameter that is related to the properties of both the semiconductor and the electrolyte. It is given by VFB = (χ + ΔEF + VH) – 4.5 = (Φsc + VH) – 4.5

(2.8)

where χ is the electron affinity; ΔEF is the difference between the Fermi level EF and the majority carrier band edge of the semiconductor (conduction band in the present case); Φsc is the work function of the semiconductor and measure the Fermi level potential with respect to vacuum; and 4.5 is the scaling factor [Nozik, 1981; Morrison, 1984; Bak et al, 2002]. Because of the adsorption equilibrium for H+ and OH− ions between the surface of semiconductors and an aqueous (aq) solution, the semiconductor surface attains the point of zero charge (PZC). The flatband potential directly determines the possibility of occurrence of gas evolution reaction at the counter electrode with sufficient illumination of the photoelectrode. For n-type photoelectrode, if VFB is more negative than the hydrogen half reaction potential, then hydrogen production can occur at the counter electrode and oxygen evolution can occur at the photoelectrode surface if the potential of the valence band edge at the surface is more positive than oxygen half

39

Chapter 2

Basics of PEC Water Splitting

reaction. For a p-type semiconductor, if VFB is more positive than the oxygen half reaction, then oxygen production can occur at the counter electrode. The flat-band potential VFB of most semiconductors is determined solely by pH and shifts proportionately with pH, with a slope of -59 mV/decade, i.e., pH, for example, Ss – OH + H+aq = Ss – OH2+

(2.9)

Where, Ss – OH refers to the OH group present at the semiconductor surface [Nakato, 2002]. Changes in pH of electrolyte can change the flatband potential of a photoelectrode by changing the adsorption layer on its surface. Changes to the adsorption layer can affect both the surface dipole and the semiconductor’s surface Fermi level. Dipole changes can be due to density of adsorbed species. Surface Fermi level changes can be caused by adsorbate dependent changes in the semiconductor’s surface reconstruction or passivation of surface states. It is common for oxide semiconductors and III-V semiconductor to have a pH dependent flatband potential that parallels the pH dependence of the hydrogen and oxygen reactions [Turner, 1983]. This is due to the formation of oxides on the semiconductor surface, which is a pH dependent process. Thus, flatband potential plays an important role in production of hydrogen in PEC cell. Flatband potential can be measured in a variety of ways [Turner, 1983]. Ideally, these methods should give the same result. In the present work Mott-Schottky plots have been utilized for determination of flatband potential.

2.3 Material of Interest Research work in this thesis is directed at development of nanostructured metal oxide semiconductors for PEC splitting of water. Long-term stability and suitable interfacial energetics of photoelectrode has been considered as the essential prerequisite for the commercial viability of photoelectrochemical water splitting process to produce hydrogen. Many metal oxides like TiO2, BaSrTiO3, SrTiO3, BaTiO3,

40

Chapter 2

Basics of PEC Water Splitting

ZnO, Fe2O3, WO3 and copper oxides have been studied as photoelectrode in PEC cell for water splitting since 1970’s [Bjorksten et al, 1994; Aroutiounian et al, 2005]. Large band gap semiconducting oxides such as TiO2, WO3, SrTiO3, BaTiO3, SnO2, ZnO etc. are stable in aqueous electrolyte but absorb in UV region which is only about 4% of the solar spectrum, whereas small band gap semiconductors such as Si, GaAs, InP, CdTe, CdSe, CuO etc. and optimum band gap semiconductor viz. Cu2O have the potential to absorb visible part of solar spectra but corrode when dipped in electrolyte [Hyun et al, 2000; Mishra et al, 2003]. Intermediate band gap semiconductor like Fe2O3 absorbs in the visible region but suffer from poor semiconductor characteristics due to redox level mismatch, low mobilities of holes and trapping of electrons by oxygen-deficient iron sites. In order to improve the lifetime of PEC system, the semiconductor material must have adequate electrochemical stability so that the charge carriers reaching at its surface drive only the water splitting reactions without any side reactions (i.e. electrode corrosion). Various modification techniques such as doping, SHI irradiation, dye sensitization, bilayered systems etc. (as discussed in section 1.8) have been used by researchers to tune the desired properties of the semiconductor material in order to enhance photoelectrochemical response in visible range of the solar spectrum with higher efficiency. In the present piece of work, bilayered modification strategy (as discussed in section 1.8.4) has been employed in order to utilize the advantages of two semiconductor metal oxides in a single photoelectrode. Such systems may cover large portion of the solar spectrum [Choudhary et al, 2012], stable and are found to be efficient in PEC splitting of water. Wide survey of literature revealed that nanostructured CuO [Singh et al, 2009], SrTiO3 [Mavroides et al, 1976], ZnO [Zhou et al, 2004; Cui et al, 2010] and WO3 [Berger et al, 2006; Wenzhang et al, 2010] to be promising candidates for photoelectrode in the PEC cell on account of their favorable band edge alignment with each other as shown in Figure 2.4.

41

Chapter 2

Basics of PEC Water Splitting

Figure 2.4: Conduction and Valence band edge positions of Metal Oxides with respect to redox potential of water.

2.3.1 Copper Oxide (CuO) Copper oxide is of great interest in semiconductor physics. Copper forms two wellknown stable oxides, cupric oxide (CuO) and cuprous oxide (Cu2O). These two oxides have different physical properties, different colors, crystal structures and electrical properties. CuO has monoclinic tenorite structure (Figure 2.5), appears black in color, abundantly available, non-hazardous source materials, and it can be prepared by low cost solution methods which is the key advantage in PEC application.

Figure 2.5: Crystal structure of Monoclinic CuO

42

Chapter 2

Basics of PEC Water Splitting

CuO has a narrow (direct) band gap of around 1.2 - 1.7 eV in bulk [Singh et al, 2009]. Copper oxide intrinsically is a p-type semiconductor because of cation-deficiency which may be due to the copper vacancies or oxygen interstitials. There are various methods of fabricating CuO, such as thermal oxidation, chemical vapour deposition etc. Thus, the possibility of low cost production methods and the good electrochemical properties make CuO to be one of the best materials for electrical, optical and sensing applications. CuO also displays high absorption in the visible regions making it a promising candidate as selective absorber material in solar cells [Kennedy, 2002].

2.3.2 Strontium Titanate (SrTiO3) Wrighton and co-workers [Wrighton et al, 1976; Bolts et al, 1976] first introduced the use of single crystal n-SrTiO3 as a photoanode and Platinum (Pt) as cathode in an electrochemical cell, resulting in a sustained conversion of water to O2 and H2 with no external bias. The photoelectrochemical properties of SrTiO3 electrodes are similar to those of TiO2, although the onset potential of photocurrent shifts negative by about 0.3-0.35 V. With a slight increase in external bias, the photocurrent increases more rapidly and sharply than that in comparison to TiO2 or SnO2. Even at zero bias a reasonable amount of photocurrent in SrTiO3 is observed that could initiate stoichiometric water splitting. The anodic photocurrent, which is proportional to the intensity of light on the photoanode, begins to flow for wavelengths less than 390 nm (~ 3.2 eV) and reaches a maximum for a wavelength of 330 nm. Mavroides et al illuminated SrTiO3 photoanode and observed splitting of water without bias in presence of 10 M NaOH [Mavroides et al, 1976]. The maximum external quantum efficiency is found to be 10 %, about an order of magnitude greater than the highest value obtained for photoelectrolysis in cells with crystalline TiO2 electrodes in the absence of a bias voltage. This increase in quantum efficiency is due to the increased band bending of SrTiO3, which is about 0.2 eV for a cell without a bias voltage. The efficiency of electron-hole pair separation increases by the application of a bias voltage to make the anode Fermi level more positive than 43

Chapter 2

Basics of PEC Water Splitting

the cathode Fermi level, which is equivalent to Ered(H+/H2) in accordance with increasing amount of band bending [Mavroides et al, 1976]. Some of the advantages of SrTiO3 as photoelectrode in hydrogen production by photoelectrochemical water splitting are: a) stability towards photocorrosion, b) easy availability at low cost, c) environmental friendliness, d) well-matched energy band edges with the redox level of water, and e) electronic properties can be varied by changing the lattice defect chemistry or the oxygen stoichiometry (Figure 2.6).

Figure 2.6: Crystal structure of Cubic SrTiO3

Although band gaps of SrTiO3 (< 420 nm) is large thereby making it poor absorber in the visible region, several approaches are used to reduce the band gap of these oxides to improve the efficiency of metal oxide based PECs. The efficiency of water splitting is determined by the band gap, band structure of the semiconductors and the electron transfer process.

2.3.3 Zinc Oxide (ZnO) ZnO is a direct band gap material with a band gap around 3.2 eV at 300K. ZnO is considered as II-VI semiconductor. It has three types of crystal structures, rocksalt, zinc blende and wurtzite. Rocksalt structure will only appear at a relatively high pressure environment. Zinc blende structure can be achieved by growing ZnO onto substrates which are cubic structure. Normally at room temperature, it stays as wurtzite, which is thermodynamically more stable. Wurtzite structure has hexagonal unit cell with zinc hexagonal sublattice and oxygen hexagonal sublattice 44

Chapter 2

Basics of PEC Water Splitting

stacking together as shown in Figure 2.7. ZnO normally has n-type semiconducting property even without doping, while p-type doping in ZnO still remains a challenge. ZnO possess various electronic and optoelectronic applications. Recently, the effect of alloying on ZnO based nanostructures has brought attention because of its potential to improve its electrical, optical and other functional properties. By alloying, the band-gap can be tuned further into the UV or down into the green spectral ranges i.e. from 2.8 to 4.0 eV [Steiner, 2004]. Studies on various binary systems, such as MgO/ZnO [Shimpi et al, 2009], Cu/ZnO [Zhou et al, 2004] and Co/ZnO [Kim et al, 2004], have been done for different applications. It has been suggested that alloying CuxO into ZnO can help in the formation of heterostructured p-n junction, which can be utilized in solar and gas sensors applications [Zhu et al, 2006; Cui et al, 2010].

Figure 2.7: Crystal structure of Hexagonal Wurtzite ZnO

Favorable aspects of ZnO are: i) non-toxic nature, ii) cheap, iii) relatively abundant source materials, iv) chemically stable, v) synthesized as a large single crystal by various methods and vi) can be grown in various morphologies (nanostructures) and dimensions. ZnO nanostructures are promising candidates in miniaturized optoelectronics and sensing devices. Though, ZnO is transparent in the visible region of electromagnetic spectrum and its absorption increases abruptly in the ultraviolet region for photon energies greater than 3.2 eV, which is the major drawback of ZnO during its application in PEC devices.

45

Chapter 2

Basics of PEC Water Splitting

2.3.4 Tungsten Oxide (WO3) WO3 was suggested as a candidate for photoelectrochemical hydrogen production as early as in 1976 [Hodes et al, 1976]. Tungsten trioxide (WO3) is n-type semiconductor with a range of potential applications including: gas sensing [Lee et al, 2000], solar energy and electrochromic devices [Jelle et al, 1999; Turyan et al, 2000], catalysis [Habazaki et al, 2002] and photooxidation of water [Miller et al, 2006]. WO3 has an indirect band gap of ~2.8 eV that absorbs a reasonable fraction of the solar spectrum, has no photo-corrosion, and has high stability in aqueous solutions under acidic conditions, making it a promising photocatalytic material. Except for the light absorption in the visible light region, the surface area of WO3 electrode is also another key factor in the efficiency of a photoelectrochemical cell. A nanocrystalline structure photoanode not only has a large surface area that provides sufficient contact between the electrode and electrolyte, but it also induces fluent charge transfer due to the short diffusion length of the minority charge carrier [Wenzhang et al, 2010]. Figure 2.8 shows orthorhombic crystal structure of WO3. Several different techniques have been reported to prepare nanostructured WO3 films, including vacuum evaporation [Colton et al, 1978], reactive sputtering [Marsen et al, 2007], sol-gel techniques [Santato et al, 2001], chemical vapour deposition [Sivakumar et al, 2004], anodization [Berger et al, 2006] and electrodeposition [Baeck et al, 2003].

Figure 2.8: Crystal structure of Orthorhombic WO3

46

Chapter 2

Basics of PEC Water Splitting

WO3 based PEC devices in a mechanical tandem stack configuration with a-Si solar cell have already demonstrated stable solar-to-hydrogen efficiencies in the range of 3-5 % [Gratzel, 2001; Marsen et al, 2007]. However, the bulk band gap of WO3 (2.6– 2.8 eV) is large to sufficiently absorb the solar spectrum. Furthermore, the electronic surface structure of WO3 is not ideal for unassisted water splitting.

2.4

Bilayered Metal Oxides as Potential Photoelectrodes in PEC

Water Splitting Bilayered systems are the most exciting and recent strategy to achieve absorption of large part of the solar spectrum for increasing the efficiency of the process. Multilayers expressed by [(A)t1 / (B)t2]n where t1 and t2 are the thicknesses of semiconductor layers deposited n times, are called as bilayers (n=1) or heterostructures. Generally bilayers encompass two main aspects: (i) the fine control of the interfaces by synthetic techniques, which may lead to unexpected properties [Ohtomo et al, 2004; Reyren et al, 2007], and (ii) bilayer of two materials may cover large portion of solar spectrum [Siripala et al, 2003; Zhang et al, 2007].

Figure 2.9: Schematic diagram of Bilayered System under illumination.

47

Chapter 2

Basics of PEC Water Splitting

Bilayered electrodes comprises of two semiconductors, one with a wide band gap while another with a small/mid band gap (as shown in Figure 2.9). Bilayers are the multifunctional materials which combine different properties of the individual semiconductor layers. The small band gap semiconductor is mainly responsible for sensitizing the large band gap semiconductor through electron or hole injection by visible light absorption. Efficient electron injection requires proper positioning of conduction and valence bands of large and small/mid band gap semiconductor [Hotchandani et al, 1992; Kamat, 2007]. The electrons transfer between the two semiconductors may enhance the charge separation and inhibit the recombination rate by forming a potential gradient at the interface. The energy layers in the bilayered devices can cover broad visible spectrum thereby offering synergistic effect. The overlying semiconducting layer absorbs photons of high energy, while being transparent to photons of lower energy. Subsequent underlying layer absorb the lower energy photons [Kale et al, 2009]. Thus, synergistic effect of both layers facilitates enhanced photocurrent and improved photoconversion efficiency [Siripala et al, 2003; Zhang et al, 2007].

2.4.1 Mechanism of Bilayered Semiconductors in PEC Splitting of Water A bilayered/heterojunction system comprises of electron-accepting semiconductor, hole-accepting semiconductor (with well aligned band edges and high electron or hole conduction ability) and the interface formed at the junction due to difference in the direction of electron or hole movement across the heterojunction [Lin et al, 2007]. The existence of interface with varying charges may lead to formation of internal electric field at the heterojunction which may induce the separation of photogenerated e--h+ pairs in both the semiconductors and at the interface, resulting into reduction in e--h+ recombination.

48

Chapter 2

Basics of PEC Water Splitting

Figure 2.10: Favourable band edge alignment of wide/mid/narrow bandgap semiconductors with redox potential of water and possible charge transfer across various heterojunctions under illumination.

The mechanism of bilayered semiconductor photoanode (in case, when the conduction band of the material deposited on substrate is lower than respective band of outer material which is in contact with electrolyte i.e. n-n junction as shown in Figure 2.10 (a) and 2.10 (b)) in the PEC cell can be understood as upon irradiation bilayered film leads to the excitation of electrons in the valence band of both the layers to the conduction band, leaving holes in the valence band. The absorbed light photons generate excitons throughout the bilayered semiconductors. The photogenerated electrons in the outer layer are injected into the conduction band of underlying material (coated on substrate) due to the effect of heterojunction 49

Chapter 2

Basics of PEC Water Splitting

interface, and the electrons easily reach at the substrate surface, and are transferred to the cathode via back contact to produce hydrogen. Photogenerated holes on the other hand move into the valence band of outer material in the opposite direction i.e. driven by the built-in field of heterojunction, and finally holes are carried out by the electrolyte (H2O) to produce oxygen at photoanode in the PEC cell. Alternatively, the electron-hole transfer across the favourable band edge position of the photocathodes (i.e. p-p junction) occur in the opposite directions compared to nn junction [Figure 2.10 (c)] producing H2 and O2 at semiconductor and counter electrode respectively. Furthermore, the favourable energetics in p-n junction leads to electron-hole transfer across the junction as shown in Figure 2.10 (d) and 2.10 (e). In bilayered semiconductors, the photogenerated charge carriers travel few nanometers before reaching the interface, which is proportionate with the diffusion length of charge carriers [Dusastre et al, 2010], which may lead to efficient charge separation in bilayered systems. Electron-hole pairs are effectively separated at the junction of wide and mid/small band gap semiconductors as shown in Figure 2.10, which leads to the enhanced photoelectrochemical activity of bilayered thin films than single layered films. The synergistic effect results in enhancement in the characteristic photoelectrochemical response as discussed earlier [Xintong et al, 1998; Zhang et al, 2007; Dang et al, 2010]. However, proper alignment of the energy levels between the wide and mid/small layers is the fundamental requirement for efficient charge transfer in PEC water splitting. Optimization of thickness of the bilayered films is yet another important criterion towards facilitation of efficient separation of photogenerated charge carriers and their movement across the interface for photocurrent improvement [Sharma et al, 2010]. To summarize, the efficiency of bilayered systems for water splitting depends on their appropriate bandgap, proper band edge alignment with each other and high stability.

50