1 Introduction

Greenhouse gases from the combustion of fossil fuels cause air pollution and climate change [1]. Green energy technology is an urgent solution required to curb the foreseen challenges associated with the use of fossil fuels [2]. Hydrogen (H2) fuel production is projected as the best alternative to clean energy technologies if obtained by solar water splitting. Photoelectrochemical (PEC) water splitting is the most viable means of generating H2 from solar radiation as it is sustainable, renewable, and eco-friendly [1, 3]. 43% of solar irradiation on the earth’s surface is in the visible range and the need for semiconducting materials that can absorb light in this range has been a focal research question for decades [4].

For practical PEC water splitting applications, a benchmark of 10% solar-to-hydrogen (S.T.H) efficiency is required from a semiconductor photo-anode [1, 5, 6]. Hematite’s theoretical S.T.H efficiency is projected at 15%, a value above the benchmarked efficiency [1, 5, 7,8,9]. Hematite is cheap, has a desirable band gap (~ 2.1–2.2 eV), is environmentally friendly, chemically stable in alkaline solutions, and non-toxic [2, 5, 7, 10, 11]. This has led to hematite being a focal research material for PEC applications [12,13,14,15,16].

However, until now, hematite has neither reached the 10% practical benchmark nor the maximum 15% efficiency. This could be due to poor electrical conductivity, the short hole-diffusion length (2–4 nm) in relation to the depth of light penetration (~ 120 nm), high photo-generated charge recombination at the surface and in the bulk due to short charge carrier life-time (~ 3–10 ps), slow oxygen evolution kinetics, inefficient separation of electron–hole pairs due to low carrier mobility (< 0.2 cm2V−1 s−1) [1,2,3,4,5, 7, 10] and a band edge mismatch with the redox potentials [2, 8, 17,18,19].

Nano-structure designs, ion doping, metal-oxide integration to form heterojunction photo-anodes, use of oxygen evolution catalysts as interfaces with hematite, forming composite structures, and surface modifications have been carried out to solve challenges concerning hematite [1,2,3, 7, 8, 18, 20,21,22,23,24].

Hematite belongs to a afamily of metal oxides and extensive work has been done to improve the performance of metal oxides for PEC devices. In this regards, song et al. assessed the PEC solar water splitting efficiency of tandem PEC cell configuration using W: BiVO4/CoPi as a working photo-electrode and CuBi2O4/CdS/TiO2/RuOX as the counter electrode highlighting the significance of a cocatalyst in hydrogen production and photo-voltage in tandem PEC cell configurations [25]. Other related work by the same research group focused p-Cu2V2O10 and yielded a photocurrent of 0.5 mAcm−2, they highlighted the importance of inserting a thin layer of Cu doped NiO between FTO substrate and CuBi2O4 photoanode which yielded a photocurrent of 2.83 mAcm−2, they further developed (CuBi2O4/TiO2/RuOx, CuBi2O4/CdS/TiO2/RuOx, CuBi2O4/BiVO4/TiO2/RuOx and CuBi2O4/Ga2O3/TiO2/RuOx) heterojunctions photoelectrodes and their performance was compared based on photocurrents, band offsets, and directed charge transfer [26,27,28,29].

There are various review articles on the the performance of hematite for PEC applications. Choudhary et al. reviewed nanostructure bilayer thin films including; CdSe/CdS, TiO2/CdSe/CdS, TiO2/CdS, TiO2/CdSe, Cu2O/CuO, Cu2O/ TiO2, TiO2/CdTe, WO3/TiO2, Cu2O/ZnO, BiOI/TiO2, TiO2/SnO2, ZnO/SnO2 [30]. The authors highlighted the mechanisms of bilayer photo-anodes for PEC water splitting. It was pointed out that bilayer structures provide proficient charge carrier transfer across the interface between the semiconductors which results in reduced charge recombination. Phuan et al. reviewed the prospects of hematite that is electrochemically synthesized [31]. The authors highlighted parameters that control pristine hematite morphology such as electrolyte precursor, electrode substrate, annealing treatment, and applied potential [31]. Esmaili et al. reviewed the importance of a nanostructure morphology on the performance of a PEC water-splitting cell [32]. The authors pointed out that the electronic structure and morphology of the semiconductor used are key performance parameters in PEC water splitting. Various morphology dimensions of 0-D, 1-D, 2-D, and 3-D were discussed in detail [32]. Saraswat et al. reviewed the synthesis schemes of photo-catalysts that result in better catalytic performance at PEC water splitting and the various corresponding challenges faced in the photo-catalyst architecture build-up [33]. The authors pointed out that, morphology can be controlled by preparation method, calcination temperature, and crystal geometry or structure, which is in agreement with the views of Phuan et al. [31, 33]. Iqbal and Siddique reviewed work on carbon-based photo-anodes used for PEC water splitting such as hematite/Graphene heterojunction nanowires, carbon and Fe2O3 quantum dots, and carbon nanotubes, among others [34]. Raza et al. recently analyzed research work on several 2-D layered photo-anodes and photocathodes [35]. The main focus was on how 2-D materials provide PEC application in energy conversion. Jun et al. recently reviewed the superiority properties exhibited by 2-D materials in comparison to their 0-D, 1-D, and 3-D counterparts [36]. The authors focused on how 2-D materials are synthesized and how they are used for PEC water splitting based on their roles as light harvesters, catalysts, interfacial transporting layers, and as passivation layers. The 2-D materials include; Graphitic carbon nitrides, layered double hydroxides, Mxenes, and transition metal dichalcogenides [35, 36]. Recently, Wan et al. reviewed the effect of electronic structures or compositions on the practical performance of hematite [21]. Li et al. reviewed heterojunctions and their advantages if employed as photo-anodes in PEC water splitting [37]. The different types of heterojunctions such as the type II and Z-scheme were introduced and their physical effects on the enhancement of carrier transfer in a material were discussed. Young and co-workers reviewed the strategies to facilitate charge transfer for enhanced oxygen evolution, reviewed articles on transitional and ternary metal oxides [38,39,40]. Figure 1 provides a summary of the work that has been published on hematite for PEC water splitting obtained from the Web of Science. Figure 1a shows research articles on hematite, Fig. 1b shows the review articles on hematite, Fig. 1c shows research articles on hematite hetero-structures and Fig. 1d shows the review articles on hematite hetero-structures. In general, the graphs depict an increasing trend for hematite as a potential PEC material. Despite the increasing trend, there are very few review articles that have been published focusing on hematite and its hetero-structures. In addition, most of the published review articles lack scientific rationale on how the measured properties are affected by the electronic properties such as electrical conductivity and charge carrier mobility of the hetero-structure/substrate. This review offers the latest developments and provides further insights into the different hematite hetero-structures in relation to their PEC performance. This review article is organized as follows: a brief PEC background, PEC performance parameters, types of PEC device configurations, and the physical and crystallographic properties of hematite. The review further discusses the performance of hematite and its hetero-structures on Fluorine-doped Tin Oxide (FTO), Sn-doped In2O3 (ITO), and other substrates.

Fig. 1
figure 1

A comparison of published work for PEC water splitting from 2010 to date; a research articles on hematite, b review articles on hematite, c research articles on hematite hetero-structures, and d review articles on hematite hetero-structures

2 Principle of operation photoelectrochemical water splitting device

The first PEC water-splitting device was created by Funjishima and Honda in 1972 and used TiO2 (the rutile phase) and Pt as photo-anode and photocathode respectively to convert UV irradiation into hydrogen [5]. Figure 2 shows the PEC system with two reference energy levels, normal hydrogen energy (NHE) and vacuum level (Fig. 2a). PEC system before equilibrium in the dark (Fig. 2b), the PEC system at illumination (Fig. 2c), and the PEC system after illumination (Fig. 2d). When the semiconductor is immersed in the electrolyte, a solid–liquid interface is created between the semiconductor and the electrolyte. As a result, electron transfer takes place between the working electrode, the electrolyte, and the counter electrode until the Fermi levels of the semiconductor, the counter electrode, and the electrolyte line up as shown in Fig. 2b [41,42,43]. Electrons will flow from the photo-anode (higher Fermi level) to the electrolyte and then to the counter electrode (lower Fermi level). After the Fermi levels have line-up, the semiconductor will have an excess of positive charge, arising from the immobile ionized impurity atoms, and the electrolyte will have an excess of negative charge. As a result of the movement of electrons, a depletion layer with width ‘‘W’’ is created in the semiconductor side and the negative charge accumulates over a much narrower region called the Helmholtz layer between the electrode and the electrolyte as shown in Fig. 2 c. On illuminating the semiconductor with photon energy greater than the band gap of the semiconductor, electron–hole pairs will be generated. The electrons will be excited to conduction band, leaving the holes in the valence band. As shown in Fig. 2d, a quasi-Fermi level is established whose gradient results into the built-in electric field or the open-circuit voltage (VOC). The electrons are transferred to the counter electrode to generate hydrogen. The holes will diffuse to the surface of the semiconductor to generate oxygen [3, 43]. For water splitting to take place, the semiconductor’s valence band (V.B.) potential must be more positive than the EOER redox of 1.23 V vs NHE to allow oxidation, whilst the conduction band (C.B.) must be more negative than the EHER redox potential of 0 V vs. NHE to carry out water reduction. This is in addition to overpotentials. The overall reaction of water splitting is illustrated by Eq. 1, and the hydrogen evolution reaction (HER) and oxygen evolution reaction (OER) are illustrated by Eqs. 2 and 3 respectively.

Fig. 2
figure 2

a Electrochemical energy levels, b energy band diagram of a PEC system before equilibrium, c energy band diagram of PEC system at equilibrium in the dark condition and d energy band diagram of PEC system under illumination

Combined PEC reaction;

$$2H_{2} O + 4hv \to O_{2} + 2H_{2}$$
(1)

Reaction at the anode;

$$4H^{ + } + 4e^{ - } \to 2H_{2}$$
(2)

Reaction at the cathode;

$$2H_{2} O + 4\left( {h^{ + } } \right) \to O_{2} + 4H^{ + }$$
(3)

The overall reaction requires the Gibbs free energy, ΔrGo of 273.2 kJ/mol [44]. In this connection, photons whose energy is equal to or greater than the band gap energy of the semiconductor material will be exited [34, 44]. Theoretically, the band gap should be larger than 1.23 eV because the difference between HER and OER is 1.23 eV. Practically, an energy gap between 1.5–1.8 eV is necessary due to HER/OER over potentials. The PEC water splitting process takes three steps, viz; absorption of light, separation and transportation of photo-generated charge carriers, and surface reduction as indicated in Fig. 2b [34, 44, 45].

2.1 The most common PEC device configurations in literature for Hematite

2.1.1 Single semiconductor photoanode light absorber configuration

This configuration is similar what has been explained in Fig. 2. When an n-type semiconductor (whose work function is lower than the work function of the metal) and an electrolyte come into contact, electrons flow from the semiconductor toward the counter electrode via the external circuit well as holes are transferred into the electrolyte until equilibrium is reached. At this point, the semiconductor/electrolyte interface is depleted of electrons compared to the electrons in the semiconductor bulk hence a depletion layer forms and the bands bend upwards as shown in Fig. 3a [46]. The electrons take part in water reduction and the holes take part in the water oxidation [47, 48]. This kind of configuration creates barrier potential, that leads to improved charge separation [36, 47, 49, 50]. The electrons then take part in HER well as the holes take part in the OER. Hematite has been used as a single junction photo-anode for PEC applications.

Fig. 3
figure 3

Schematic band diagrams illustrating a single n-type semiconductor photo-anode/metal counter-electrode, b semiconductor heterostructure photo-anode/metal counter electrode, c with bias, and d tandem PEC cell configuration without

2.1.2 Heterojunction semiconductor photoanode light absorber configuration

In this configuration, secondary semiconductors are developed onto the primary absorber to improve light absorption and charge separation [46]. In this configuration, the top layer will absorb high-energy photons but allow low-energy photons to go pass through as illustrated in Fig. 3b [51]. The low-energy photons will be absorbed by the subsequent layer hence improving photo-absorption. For this configuration to form, band gap variation and lattice matching are key [52]. Hetero-structuring provides properties that complement each semiconductor such as increased spectral range, enhanced interfacial charge carrier mobility, strong interfacial built in electric field and thus effective interfacial charge carrier transfer[53,54,55,56].

2.1.3 Series connected photoanode and photocathode (tandem)

In this configuration, two light absorbers are utilized as a Photoanode for water oxidation and a photocathode for water reduction. This configuration is subdivided into wired (operates under bias) and wireless configurations (operates without bias). One conductive metal wire can be used for the wired configuration and a transparent conductive substrate can be used as an electron–hole recombination layer for the wireless configuration. The separation membrane separates O2 from H2 [57]. Tandem cell configurations without bias minimize on the challenges of overpotential and Ohmic losses faced by single light absorber PEC devices [58]. The tandem cell in this configuration must have a photo-voltage at the photo-anode and photo-cathode that exceeds HER and OER potentials for overall water splitting (> 1.23 V) [47, 59]. Figure 3c shows an n-type photo-anode with a p-type photo-cathode connected by an external cicuit when operating with bias. On illuminating the photo-anode with photon energy greater than its band gap enegy, electron–hole pairs are photogenerated and separated, the electrons flow to the C.B. from the V.B. and via the external circuit to the photo-anode’s C.B. well as the holes drift to the photo-anode surface [58]. Figure 3d indicates the operation of a Tandem PEC cell without bias where on irradiation, a space charge field is set up enabling separation of the photogenerated electron–hole pairs where the p-type photo-cathode electrons drift to the semiconductor/electrolyte interface to take part in the reduction process well as the holes in the n-type photo-anode drift to the semiconductor/electrolyte interface to take part in the oxidation process [25, 60].

3 Properties of hematite

3.1 Crystallography

Hematite has rhombohedral lattice structure with a space group of \(R\overline{3 }c\). The lattice parameters at room temperature are a = b = 0.5034 nm, c = 1.375 nm, and γ = 120° [4, 5, 21, 37, 61,62,63]. The XRD peaks are indexed at (110), (101), (200), and (211) at 2θ values of 26.5°, 33.5°, 37.5°, and 51.5° respectively [64,65,66,67]. Its space group is P42/mnm with lattice parameters, a and c of about 4.74 Å and 3.19 Å respectively [65, 68]. Hematite when deposited on an FTO substrate gives peaks at 2θ ~ 24.1°, 33.1°, 35.6°, 40.9°, 49.4°, 54.0°, and 64.0° corresponding to (012), (104), (110), (113), (024), (116), and (300) planes respectively [69]. Of these, the peaks at (104) and (110) are predominant well as other FTO peaks are suppressed—an indication that the film has water in it and so the Fe3+ ions in the hematite film interact with the hydroxyl groups [66, 70]. FTO peaks may not be observed on the XRD chart if the film grown on them is thick, though when observed for such films, it is an indication that all water has been removed from the film and the extent of water removal is depicted in the intensity of the observed peaks [70, 71]. Hematite on FTO has a lattice parameter, a of 5.04 Å, and c as 13.74 Å (Fig. 4).

Fig. 4
figure 4

(Reproduced with permission from reference [72], copyright 2014, AIP publishing)

A schematic diagram showing the crystal structure of hematite where the large circles indicate Fe ions and the small circles indicate the O ions

3.2 Optical properties

Hematite has a band gap that ranges between 1.9–2.2 eV and absorbs both partial visible light and ultraviolet light [7, 9, 21, 60, 73]. The photon penetration depth ranges between 100–120 nm at a wavelength of about 550 nm [74, 75]. This results in the generation of charge carriers relatively far from the hematite/electrolyte interface, hence not contributing significantly towards the water oxidation reaction. Also, at a wavelength of about 550 nm, hematite experiences low absorptivity hence the preference of synthesis of thick hematite films so as to absorb a sufficient amount of the sunlight for optimum PEC performance [9].

3.3 Electrical properties

Hematite is an n-type semiconductor with an electrical conductivity approximated at about 10–4–10–6 Ω−1 cm−1, charge carrier mobility of approximately 10–2 cm2V−1 s−1, and a carrier decay rate of < 1 ps [3, 5, 9, 31, 75,76,77].

4 Surface morphology of hematite on different substrate

1 D Nano-porous films, Nano flakes, Nano-tubes, or Nano-wires of hematite have been formed on an Fe foil, 3 D hematite like-Nano-flower- films have been formed on an FTO glass substrate as shown in Fig. 5a–d [10, 19, 31, 78]. 3-D structure arrays like Nano-spikes, Nano-wells, Nano-cones, and Nano-pillars enlarge the materials’ surface area and hence enhance light absorption [79, 80].

Fig. 5
figure 5

(Reproduced with permission [10] copyright 2020, Elsevier)

SEM images showing the morphology of hematite on different substrates: a a 3 D flower-like film of pristine hematite on FTO glass substrate (Reproduced with permission [19] copyright 2016, Elsevier), b a porous morphology of Ti-Fe2O3/Cu2O heterojunction on an ITO glass substrate b (Reproduced with permission [78] copyright 2014, Elsevier B.V), c nano-flakes of pristine α-Fe2O3 sintered at 400 °C on an Fe foil substrate, d a coral-like morphology is obtained when sintered at temperatures above 600 °C

5 PEC performance of hematite on different substrates

Different substrates have different physical properties. Fluorine-doped tin oxide (FTO) and tin (Sn)-doped In2O3 (ITO) are the commonly used TCOs [81].

5.1 Physical properties of FTO glass substrate

FTO is an n-type semiconductor with a band gap of 3.8 eV [82] and a work function of 4.8–5.0 eV vs vacuum (0.3–0.5 eV vs R.H.E) [28, 30].The optical transmittance ranges between 60–90%, chemically and thermally stable, and has a high resistance to physical etching [67, 83]. FTO has a resistivity of about 3.5 × 10–4 Ωm, charge mobility of about 25 cm2V−1 s−1, and a carrier density of about 4.5–6 × 1020 cm−3 for thickness greater than 500 nm [84]. If doped with 2.5% F concentration, the electrical resistivity is about 6 × 10–4 Ωcm and a carrier density is of about 1.84 × 1020 cm−3 [65].

5.2 PEC performance of hematite on FTO substrates

In this section, we examine the performance of various morphologies of pristine and heterojunction hematite on FTO substrates.

5.2.1 FTO/Hematite/electrolyte/Pt configuration

Farhoosh et al. synthesized a Fe2O3/FTO photo-anode of nano-rod morphology via hydrothermal synthesis obtaining a 3.34 × 1019 cm−3 donor density, a 0.5 V flat band potential and a less than 1% IPCE value with a 0.003% ABPE value taken at 1.1 V vs RHE [1]. Zhang et al. fabricated a bead array α-Fe2O3/FTO photo-anode and only managed to obtain a 0.46 mAcm−2 photocurrent density, a 0.461 V flat band potential with an IPCE value of 7.2% [85]. This shows very low performance exhibited by hematite if used in a single photo-absorber PEC device configuration. Table 1 shows a number of study findings for FTO/Hematite/Electrolyte/Pt PEC device set ups.

Table 1 Work done on pristine hematite developed on FTO glass substrate

5.2.2 FTO/hematite heterojunction/electrolyte/Pt configuration

The formation of heterojunctions provides improved length of hole diffusion and lead to reduced charge recombination in the photo-electrode. This results in improved donor density and photocurrent density hence better overall conductivity of the photo-electrode.

The electrical conductivity of a semiconductor is related to carrier concentration (n), charge (e) and carrier mobility (\(\mu )\) by

$$\sigma = ne\mu$$
(4)

and thus the increase in donor density increases the electrical conductivity. Likewise, photocurrent density (J) is related to electrical conductivity and electric field (\(\overrightarrow{E}\)) by

$$J = \sigma \vec{E}$$
(5)

This equation emphasizes that an increase in electrical conductivity increases the current density and thus the observed effects [98]

Where multiple junctions have been studied, it is noted that these enhance charge carrier photo-generation at the depletion layer leading to increased PEC performance of the heterojunction hematite photo-electrode. For instance, Farhoosh et al. synthesized Ti–Fe2O3/Fe2TiO5/FeOOH/FTO photo-anode through hydrothermal synthesis realizing a photocurrent density of ~ 1.63 macm−2 [1]. Fe2TiO5 and α-Fe2O3 have similar band gaps (2.1 eV for both) and were arranged with hematite as the under-layer onto an FTO substrate to form an n–n junction. As shown in the schematics of Fig. 6a(i), on illumination through the FTO side, electrons drift across the α-Fe2O3/FTO junction towards the Pt counter electrode via the external circuit while the holes drift across the Fe2TiO5/α-Fe2O3 junction to take part in the water oxidation process at the photo-anode/electrolyte interface. The Fe2TiO5 layer improved on the electrical conductivity by acting as a charge separator between the Ti-Fe2O3 and FeOOH layers.

Fig. 6
figure 6

Images of schematic diagrams showing hetero-structure build-ups and their corresponding photocurrent density comparisons with pristine hematite photo-anode performances on FTO glass substrates in photo-absorber photoanode/metal counter electrode configuration, (Adopted with permission a [1] copyright 2021, Elsevier B.V, b [87] copyright 2015, ACS Publications, c [18] copyright 2017, Elsevier, d [8] copyright 2019, ACS Publications, e [7] copyright 2013, RSC Publishing, f [2] copyright 2017, ACS Publications, g [88] copyright 2021, Elsevier, h [78] copyright 2014, Elsevier B.V, i [69] copyright 2017, IOP Publishing, and j [99] copyright 2018, Elsevier B.V

For photo-generation of charge carriers to be effective, the C.B. and V.B. edges must suitably align. Therefore, band bending is inevitable. Here, the different Fermi levels of the semiconductors align through adjustment of the energy bands once the semiconductors come into contact. This results in a single Fermi-level for the hetero-structure as shown in Fig. 6(i) for each case. In this case, the positioning of the C.B. and V.B. is key. The electrons drift from a higher C.B. to a lower C.B. well as holes drift from a lower V.B. to a higher one. The drift through more junctions helps in improving the length the holes drift through to hit the photo-anode/electrolyte interface to take part in water oxidation. Thus better photocurrent density in the heterojunction photo-anode is obtained as indicated in Fig. 6(ii) for each case. With these band edge positions in mind, for instance, Fig. 6 (i), (b) with p-CaFe2O4/n-Fe2O3 at a band gap of 1.9 eV and 2.1 eV respectively, (d) with FeNbO4 at 1.91 eV and Fe2O3 at 2.1 eV, and (e) with SnO2 at 3.8 eV and Fe2O3 at 2.1 eV, it is noted that the built-up hetero-structures used the semiconductor with a higher band gap as the under-layer on FTO substrate, well as (g), (h), and (i) indicate the reverse. With p-CaFe2O4/n-Fe2O3, a p–n junction photo-anode was formed where on illumination, holes in the n-Fe2O3 were reported to be excited into the V.B. of p-CaFe2O4 well as electrons were transferred from the C.B. of p-CaFe2O4 to the C.B. of n-Fe2O3 to reach the back contact as shown in Fig. 6b(i) [87]. The presence of the heterojunction in p-CaFe2O4/n-Fe2O3/FTO photo-anode allows for efficient electron–hole pair separation at the heterojunction. With the presence of a Co-Pi, hole capturing at the photo-anode/electrolyte interface resulting in better charge separation hence reducing their accumulation which usually results in recombination experienced in the n-Fe2O3/FTO photo-anode. In Fig. 6f(ii), with p-LaFeO3/n-Fe2O3/FTO p–n junction photo-anode, effective charge separation at the p–n junction was realized giving a photocurrent density of 0.58 mAcm−2 which was as well not impressive [2]. Here, during band edge matching (band bending at the interface), the V.B. of LaFeO3 is more negative than that for Fe2O3 which results in reduced oxygen evolution driving force at the junction. This in addition gives slow reaction kinetics at the interface hence low performance even with the heterojunction. Applying a CoOx layer (catalyst) on the p-LaFeO3 surface to form CoOx/p-LaFeO3/n-Fe2O3/FTO results in a better photocurrent density value recorded at 1.12 mAcm−2. From Fig. 6g(i), in the heterojunction photo-anode (MoS2/BiVO4/α-Fe2O3/FTO) electrons move from the C.B. of MoS2 to that of BiVO4 then to the C.B. of α-Fe2O3 through the FTO substrate then via the external circuit to the counter-electrode, while the electrostatic field drives the holes from the V.B. of α-Fe2O3 to that of BiVO4 and then to that of MoS2 this takes part in the oxidation reaction [88].

Tremendously improved results for hematite PEC performance are realized once more than one strategy is applied in the formation of a hematite photo-anode. For instance, Sharma et al. strategized with doping Fe2O3 with Ti and then hetero-structuring it with Cu2O and gained a big improvement in the performance of hematite at PEC water splitting [78]. An IPCE value of 6.5% for the doped hematite hetero-structure (Cu2O/Ti-Fe2O3/FTO) with respect to 0.18% for pristine hematite photo-anode were recorded. Also a photo-anode of FeNbO4/Fe2O3/FTO, Sn and Nb doping into Fe2O3 at the Fe2O3/FTO and FeNbO4/Fe2O3 interfaces respectively developed by Zhang et al. results in better charge carrier density in the doped heterojunction photo-anode relative to the pristine hematite photo-anode that is at 1.77 × 1019 cm−3 and 0.81 × 1019 cm−3 respectively [8].

In another recent study, Zhang et al. fabricated a Voα-Fe2O3/AuNSs/FTO heterojunction photo-anode realizing a better photocurrent density of 1.46 mAcm−2 in relation to the 0.46 mAcm−2 from the α-Fe2O3/FTO photo-anode [85]. This improvement was attributed to improved bulk conductivity and exhibition of surface electrocatalytic ability by the heterojunction photo-anode. Here, charge recombination is suppressed due to the strong AuNSs and α-Fe2O3 interface contact which improves on band bending in addition to the oxygen vacancy (Vo) presensence reducing on the activation energy provided by increased holes. Table 2 shows the various hematite heterojunction photo-absorber PEC performances.

Table 2 Work done on hematite heterojunction photoabsorber for PEC water splitting

5.2.3 FTO/hematite tandem configuration

Vilanova et al. designed a compact optimized open light PEC cell (CoolPEC cell) of a tandem configuration in which hematite was used as a working-electrode, an anion as the exchange membrane, and platinized-Ti mesh as a counter-electrode [100]. Using cyclic holders, the platinized-Ti mesh was held firmly to the membrane and the hematite photo-anode to reduce the Ohmic losses at the contact points resulting in a 0.45 mAcm−2 photocurrent density. The test was done using a 2-electrode configuration under 1000 Wm−2 solar simulation. According to Landman et al., this single-cell configuration is efficient for small-scale hydrogen production rather than large-scale due to the high maintenance and membrane costs involved [101]. Therefore, the researchers recommend setting up separate cell configurations for oxygen and hydrogen production known as the PEC-PV configuration where a membrane is not necessary since the two gases will be produced in different cells. Gurudayal et al. examined the performance of pristine hematite in tandem connection with a mesoporous perovskite solar cell as shown in the schematic in Fig. 7a recording a photocurrent density of 1.33 mAcm−2 shown in Fig. 7b with a fill factor of the perovskite solar cell at 0.62, and a STH efficiency of 1.5 [102]. in another study, Gurudayal et al. modified hematite photoanodes with SnOx via hydrothermal synthesis followed with the atomic layer deposition technique [103]. These modified Sn-doped Fe2O3 Nws photoanodes were set-up in tandem with a CH3NH3PbI3 perovskite solar cell realizing a 4.5 mAcm−2 photocurrent density and a 3.4% STH efficiency. Table 3 shows studies on hematite in a Tandem PEC cell set-up.

Fig. 7
figure 7

(Adopted with permission [102] copy right 2015, NANO letters)

a Image of a schematic showing a PEC-PV tandem cell set-up, and b corresponding performance at PEC water splitting

Table 3 Work done on FTO/Hematite in Tandem PEC cell configuration

5.3 Physical properties of ITO glass substrate

An ITO glass substrate is an n-type semiconductor with a relatively wide band gap energy approximated at 3.5–4.3 eV [82, 104]. ITO glass substrates bear good optical properties, that is an optical transmittance of about 80–90% [67]. An ITO glass substrate exhibits a cubic structure of the Ia-3 space group with lattice parameter a ~ 10.130 Å [64, 65, 105, 106] and with a crystalline grain morphology [65, 107]. For a thickness above 100 nm at Sn/In concentrations of 2.4%/4.8% respectively, an ITO glass substrate has an electrical resistivity of about 3–4 × 10–4 Ω, a carrier density of about 4–10 × 1020 cm−3 and a mobility of about 35 cm2V−1 s−1 [84]. With a Sn concentration at 6%, the electrical resistivity is about 8 × 10–4 Ω, and the carrier density is about 3.67 × 1020 cm−3. This increased resistivity is attributed to high free electron concentrations due to substitutional Sn and oxygen vacancies in addition to ITO having a cubic structure [65]. The mobility of ITO can be reduced to about 20 cm2V−1 s1 at temperatures above 400 °C.

5.4 PEC performance of hematite on ITO

Riha et al. and Kaouk et al. agree that annealing at such high temperatures results in reduced conductivity of hematite on ITO substrates since such temperatures result in increased resistance of the hematite/ITO photo-anode [108, 109]. Here, more Sn atoms in the crystal are localized on the interstitial lattice sites, or a radius mismatch is observed which results in distortion of the crystal structure. Also, such temperatures reduce the oxide layers into metallic indium layers [84]. This results in localized Sn atoms acting as ionized scattering centers.

5.4.1 ITO/hematite/electrolyte/Pt configuration

According to Zandi and Haman, for efficient water oxidation, the photo-anode used should be of an n-type semiconductor since this provides efficient band bending for charge carrier separation [110]. In addition, He et al. suggest that homo-junctions can be formed to improve photon absorption [111]. In this regard, hematite/ITO build-ups produce a photo-anode of the n/n-type homo-junction since both hematite and the ITO substrate are small band gap and wide band gap n-type semiconductors respectively. The expectation here is that when hematite is grown on ITO, a balance of electrons and holes will take place at the interface hence a space charge zone will be created and an electric field will flow [111]. For instance, Yu et al. fabricated hematite homo-junctions with an ITO substrate by oxygen vacancy control and concluded that hematite bilayer films increased carrier concentrations in hematite resulting in modulation of the Fermi-level of hematite leading to the formation of the n–n hematite/ITO homo-junction [112].

Band gap energy is one of the fundamental materials’ properties that determines the portion of solar spectrum absorbed by a semiconductor photo-absorber hence controlling its photo conversion efficiency [113]. The grain size for hematite nanoparticles is about 29 nm from HRSEM micrographs with a microstrain ~ 0.3492 × 10–3 from XRD peaks well as its band gap is increased to about 2.41 eV (outside the theoretical band gap range for hematite of 1.9–2.1 eV) when deposited on an ITO substrate by spin coating for 6 min in a study conducted by Kumar and Yadav—an indication of the presence of the Fe3+ ions in a low spin state [70]. On illumination, the absorbed photons excite electronic transitions though some do not lead to the generation of electron–hole pairs for instance if the d-d ligand transition is excited [113]. With this good optical property of hematite, Wang et al. also performed a theoretical optical simulation assuming an air mass of 1.5 solar spectra with a 20 nm film thickness (which is almost the hole diffusion length in hematite) and managed to derive the theoretically expected maximum photocurrent density of 12.5 mAcm−2 for hematite grown on an ITO Nano-cone substrate [114].

Landolo et al. also carried out a study of α-Fe2O3 deposited on ITO and FTO by physical vapor deposition with heat treatment at 350 °C for a period ranging from 0 to 24 h [115]. For heat treatment above 12 h at 350 °C, the hematite/ITO samples showed negligible photo activity attributed to electrical conductivity loss by the ITO substrate under prolonged heat exposure. A more cathodic potential shift was observed as the time for heat treatment increased from 0, 3, 6, and 12 h and were recorded as 1.28, 1.19, 1.13, and 1.08 V vs RHE respectively. In addition, an increased donor density for hematite/ITO at 12 h was recorded at 2.52 × 1018 cm−3 compared to that at 2.28 × 1018 cm−3 for 0-h exposure [115]. A higher donor density implies reduced photo-generated charge recombination [116]. When the same experiment was carried out on FTO substrates, a similar cathodic potential shift was observed for increased heat treatment time 0 to 12 h of exposure [115]. This indicates that the substrate type does not affect the observed onset potential obtained for hematite thin films. Table 4 shows the various findings of research on ITO/pristine hematite photo-anodes.

Table 4 Work done on ITO/ hematite/Electrolyte/Pt Configuration

5.4.2 ITO/hematite heterojunction/electrolyte/Pt

The annealing treatment given to a hetero-structure of hematite on ITO substrate has an impact on the properties realized of the heterojunction formed. In a study conducted by Palmolahti et al., the effect of annealing on TiO2/α-Fe2O3/ITO material was carried out [119]. XPS results after annealing indicated that the Fe 2p3/2 binding energy peak shapes were unchanged due to the presence of Fe3+ oxidation states. On the other hand, the Ti4+ 2p3/2 binding energy peaks showed a − 0.31 eV shift with reduced width size after annealing. The concentration of Ti at annealing temperatures above 500 °C was reduced well as that for In and Sn was increased—an indication of diffusion of Ti in the hematite top layer resulting in the formation of Fe3+ Ti4+ surface oxide. The researchers argue that much as In and Sn diffuse into hematite at high annealing temperatures, their binding energy peaks corresponding to In 3d5/2 and Sn 3d5/2 are unchanged—an indication of a stable chemical environment of hematite/ITO. In and Sn can diffuse from ITO into hematite at high annealing temperatures (above 500 °C) thus oxygen vacancies can be created on the surface of hematite [119, 120, 122]. Table 3 shows the various hetero-structure performances realized with hematite on ITO glass substrates.

Diab and Mokari managed to grow a uniform layer of magnetite (Fe3O4) on an ITO glass substrate by thermal decomposition of Fe(ac)z in an inert environment at 270 °C for 15 min sintered at 400 °C for 5 h followed by annealing at 700 °C for 12 min hence converting the Nanoporous magnetite into hematite (Fe2O3) mesoporous film [17]. In the study, modified hematite/CoO films were also synthesized by thermal decomposition followed by drop casting the mesoporous hematite films with Co(NO3)2 solution. A photocurrent density of 0.11 mAcm−2 at 1.82 V vs RHE was recorded for mesoporous hematite without any surface modification with an offset potential of 1.6 V vs RHE. Photocurrent densities of 1.17 mAcm−2 and 0.66 mAcm−2 were recorded for CoO-modified hematite and Co(NO3)2 modified hematite films respectively. The CoO-modified hematite film also exhibited a cathodic onset potential of 130 mV at 1.82 V vs RHE well as the Co(NO3)2 modified hematite film showed an onset potential shift of 70 mV at 1.82 V vs RHE which shows a significant improvement attributed to surface trap passivation in the modified films [17].

In another study, the TiO2/Fe2O3/ITO hetero-structure, its intermediate surface states (i-ss), and recombination surface states (r-ss) were studied at different annealing temperatures (300–700 °C) [119]. At an annealing temperature of 300 °C, the least recombination of free charge carriers was recorded in addition to having the highest population of r-ss at a delay time of 1 ns hence resulting in better PEC performance as shown in Fig. 8a (ii). The researchers also highlighted that the TiO2 layer on hematite modifies its r-ss and hence improves on charge carrier life-time thus more will take part in the reaction processes for water splitting. Yu et al. compared the performance of pristine hematite on an ITO with that when an FeOOH layer is deposited on hematite forming FeOOH/Fe2O3/ITO photo-anode [121]. Here, a photocurrent density of 217 μAcm−2 for Fe2O3/ITO photo-anode was recorded which turns into 850 μAcm−2 when decorated with FeOOH as a co-catalyst to form FeOOH/Fe2O3/ITO as indicated in Fig. 8b (ii). This improvement is attributed to the small resistance to charge flow offered at the FeOOH-electrolyte interface which results in a positive effect to charge transfer at the interface. This is because the charge transfer barrier is reduced hence enhancing PEC performance. In another study, a CZTS p-type semiconductor with α-Fe2O3 an n-type formed a p–n junction as shown in Fig. 8c (i) after band bending to match the Fermi levels of the semiconductors which led to the generation of an electric field that enables the charge carrier separation process [120]. The graph shown in Fig. 8c (ii) shows a comparative performance of the pristine Hematite/ITO (H), Hematite-Graphene/ITO (HG), CZTS QDs/Hematite/ITO (HC), and CZTS QDs/Hematite-Graphene/ITO (HGC) thin films at water splitting. High vacancy and low vacancy hematite were built to form the HV/LV-Fe2O3 junction where an electric field forms enabling accumulation of holes on the LV-Fe2O3 side which take part in the oxidation reaction at the Ohmic contact between the ITO substrate and the electrolyte hence the realized better PEC performance of the HV/LV-Fe2O3/ITO photo-anode compared to the LV-Fe2O3/ITO and the HV-Fe2O3/ITO photo-anodes as shown in Fig. 8d (ii) [112]. Table 5 shows other studies carried out on hematite heterostructures set up in an ITO/Hematite heterojunction/Electrolyte/Pt configuration.

Fig. 8
figure 8

(Adopted with permission a [119] copyright 2020, ACS Publications, b [112] copyright 2016, Royal Society of Chemistry, c [120] copyright 2022, Elsevier, d [121] copyright 2015, Willey)

Images of schematic diagrams showing hetero-structure build-ups and their corresponding photocurrent density comparisons with pristine hematite photo-anode performances on ITO glass substrates in photo-absorber photoanode/photo-absorber photocathode configuration

Table 5 Various works done on ITO/Hematite heterojunction/Electrolyte/Pt configuration

5.4.3 ITO/hematite tandem configuration

A FeNiOOH/Fe2TiO5/Fe2O3/ITO photo-anode was built for use in a Tandem PEC cell set-up as shown in Fig. 9a [123]. Here, the thin film silicon photocathode together with the Ni-based catalyst powers the Tandem set-up. No external bias was used in the set-up yet a photocurrent density of 2.2 mAcm−2 was realized as indicated in Fig. 9b. Table 6 shows the PEC performance realized from the study’s Tandem configuration.

Fig. 9
figure 9

(Adopted with permission [123] copyright 2019, Willey)

Images of schematic diagrams showing a a hetero-structure build-up and b the corresponding photocurrent density comparisons with pristine hematite photo-anode performances on ITO glass substrates in tandem cell configuration

Table 6 Various works done on hematite hetero-structures in an ITO/Hematite Tandem Configuration

5.5 PEC performance of hematite on other substrate types

Noble metal films like Pt films are recommended for use as substrates in times when high temperatures in addition to long sintering time, however, optical properties such as transparency may be affected [81]. These affect the nanostructure formed of hematite on deposition and hence affect the PEC performance of the structure formed. In a study conducted by Koziot-Rachwal et al. on the magnetic properties exhibited by hematite/Pt (111) hetero-structure, hematite was grown onto a Pt (111) seed layer by oxidation of magnetite (Fe3O4) (111) films [124]. Here exposure of Fe3O4 layers to oxygen at a pressure and high temperature (3 × 10–5 mbar and 770 K respectively) was done during the annealing process for 30 min after which magnetite turned into hematite (Fe2O3).

Fe substrates like Fe rods and Fe foils are also used as substrates over which hematite is deposited [31]. When an iron substrate was used, Nano-flakes were grown onto the iron substrate. On introducing Pt and FeOOH, a hetero-structure photo-anode of FeOOH/α-Fe2O3 NFs/Pt/iron substrate was formed which resulted in a cathodic photocurrent density of − 2 mAcm−2, a better photo-response compared to the α-Fe2O3 nFs/iron substrate photo-anode as shown in Fig. 10a (ii) [125]. In a CuO/α-Fe2O3/Fe-foil photo-anode, other than changing the surface morphology of the photo-anode from Nano-flakes to a coral-like morphology, the presence of a CuO layer forms a p–n junction resulting in an improved photocurrent density to a tune of 1.5 mAcm−2 from 0.7 for pristine hematite grown on a Fe-foil at a 1.6 V applied potential as shown in Fig. 10b (ii) [10].

Fig. 10
figure 10

(Adopted with permission a [125] copyright 2017, RSC Publishing and b [10] copyright 2020, Elsevier)

Schematic diagrams showing performances of hematite photo-anode hetero-structure and their corresponding photocurrent density in comparisons with pristine hematite on other substrates

Table 7 illustrates the different research works done on hematite hetero-structures on other substrates.

Table 7 Various works done on pristine hematite and hetero-structures on other substrates

6 Conclusion and outlook

The single photoanode light absorber and the heterojunction semiconductor photoanode light absorber configurations are the commonly used PEC device configurations in hematite studies. Pt is the common counter electrode used citing its electrocatalytic property which helps in reducing mass-transport over-potential [128]. All PEC device configurations require Ohmic contacts where the electronic junction is formed [129]. Hematite’s PEC water-splitting performance is still low due to its poor C.B edge match with the H2O reduction potential. Much as the series connected photoanode and photocathode (Tandem) PEC device configurations enable bias-free PEC cell operation, challenges of limited light absorption and low current density are still hindering their performance [36]. From the reviewed studies, the best photoanode light absorber configuration performance is at a photocurrent density of 3.56 mAcm−1 with a heterojunction photo-anode of H:TiO3/Fe2O3/ITO, and 4.2 mAcm−1 for Tandem configuration obtained from μc-Si:H/μc-Si:H/α-Fe2O3/ITO which is bias-free [118, 123]. Formation of a homo-junction between hematite and ITO leads to improved absorption of the sun’s photo energy [111]. Formation of a heterostructure increases the donor density and leads to positive shift in the flat band potential and results in band bending at the heterojunction interfaces improving on the charge separation [1]. Also surface catalysis exhibited by heterojunction photo-electrodes enables the material to perform better at charge separation in the bulk as well as increases on its charge carrier concentration [85]. The best Hematite/FTO performance is at a photocurrent density of 1.02 mAcm−1 in Tandem configuration, 1.04 mAcm−1 for Hematite/ITO in single photoanode light absorber configuration, and 2.8 mAcm−1 for Hematite/Ceramic in single photoanode light absorber configuration [18, 120, 127]. An indication that substrates affect the hematite’s PEC performance in different PEC device configurations.

Conducting substrates are used to form PEC photo-electrodes and preference is given to TCOs since these enable creation of an Ohmic contact which enhances transportation of charge carriers—the reason for the common use of FTO and ITO substrates [81]. Transparent substrates also provide for both back-side and front-side illumination. Front-side illumination reduces on the length the holes have to diffuse after photo-activity to take part in the reaction process well as back-side illumination reduce on the electrons’ length traveled without moving through the bulk of the electrode [48]. Therefore, balancing of hematite’s thickness with light absorption to favor particular charge carrier transportation is necessary.

As much as longer annealing time at higher annealing temperatures leads to reduced recombination centers in a hematite electrode, care has to be taken on the substrate used. This is because FTO substrates may deform in addition to Sn4+ cations diffusing into hematite resulting in doping hematite with Sn, well as for an ITO substrate, the electrical conductivity is lost [81].

More studies on the performance of hematite hetero-structures on an ITO substrate in a Tandem configuration (bias-free) with a suitable membrane can be focused on aiming at improving light absorption to enhance hematite’s PEC performance.

Research on hematite photo-anodes for other photoelectrocatalytic reactions like the production of H2O2, degradation of organic pollutants and organic synthesis can be explored.

Furthermore, the introduction of out-fields like magnetic and photothermal fields can be explored to enhance the PEC efficiency of hematite.