1 Introduction

In the last years, the growth of the world population and expansion of industrial centers have raised serious concerns caused by pollution and demands for clean water (H2O) and energy [1]. Versatile, low-cost, and environment-friendly treatment technologies have been investigated to mitigate these issues. Besides that, recent developments in research and application of heterogeneous photocatalysis for light (ultraviolet, visible, or solar irradiation) conversion have been used as the most advanced and greenest method [2]. Among various semiconductors applications, photoelectrochemical (PEC) cell for H2O splitting based on semiconductor oxides materials is a promising and environmentally friendly alternative approach to convert solar energy into a fuel source [3]. In addition, persistent organic pollutants or carbon dioxide (CO2) can be converted into less inert compounds by photo(electro)catalytic processes [4].

In the photoelectrocatalysis process, the semiconductor material must be supported on the conductive substrate as a film, usually called photoelectrode. The photocatalytic process starts when the material absorbs light with energy equal to or superior to the semiconductor bandgap energy (Ebg) value [5]. This process results in photogenerated electrons \(({e}_{\left(CB\right)}^{-})\) in the conduction band (CB) and holes \({(h}_{(VB)}^{+})\) in the valence band (VB). These photogenerated charges can reach the semiconductor surface and then react with chemical species present in an electrolyte solution. It is typically considered that an electron can be transferred to an acceptor molecule if its redox potential lies below the CB of the photocatalyst, while a hole can be transferred to a donor molecule if its redox potential lies above the VB of the photocatalyst [6, 7].

Several types of metal oxide semiconductors such as titanium dioxide (TiO2) [8, 9], tungsten trioxide (WO3) [10], hematite (α-Fe2O3) [11], and zinc oxide (ZnO) [12], have been investigated in many photocatalytic applications. However, although these materials exhibit substantial photocatalytic activity, TiO2 and ZnO have limited application due to their high Ebg values. On the other hand, WO3, α-Fe2O3 and Ag3PO4 oxides are colored, but have some drawbacks, which include reduced chemical stability at alkaline condition, acid medium and weak photostability, respectively [10,11,12,13,14]. In general, colored semiconductor oxides undergo photocorrosion, while white oxides like TiO2 and ZnO are photoactive only under ultraviolet radiation. In addition, the main disadvantages presented by oxides in catalytic applications include a short carrier path length and, high charge carrier recombination [15,16,17]. Thus, ideally, it would be important to obtain oxides materials that present excellent chemical stability and photoactivity under visible irradiation. This way, it is important to research semiconductors with small Ebg, facile surface reaction kinetic, abundant and stable chemically semiconductors for efficient PEC H2O splitting. Recently, our research group has investigated the properties of copper tungstate (CuWO4) films as photoelectrode for photocatalysis applications [18, 19]. CuWO4 is a transition metal tungstate, displays great chemical stability, and can harvest sunlight owing to its low bandgap energy.

Transition metal tungstates like AWO4 (with A2+ = Fe, Co, Ni, and Cu) are ternary oxide semiconductors that have attracted considerable attention due to their interesting photoelectrochemical, electrocatalytic, luminescent, and photocatalytic properties [20, 21] Wolframite-type tungstates are favored by metal cations with radii less than 0.077 nm, while metal ions greater than 0.099 nm favor scheelite-type tungstates [22]. In tungstates with the wolframite-type monoclinic structure the transition metal (A) and W atoms are both bonded with six oxygen atoms to form [AO6]/[WO6] clusters. In scheelite structures, the transition metal is combined with eight oxygen and the W atoms are bonded with four oxygen, which forms the [AO8]/[WO4] clusters [23,24,25].

The iron tungstate (FeWO4) is a semiconductor with a bandgap of approximately varying of 1.8 to 2.0 eV, absolving a wide range of light responses and playing an important role in the magnetic and photocatalytic potential application [26,27,28]. Cobalt tungstate (CoWO4) has a well-defined magnetic structure and bandgap of about 2.8 eV, exhibiting excellent electrical and magnetic properties [29]. Nickel tungstate (NiWO4) demonstrates an indirect bandgap of 2.0 eV [30] and the remarkable quantum size effect [31], which make it a potential photoanode candidate in the H2O splitting; and, others applications including photocatalysis [32], antimicrobial [33] and electrocatalysis [34]. Copper tungstate (CuWO4) has an indirect bandgap energy (Ebg) in the range of 2.2 to 2.45 eV [35,36,37], owing to their sufficient chemical stability, efficient sunlight utilization, and nontoxicity has seen applicated highly photoelectrocatalytic activity and photoelectrochemical H2O splitting [38,39,40]. In addition, others applications have aroused technological and scientific interest such as scintillation detectors [41], humidity and gas sensors [42, 43], supercapacitors [44], photoanodes [35], supercapacitor [45, 46], photoelectrolysis electrodes for lithium-ion batteries [47,48,49], electrochromic devices [44, 50] and photoelectrodes for hydrogen treatment [51].

Recently, some synthesis methods were employed for obtention of metal tungstate-based photoanode films as hydrothermal method [40], template-assisted synthesis method [52], co-precipitation [20], spray pyrolysis [53] and electrochemical deposition [54]. The polymeric precursor’s method (PPM) has been very effective about conventional methods principally for wastewater treatment [55]. The main advantages of PPM have been seen allowing the control of the composition, the grain size, and the homogeneity of powders [56]. There is a variety of studies using AWO4 as dispersed photocatalyst particles in the reaction medium for organic pollutants degradation. However, there are still few reports of preparing films from these materials. In addition, most film preparation methods are time-consuming and require an expensive structure. Thus, simple methodology, low cost, and high-performance AWO4 films still need to be investigated. To the best of our knowledge, the PPM has not been used to obtain transition metal tungstate AWO4 (A2+ = Fe, Co, Ni, and Cu) thin films. PPM has been studied in preparation of AWO4 oxides semicondutores at powder form. In addition, this is the first time that a detailed photoelectrochemical study of these materials has been presented.

Therefore, in this paper, AWO4 (A2+ = Fe, Co, Ni, and Cu) thin films were synthesized by the PPM and deposited on the transparent conductive substrate (FTO-glass) using a simple drop-casting method. After thermal treatment, the film’s photoelectrochemical behaviors were investigated. These studies are important and can provide information about the potential use of films in photocatalytic applications, such as H2O splitting. Besides, all films were structurally characterized using X-ray diffraction (XRD) patterns; the composition was analyzed by X-ray photoelectron spectroscopy (XPS); the morphological features were examined by field emission scanning electron microscope (FE-SEM), and the optical bandgap energies were estimated using UV–Vis spectroscopy.

2 Experimental section

2.1 Materials and procedure

The citric acid (C6H8O7; 99% purity), tungstic acid (H2WO4; 99% purity), ammonium iron (II) sulfate hexahydrate [(NH4)2Fe(SO4)2, 99,9% purity], cobalt (II) nitrate hexahydrate [Co(NO3)2·6H2O, 99.99% purity], nickel (II) nitrate hexahydrate [Ni(NO3)2·6H2O, 99.99% purity] and copper nitrate trihydrate [Cu(NO3)2.3H2O; 99% purity] were purchased from Sigma-Aldrich®. Ammonium hydroxide (NH4OH, 35% purity) and ethylene glycol (C2H6O2, 99.5% purity) were purchased from Dynamic®.

AWO4 (A2+ = Fe, Co, Ni, Cu) films were obtained from the resin by thePPM [57]. This is a modified sol–gel method that is based on the complexation of metal alkoxides, and the esterification/polymerization reaction [58]. Recently, this synthesis method has been successfully used to obtain crystalline ceramic powders when calcined at temperatures between 500 and 700 °C [55, 59, 60]. For first time here was used to obtain metal tungstate thin films.

2.2 Synthesis of AWO4 thin film by polymeric precursor method

A tungsten metal citrate (Cit-W, pH = 7) was initially prepared using deionized water (DI-H2O) as a solvent, 0.135 mol citric acid (C6H8O7) as a complexing agent and 0.045 mol tungstic acid (H2WO4) as a network modifier forming a transparent solution with concentration of 0.4256 μmol L−1. After homogenization, bivalent cation precursors (A = Fe2+, Co2+, Ni2+, and Cu2+) were added to citrate, respectively. The 150 g citrate with each bivalent cations was heat-treated and stirred at 90 °C for 2 h and obtained a homogeneous brown, wine, green, and blue color solution for iron, cobalt, nickel, and copper, respectively. Ammonium hydroxide (NH4OH) was used to adjust the pH ~ 7. This solution was polymerized with the addition of 2.467 g ethylene glycol (C2H6O2, polyalcohol); after reaction at a temperature of approximately 90 °C is formed a polyester and H2O. The citric acid/metal molar ratio was set at 3:1 and 60:40 for citric acid/ethylene glycol to promote citrate polymerization. The polymerized solution was heated until H2O evaporated and a polymeric resin formed about 24 h and cooled to 16 °C. Before the film’s preparation, the FTO glasses (1.0 × 2.5 cm) were cleaned by a sonication bath, successively, with H2O and a neutral liquid soap, DI–H2O, and finally with isopropyl alcohol for 15 min in each step. For films preparation, an adhesive tape was used to define the geometrical area of the films (1.0 cm2); then, 80 μL resin was dripped onto the FTO-glass (Sigma-Aldrich® R-7Ω cm−2) by dropping casting method, heated at 100 °C for 1 h (10 °C /min) and heat-treated at 500 °C for 2 h (2 °C /min) in a muffle furnace. To form a second and third layer, the resin was dropped again, heated and one unique thermal treatment was realized.

2.3 Structural, morphological, optical, and composition characterization of AWO4 films

AWO4 (A = Fe2+, Co2+, Ni2+, and Cu2+) thin films were characterized by the X-ray diffraction (XRD) pattern (Rigaku-DMax 2500PC, Japan with Cu Kα radiation (λ = 0.15406 nm) in the 2θ range from 10 to 80° with a scan speed of 0.02°/min. The diffraction patterns were compared to the data from the Inorganic Crystal Structure Database (ICSD). Optical analyzes were obtained by UV–Vis transmittance spectra using the UV–Vis spectrophotometer (Shimadzu, 2600). The Ebg values of the samples was obtained by the transmittance spectrum in the UV–Vis region, using FTO-glass as a reference, applying the Wood-Tauc plot method [61]. The surface and average grain size were characterized by field emission–scanning electron microscopy (FE-SEM, Supra 35-VP Carl Zeiss (Germany)). The images were obtained with a working voltage of 15 kV. The composition of the heterojunction film surface was analyzed by X-ray photoelectron spectroscopy (XPS; Scientia Omicron ESCA, Germany) using a monochromatic X-ray Al Kα (1486.7 eV).

2.4 Photoelectrochemical analysis of AWO4 photoelectrodes

The photoelectrochemical properties were investigated using an electrochemical cell fabricated with an optic glass window (100% transmittance for λ > 360 nm). Working electrodes were the AWO4 (A2+  = Fe, Co, Ni, and Cu) films, a Pt wire as the counter electrode, and Ag/AgCl (in 3.0 mol L−1 KCl saturated aqueous solution) was the reference electrode (in Luggin capillary). A three-electrode system configuration was used to measure; and so, was investigated in an aqueous solution of 0.1 mol L−1 Na2SO4 (pH 5.6) as the inert support electrolyte.

The measurements were performed using an Autolab Potentiostat/Galvanostat (PGSTAT302N Metrohm) and recorded by the NOVA 1.7 software, in the absence of light (dark) and under backside polychromatic irradiation. A metallic vapor discharge lamp (HQI-TS NDL) with a nominal potency of 150 W was used to irradiate the system. All our measurements were carried out with an irradiance of 100 mW cm−2, with the AWO4 films placed 10 cm from the irradiation source, similar to previous studies carried out by our group [9, 62]. The irradiance was measured with a Newport Power Meter, model 843-R. For more details on the behavior of the radiation source, see the supplementary material section (Fig. S1). Photoelectrochemical analyses consisted initially in the determination of layers which were using linear sweep voltammetry (LSV). Thus, the determination of the flat band potential (Efb) was carried out with the data recorded in the anodic potential range of 0.1–1.3 V at a scan rate of 1.0 mV s−1, using a chopped illumination at 0.1 Hz following the Gärtner–Butler model [63, 64]. The photocurrent-potential was measured using cyclic voltammetry (CV) with a scan rate of 20 mV/s in the dark and under polychromatic irradiation. Also, the chronoamperometric curves were registered under dark or light conditions with the electrodes polarized at + 0.7 V (vs Ag/AgCl). For comparison, the potential registered using the Ag/AgCl reference electrode was adjusted to a reversible hydrogen electrode (RHE), according to the Eq. (1) [37, 54, 65]:

$${\text{E }}\left( {{\text{vs}}.{\text{ RHE}}} \right)\, = \,{\text{E }}\left( {{\text{vs}}.{\text{ Ag}}/{\text{AgCl}}} \right)\, + \,0.0{591} {\text{Vx pH}}\, + \,0.{199} {\text{V}}{.}$$
(1)

Further, the potentials after adjustment with respect to the RHE (in volts) were converted to energy levels (eV) relative to vacuum level using Eq. (2) [66]:

$${\text{E}}_{{({\text{eV}})}} \, = \,\left[ { - {4}.{5} {\text{eV}} - {\text{eE}}_{{({\text{RHE}})}} } \right]$$
(2)

From the photoresponse, it was possible to calculate the recombination lifetime from the time constant (τ) obtained from Eq. (3), which describes the kinetics of the transient currents as follows below:

$$R=\mathrm{exp}(-t/\tau )$$
(3)

where the ratio R is \(R=\frac{Jt-Jst }{Jin-Jst}\), Jt is the photocurrent at time t, Jst is the steady-state photocurrent, and Jin is the photocurrent at the initial time [67, 68].

3 Results and discussion

3.1 Structural, morphological and optical characterization of AWO4 films

The crystal structures of the resulting samples were identified by the X-ray diffraction (XRD) patterns. The XRD patterns are recorded in the range of 10–80º as shown in Fig. 1a–d. Diffraction signals suppressed and marked with the symbol “•” correspond to the FTO transparent conductive glass substrate.

Fig. 1
figure 1

XRD patterns of (a) FeWO4, (b) CoWO4, (c) NiWO4, and (d) CuWO4 thin films synthesized by the PPM and heated at 500 °C for 2 h. The “•” indicates the FTO-glass substrate. The vertical bars represent planes and intensity of the ICSD card No. 15952, 15851, 15852, and 16009 for the pure monoclinic FeWO4, CoWO4, NiWO4 phase, and pure triclinic CuWO4 phase, respectively

Full size image

For FeWO4, CoWO4, and NiWO4 films displayed in Fig. 1a–c, the XRD patterns show diffraction peaks which are characteristic and attributed to (010), (100), (110), (011), (111), (\(\overline{11 }\) 1), (020), (002), (021), and (200) planes, at 2θ equal to 15.6º, 19º, 24º, 25º, 30.6°, 31°, 36°, 37°, and 38°, respectively. These XRD patterns are characteristic of wolframite-type monoclinic phase structure with space group (P2/c) as determined by ICSD card No. 15952, 15851, and 15852 for FeWO4, CoWO4, and NiWO4 crystal, respectively [69, 70]. On the other hand, the XRD pattern for the CuWO4 sample (Fig. 1)d shows typical diffraction peaks for triclinic structure (ICSD No. 16009) with space group (P\(\overline{1 }\)) [71]. The XRD diffraction were registered at 2θ of 15.29º, 18.19º, 19.02º, 22.94º, 23.56º, 24.13º, 25.96º, 26.97º, 28.70º, 30.14º, 30.85º, 31.65º, 32.11º, 34.34º, 35.66º, 36.44º and 36.86º, which correspond to the planes (010), (001), (100), (110), (0 \(\overline{1 }\) 1), (011), (\(\overline{1 }\) 10), (101),\((\overline{11 }\) 1), (111), (020), (\(\overline{1 }\) 11), (1 \(\overline{1 }\) 1), (120), (0\(\overline{2 }1\)), (021), and (002), respectively. The diffraction peaks without other impurities or deleterious phases, which suggests that films can be prepared successfully by means of PPM.

FE-SEM images for the morphology, particles size, and surface of AWO4 (A2+ = Fe, Co, Ni, and Cu) films are shown in Fig. 2a–d, respectively. In Fig. 2a–d, the FE-SEM images revealed the morphology and particles size of the FeWO4, CoWO4, NiWO4 and CuWO4 films. The deposition of the FeWO4, CoWO4, NiWO4, and CuWO4 films on a conductive substrate (FTO-glass), followed by annealing at 500 °C, led to an immobilized particle density of ca. 1.6 ± 0.2 mg cm−2, 1.2 ± 0.2 mg cm−2, 0.9 ± 0.2 mg cm−2 and 1.4 ± 0.1 mg cm−2, respectively. The FeWO4 film presented in Fig. 2a indicates that has a well-sintered nanoplates-like structure with an average particles size of 51.3 nm ± 1.343 (See histogram in Fig. S2(a)). Yu et al. reported that FeWO4 synthesized by hydrothermal method present morphology influenced by pH condition, which can change from hexagonal flakes to nano-rods [72]. Figure 2b shows the morphological surface for CoWO4 film with irregular spherical nanoparticles and an average particles size of ca. 27.7 nm ± 0.9799 (See histogram in Fig. S2(b)). Whereas, the NiWO4 film exhibits an irregular surface morphology formed by polyhedron interconnected particles (Fig. 2c) and an average particles size of ca. 74.75 nm ± 0.8185 (See histogram in Fig. S2(c)). Figure 2d shows that CuWO4 film is formed by irregular nanoparticles with an average size of 24.77 nm ± 2.537 (See histogram in Fig. S2(d)). Thus, the presence of several grains in the CuWO4 film agglomerated is very similar to that spectated to triclinic structure crystals [37]. In this case, a porous structure with a non-uniform pore distribution is formed. This porous nanostructure can contribute to superior charge mobility and a large specific surface area [73]. In all cases, the average crystal size was calculated by equatorial diameter and estimated using the GNU Image Manipulation Program (GIMP 2.10.10 version for Windows 7–64 bits) [74].

Fig. 2
figure 2

FE-SEM images of the surface of the (a) FeWO4, (b) CoWO4, (c) NiWO4, (d) and CuWO4 films synthesized by PPM and heated at 500 °C for 2 h

Full size image

The optical properties of these films were examined with UV–Vis spectra registered in transmittance mode and estimated the bandgap energy (Ebg) by Wood-Tauc plot method as shown in Fig. 3. Figure 3a shows that the AWO4 (A2+ = Fe, Co, Ni, and Cu) films had a maximum transmittance of ~ 72% in the visible region and interference fringes can be registered in FeWO4, and CoWO4 thin films. Considering an indirect gap transition, the optical Ebg for AWO4 (A2+ = Fe, Co, Ni, and Cu) films were calculated with the following Eq. (4) [61]:

$$ (\alpha {\text{h}}\upsilon )^{{0.{5}}} \, = \,{\text{C}}({\text{h}}\upsilon - \it {\text{ E}}_{{{\text{bg}}}} ).$$
(4)
Fig. 3
figure 3

a UV-Vis transmittance curves of films and bandgap energy estimated by Wood-Tauc plot method for (b) FeWO4, (c) CoWO4, (d) NiWO4 and (e) CuWO4 film deposited on FTO substrate.

Full size image

In this Eq. (4), the hυ is the incident photon energy, C is a constant of proportionality of semiconductors, Ebg is the optical bandgap energy and α is the optical absorption coefficient of the film. According to the equation, (αhν)0.5 has a linear relation with the hν [75]. The optical Ebg values of AWO4 (A = Fe, Co, Ni, Cu) films were determined by extrapolation of the linear relation to (αhv)0.5 = 0. As illustrated in Fig. 3 b–e, the Ebg of FeWO4, CoWO4, NiWO4 and CuWO4 films are estimated to be about 1.87, 2.40, 2.45, and 2.19 eV, respectively. These Ebg values are slightly close that on registered in other studies [30, 54, 76,77,78,79,80]. Slight deviations in the Ebg values can be attributed to the effects of film morphology and particle sizes. These results indicate that the materials are able to harvest sunlight. Also, UV–Vis measurements using an integrating sphere were made. In general, there are no significant differences for the curves registered in both modes (See Fig. S3 on the Supporting Information).

An X-ray photoelectron spectroscopy (XPS) analyses were utilized to investigate the elemental and chemical composition of samples, as shown in Figs. S4, and 4. A full survey spectrum shown in Fig. S4 indicate the presence of Fe 2p [81], Co 2p [82], Ni 2p [83], and Cu 2p states [51] signals due to presence these elements in FeWO4, CoWO4, NiWO4, and CuWO4 films samples. All XPS spectra are attributed to the constituent element core levels or Auger lines. Also, in all survey analyses, it was possible to observe the presence of corresponding peaks O 1 s and W 4f. The C 1 s peak can be attributed to carbon adsorbed on the surface of the AWO4 films, probably due to polymeric precursors method not completely eliminated during the thermal treatment. Table S1 displays the values of XPS primary peak for all A divalent ions transitions metal (A2+  = Fe, Co, Ni, and Cu), W 4f, and O 1 s core-level binding energies.

Metal ions (A2+  = Fe, Co, Ni, and Cu) 2p state in XPS spectra of each AWO4 film are presented in Fig. 4a–d. The deconvoluted Fe 2p XPS (Fig. 4)a consists of two peaks which corresponds the binding energy of Fe 2p3/2 (710.70 and 712.48 eV), which higher binding energy of 712.48 eV can be attributed to + 3 oxidation state [84], and Fe 2p1/2 (724.88 eV) due to Fe2+ in the lattice sites of FeWO4 film [85]. In addition, there are two shake-up satellite peaks at Fe 2p3/2 (715.71 and 719.81 eV) and one satellite peak at Fe 2p1/2 (732.07 eV), which are typical Fe2+ oxidation states [86]. By using a Gaussian fitting method, the high-resolution Co 2p spectra is deconvoluted into two spin–orbit doublets and two shake-up satellites peaks (802.08 and 785.28 eV) [87]. The XPS Co 2p core-level spectra of CoWO4 film is presented in Fig. 4b. The deconvolution of Co 2p peak is divided in Co 2p3/2 (779.94 and 781.31 eV) and Co 2p1/2 (796.57 eV) understanding Co2+ oxidation state. In same way that Fe for peaks of Co 2p (higher binding energy of 781.31 eV) attributed to +3 oxidation state [88].

Fig. 4
figure 4

High—resolution XPS spectra for (a) Fe 2p, (b) Co 2p, (c) Ni 2p, (d) Cu 2p, (e) W 4f, and (f) O 1 s core-level spectra of FeWO4, CoWO4, NiWO4, and CuWO4 film, respectively

Full size image

Similarly, an XPS Ni 2p spectrum of NiWO4 film is seen in Fig. 4c. The binding energy of the Ni 2p3/2 (855.08 and 856.59 eV) and Ni 2p1/2 (872.66 and 874.28 eV) are resulted of spin–orbit doublets deconvolution which characteristic of Ni2+ and two shake-up satellite peaks (856.59 and 874.28 eV) and good agreement with the literature data [34]. Two deconvolution of Ni 2p (higher binding energy of 874.28 eV and 856.59 eV) can be attributed to + 3 oxidation state [89]. The satellites peaks are typical shake-up satellites which were due to the multielectron transitions and the decrease of the kinetic energy of emitted photoelectrons [90, 91]. Also, Fig. 4d shows XPS core-level spectrum of Cu 2p, which displays two sets of Gaussian doublets (Cu 2p3/2 and Cu 2p1/2) located at 931.33/933.61 eV, and 951.42/953.54 eV and shake-up satellite peaks located at 940.52, 942.80 and 961.40 eV assigned to a typical feature for Cu2+ oxidation state [92]. On the other hand, Tang and co-works have shown in previous works that signals which occur at lower binding energy values (951.42 and 931.33 eV) can be attributed to Cu1+ oxidation state [51]. The appearance of doublet peaks in bivalent cations can indicate presence of monovalent or trivalent cations (A+ or A3+) [91, 93]. Figure 4e displays the W 4f core-level spectra of AWO4 (A2+ = Fe, Co, Ni, and Cu) films. The deconvolution of W 4f spectra into one pair of spin–orbit doublet peaks dominated by W 4f5/2 and W 4f7/2 states corresponding to the W6+ valence state in tungstate-based crystal structure [7, 94]. The binding energy values of W 4f show doublet of W 4f7/2 and W 4f5/2, besides W 5p3/2 at 33.64/34.28, 35.95/37.22 and 39.83 eV for FeWO4, 34.26/34.96, 36.45/36.65 and 40.36 eV for CoWO4, 34.14/34.72, 36.57/36.71 and 40.37 eV for NiWO4 and 34.42/35.07, 36.73/37.91 and 40.19 eV for CuWO4 [95]. The appearance of doublet peaks of W 4f and W5p in XPS spectra indicates the presence of W5+ in the film [96]. As the presence of W5+ (or A+) in AWO4 films can induce oxygen vacancies that are regarded as shallow donors [52], a higher concentration of W5+ will result in a larger electron density in metal tungstate-based films [51]. Figure 4f shows the XPS O 1s core-level spectra of AWO4 films annealed at 500 °C for 2 h. The O1s spectra was decomposed into three peaks located at 529.89, 530.88 and 531.60 eV, and were assigned to the O–W–O bonds, lattice oxygen of Fe–O–Fe, metallic oxide of Fe–O bonds in FeWO4, respectively [2, 26]. Similarly, the XPS O 1s spectra of CoWO4 was analyzed and then, was observed three peaks located at 529.32, 530.23 and 531.40 eV. For NiWO4, the O 1s core-level spectra deconvoluted into three peaks are located at 529.46, 530.27, and 531.57 eV. Similar to others films, the O 1s core-level spectra of CuWO4 was also deconvoluted in three peaks located at 529.76, 530.56 and 531.65 eV corresponding to same description as previously presented. The XPS spectra confirm the formation of possible defect states on the surface and suggest the presence of oxygen vacancies created during production and incorporated in the crystal analyzed [15]. Here, the XPS data suggest that oxygen vacancies cause lower oxidation states in metals, in order to maintain the electroneutrality of the material. Thus, these defects may be responsible for the n-type conductivity of semiconductors, as well as the non-stoichiometry at surface FeWO4, CoWO4, NiWO4, and CuWO4 films [97]. Hence, the XPS spectra confirm the formation of good-quality tungstate-based films.

3.2 Photoelectrochemical behavior of AWO4films

Figures 5a–d show the photoelectrochemical behavior of AWO4 (A2+ = Fe, Co, Ni, and Cu) films with different deposited layers on FTO-glass was examined by linear sweep voltammetry (LSV) curves registered at 1mVs−1 under chopped 0.10 Hz polychromatic irradiation. All AWO4 films displayed a more negative photo-potential after irradiation condition, which means, in the dark the open-circuit-potential (OCP) value is more positive than OCP registered under irradiation condition (Elight-on< Elight-off vs. Ag/AgCl). This negative photo-potential is characteristic of n-type semiconductor oxides [98]. Thus, LSV curves were registered toward anodic direction and positive photocurrents were observed as applied bias until oxygen evolution reaction (OER) signals under dark condition. To investigate the influence of AWO4 film layers on photocurrent response, samples were prepared with one, two, and three layers.

Fig. 5
figure 5

LSV curves plots of (a) FeWO4, (b) CoWO4, (c) NiWO4, and (d) CuWO4 photoanodes in inert 0.1 mol L−1 Na2SO4 solution (pH ~ 5.6) irradiated with chopped polychromatic light (0.1 Hz), respectively

Full size image

For FeWO4 and CoWO4 materials, films with two layers showed the best photoresponse values, which can be with associated adequate quantity and thickness of semiconductors, as shown in Fig. 5a,b. Moreover, the FeWO4 and CoWO4 films with three layers showed lower values of photocurrents, suggesting that these films present greater resistance in the flow of electrons. From Fig. 5c,d, it is possible to observe that this reduction in photocurrents values was not registered for NiWO4, and CuWO4 films. In NiWO4 film with three layers, higher photocurrents were observed, while in CuWO4 films this photoresponse value was maintained. For comparison effect, all other photoelectrochemical measures registered hereinafter were developed with two layers for FeWO4, CoWO4, NiWO4, and CuWO4 films.

From Fig. 5a, it is possible to observe that an enhancement in photocurrent density of FeWO4 photoelectrode started from 0.95 V vs. Ag/AgCl (1.48 V vs. RHE). For CoWO4, NiWO4, and CuWO4 the photocurrent values evolve in potentials closer to OCP, suggesting that in these films the electrons can reach the conductive substrate more easily. In all photoelectrodes, some positive spikes of current were observed immediately after the light was turned on, followed by an exponential drop in current values, which can be related to transient effects in the semiconductor material [65]. These spikes can be caused by the rapid injection of electrons into the CB (peak anode current), while the exponential drop is related to the electron/hole charges recombination process, even under polarization conditions. When the light is turned off, rapid recombination in VB happened and cathodic spikes can be registered, as displayed in Fig. 5b [99]. Current spike behavior also depends on the degree of crystallinity. Current spike diminishes by increasing the annealing temperature, but the conductivity of FTO substrate deteriorates at temperature above 500 °C. The NiWO4 shows a beautiful LSV curve (Fig. 5c) due to a good crystallinity as shown in FE-SEM (Fig. 2c) and large particle size (Fig. S2c). Also, this transient cathodic peak can be associated to back electron transfer from the external circuit to the photoanode, which resulted in a surface recombination process [100]. According to Le Chatelier’s principle, a system at equilibrium that is subjected to an alteration in concentration will readjust itself to counteract the alteration and establish a new equilibrium caused by an excess of electrons. This would explain the transient current spikes in the AWO4 photoanodes.

The photocurrent density measurements by LSV curves were utilized to estimate the flat band potential (Efb) of these semiconductors, following the Butler-Gärtner model [63, 64], as displayed in Fig. 6(a–d). Gärtner–Butler analysis is an alternative method of determining the Efb is based on the measurement of the square of the photocurrent density as a function of electrode potential. The Efb is predicted to be at the intercept of the square of the photocurrent density with the potential axis [101]. This model assumes that the photocurrent (Iph) is observed only when the electrode potential is more positive than the semiconductor flat band potential (i.e., when E > Efb and Iph > 0). Thus, the Butler-Gärtner model can be represented by Eq. (5) below:

$${I}_{ph} = \alpha {W}_{0}q{\varphi }_{0}\sqrt{E-{E}_{fb}}$$
(5)

where α is the absorption coefficient, W0 is the depletion layer width, φ0 is the radiation intensity, q is the electron charge, and Efb is the flat band energy. This methodology already was employed to determine Efb values for other photoelectrodes by our research group [11, 37, 102]. Therefore, Efb values obtained by extrapolation of photocurrent initial potential were estimated at 0.23, 0.25, 0.28, and 0.38 V (vs. Ag/AgCl) for FeWO4, CoWO4, NiWO4, and CuWO4 electrodes, respectively. The Efb value is strongly dependent on the medium conditions such as pH and electrolyte composition, as well as film thickness, crystal structure, and morphology of the semiconductor [103]. It is very important to determine the Efb value of semiconductor materials before its application in any photoelectrochemical proposition [97, 104,105,106], since this electrode potential can be related to the Fermi energy level and the CB band edge position in n-type semiconductors [107]. The flat band potential and CB must be related to electron donor and acceptor states using Eq. (6):

$${E}_{CB}={E}_{fb}-\frac{kT}{e}ln\frac{{N}_{C}}{{N}_{D}}$$
(6)

where k is the Boltzmann constant, T is the absolute temperature, e is the electron charge, NC is the effective density of states in the CB, and ND is the effective charge density [108]. Considering that the values of ND and NC are similar, the second term in Eq. (6) should be relatively small and then conduction band potential (ECB) can be approximated to Efb [109]. Thus, the relative positions of VB and CB edge potentials of the semiconductors were estimated from Ebg and Efb values and using the Eqs. (5, 6), as shown in Fig. 6e. All electrodes show sufficient CB potential values to oxidize H2O/O2 (see dashed line in Fig. 6e. The photocatalytic activity of each material is dependent on the relative positions of the conduction and valence bands. Previous studies reveal that it is important to determine the oxidation potential of species present in solution, relating this potential to the positions of CB and VB to predict the photocatalytic capacity of the semiconductor material [11, 102]. Similarly, Efb values were estimated by Mott-Schottky plot through Electrochemical Impedance Spectroscopy. Fig S5(a) shows the Mott-Schottky curves and relative positions of the CB and VB obtained for all samples. As observed at Fig S5(b), there are no significant differences for the relative positions of the bands for curves obtained by Mott-Schottky or by Butler-Gärtner model.

Fig. 6
figure 6

Variation in the square of the photocurrent density with applied potential, for (a) FeWO4, (b) CoWO4, (c) NiWO4, and (d) CuWO4 electrode in 0.1 mol L−1 Na2SO4 aqueous solution, under polychromatic irradiation obeying Butler-Gärtner model and (e) the experimentally determined band diagrams

Full size image

Figures S6(a–d) display the cyclic voltammetry (CV) curves of AWO4 (A2+ = Fe, Co, Ni, and Cu) films registered at 20 mV s−1 in the dark and under polychromatic irradiation. In the dark, the electrodes presented an almost zero capacitive current. As mentioned before, the potential value associated with OER serves to limit the potential window toward anodic reactions. The OCP shifted toward more negative values under irradiation conditions, resulting in a photopotential of − 20, − 60, − 80, and − 50 mV for FeWO4, CoWO4, NiWO4, and CuWO4 photoelectrodes, respectively. These differences recorded in the OPC before and after illumination may be attributed to electron accumulation in the conduction band (CB), due to the electron–hole charge separation [110]. According to theory, after adequate illumination of a n-type semiconductor oxide, the electrons accumulated in the CB are collected by circuit until the Pt counter-electrode, while the photogenerated holes are moved to the electrode/electrolyte interface, thus generating the anodic photocurrent [111]. The photocurrent density of electrodes was respectively 15 μA cm−2, 30 μA cm−2, 20 μA cm−2, and 48 μA cm−2 at 1.0 V vs. Ag/AgCl (1.53 V vs. RHE) for FeWO4, CoWO4, NiWO4, and CuWO4 electrodes, as displayed in Figs. S6(a–d). The photoelectrochemical performance of photocatalyst electrodes fabricated in different synthesis methods reported in recent years were compared with this work, as shown in Table S2. Table S2 relating synthesis methods, photocurrent values, type of irradiation and electrolyte was presented [18, 19, 112,113,114,115,116,117,118,119,120,121,122,123,124,125]. Photocurrent values in photoelectrochemical measurements are dependent on several factors, including the material’s bandgap energy value, material defects, and the ability of electron charges to reach the conductive substrate. Even with a higher Ebg value, CuWO4 film presented higher photocurrent value, indicating that the photogenerated charges reach the conductive substrate more easily than the other AWO4 investigated.

To investigate the photocurrent stability and \({e}_{CB}^{-}/{h}_{VB}^{+}\) recombination lifetime of electrodes in an aqueous medium, a photoelectrochemical study was carried out by measuring the photocurrent density over time at a constant potential of ca. 0.7 V vs. Ag/AgCl (1.23 V vs. RHE). This study is based on transient photocurrent behavior, in which the current response of a system is monitored while a light source is modulated periodically (usually light-on and light-off). Based on this measure, it is possible to study charge carrier dynamics and their traps in a semiconductor material as photoelectrode. The curves registered for AWO4 films are displayed in Fig. 7. In the dark, all films displayed the current was almost null, while under irradiation, the photocurrent increased abruptly because of charge separation and electrons accumulating in the conduction band, a typical characteristic of n-type semiconductor oxides, as seen in Fig. 7a. In addition, the films demonstrated excellent photo-stability for 2100s, except the FeWO4 film, whose, its photocurrent density has decreased continuously over time. The photocurrent density values followed the same results exposed previously by LSV and cyclic voltammetry measurements, see Fig. 5 and S6.

Fig. 7
figure 7

a Photocurrent density-time (J-t) curves of the annealed films measured under polychromatic illumination with chopped 300 s light on/off at 0.7 V vs. Ag/AgCl, b schematic representation of photocurrent transient curve and c normalized plot of current–time dependence for AWO4 (A2+  = Fe,Co, Ni, and Cu)

Full size image

To obtain the electron–hole recombination lifetimes for the photoelectrodes, photocurrent-time curves, following the schematic representation in Fig. 7b, were analyzed according to Eq. (3). This equation shows that the kinetics of the transient photocurrent can be obtained from slope of the ln R vs. time (t) plot, as shown in Fig. 7c for FeWO4, CoWO4, NiWO4, and CuWO4 photoelectrodes, biased at 0.70 V. The reciprocal 1/τ is related to the electron–hole lifetime. In previous studies, we demonstrated that the mass transport mechanism (i.e. migration or diffusion) in the system does not alter the recombination exponential profile of transient photocurrent utilized to determine the lifetime [102]. As observed in Fig. 7a and schematically represented in Fig. 7b, an initial anodic photocurrent spike Jin indicates the injection of electrons in the CB owing to the electron/hole charge separation process. At the same time, electrons are transported to the conductor substrate to be collected by the external circuit; the holes are moved towards the semiconductor surface where they are reduced by the species in the electrolyte. Simultaneously, the photocurrent decreases exponentially with time until it reaches a steady-state photocurrent density (Jst) because of charge recombination.

The slope of the plot provides photocurrent transient (τ), which is related to the charge recombination lifetime in each electrode. The recombination lifetime results are displayed in Table S3. The τ values were calculated at about 83 ± 7, 32 ± 3, 25 ± 4, and 40 ± 1 s for the FeWO4, CoWO4, NiWO4 and CuWO4 films, respectively. These τ values for the CoWO4 and NiWO4 films are smaller than the FeWO4 and CuWO4 films. However, these charge recombination time values are not in agreement with the photocurrent values shown in Fig. 5. Thus, Nyquist plots were used to assess the effective parallel resistance of charge transfer in the samples. Fig S7 shows the Nyquist plots. Even not displaying a well-defined arc as a semicircle, the maximum values observed for the Z' axis were used to verify the greater or lesser tendency in charge transfers. The reduction of the effective parallel resistance indicates a better charge transfer for CuWO4, CoWO4, NiWO4, and FeWO4 samples, respectively. This sequence agrees with the photocurrent values. Thus, probably, the electrons promoted to the conduction band do not return to VB. However, we have suppressed defects in the crystal lattice and we believed the holes are captured by traps present in the grain boundaries of the material. Therefore, the photoelectrochemical properties were observed in different wolframite films with anodic photocurrent and capable of harvesting visible light allowing several photoelectrochemical applications.

4 Conclusion

In summary, FeWO4, CoWO4, NiWO4, and CuWO4 thin films onto FTO-glass were synthesized with success by the PPM method via drop-casting after heat treatment at 500 °C for 2 h. XRD patterns analyses confirmed the presence of wolframite-type monoclinic structure for FeWO4, CoWO4, and NiWO4 films, while the CuWO4 film exhibits a triclinic structure. The AWO4 (A2+ = Fe, Co, Ni, and Cu) films showed an ability to absorb in the visible light region, which suggests being excellent harvesters of sunlight for energy conversion applications, such as H2O oxidation. Morphological features showed by means of FE-SEM images revealed the presence of nanostructured oxides with irregular shapes and the diverse average size of particles. The PEC testing under illumination revealed anodic photocurrent and negative photopotential characteristics of n-type semiconductors oxides. In inert electrolyte, the CuWO4 electrode showed the highest photocurrent density value of 48 μA cm−2 at 1.0 V vs. Ag/AgCl (1.53 V vs. RHE). In addition, all electrodes demonstrated excellent chemical stability for photoelectrochemical oxygen evolution reaction with constant bias potential at 0.7 V vs. Ag/AgCl (1.23 V vs. RHE). From Ebg and Efb values were possible to assemble a diagram of (CB and VB) positions, which leads to deducting the greatest applicability of electrodes in PEC, exploring these AWO4 (A2+ = Fe, Co, Ni, and Cu) films as electrodes with a high potential to H2O-splitting photocatalysts.