Electrochromic solar water splitting using a cathodic WO 3 electrocatalyst

Solar-driven water splitting is an emerging technology with high potential to generate fuel cleanly and sus-tainably. In this work, we show that WO 3 can be used as a cathodic electrocatalyst in combination with (Ag,Cu) InGaSe 2 solar cell modules to produce hydrogen and provide electrochromic functionality to water splitting devices. This electrochromic effect can be used to monitor the charge state or performance of the catalyst for process control or for controlling the temperature and absorbed heat due to tunable optical modulation of the electrocatalyst. WO 3 films coated on Ni foam, using a wide range of different sputtering conditions, were investigated as cathodic electrocatalysts for the water splitting reaction. The solar-to-hydrogen ( STH ) efficiency of solar-driven water electrolysis was extracted using (Ag,Cu)InGaSe 2 solar cell modules with a cell band gap varied in between 1.15 and 1.25 eV with WO 3 on Ni foam-based electrolyzers and yielded up to 13% STH efficiency. Electrochromic properties during water electrolysis were characterized for the WO 3 films on transparent substrate (indium tin oxide). Transmittance varied between 10% and 78% and the coloration efficiency at a wavelength of 528 nm and the overpotential of 400 mV was 40 cm 2 C (cid:0) 1 . Hydrogen ion consumption in ion intercalation for electrochromic and hydrogen gas production for water electrolysis processes was discussed.


Introduction
Concerns about climate change, urban air pollution, and energy security have led to an increased interest of scientists, industry, and political leaders to develop methods for production, usage, and storage of energy in a clean and sustainable way.Molecular hydrogen is here a promising and carbon-free energy carrier for energy storage for later utilization as a source for heat, electricity, or as a chemical feedstock for industry.Electrolysis of water is a technique that yields hydrogen with high purity and would form a sustainable solution if electricity from wind-, hydro-, or solar-harvesting were used.A direct production in a photocatalytic or photovoltaic (PV) system coupled to an electrolyzer can be readily used today where molecular hydrogen stores the solar energy in a carbon-free fuel for later use in vehicles or stationary applications.The stored energy can be used as heat by combustion or as electrical power via a hydrogen fuel cell on demand or when the solar power is intermittent.The history of electrolysis began at the end of the 18th century (1789) when Deiman and van Troostwijk dissociated water using an electrostatic generator and two gold electrodes immersed in water [1].Subsequently, Nicholson and Carlisle were the first who described the process of using electricity to dissociate water into hydrogen and oxygen in their experiments on the electrolytic splitting of water using voltaic piles in the year 1800 [2].Following the electrification of society, over 400 industrial water electrolyzers were in use in 1902 [3].Light-driven electrolysis was first described by Fujishima and Honda in 1972 using TiO 2 as a photoanode and platinum as a cathode [4] and has been an active field of research since then.
In electrocatalytic water splitting, hydrogen and oxygen gases evolve from reduction and oxidation of water at the cathodic and anodic electrodes, respectively.For photocatalytic water splitting using photoelectrochemical cells (PECs), the charge carriers are created from the photovoltaic effect close to the catalytic site.Solar cell driven electrocatalysis instead has advantages over direct photocatalysts in terms of higher efficiency, high power per area, and the possibility to separate the light absorption process from the electrolysis process [5] allowing optimization of each function independently, and not the least, enable the removal of the photo absorbing material from the electrolyte [6].Proton exchange membrane (PEM) electrolysis, which occurs in acidic electrolytes (pH 0-7), has better efficiency and enhanced ramping capability over other types of electrolysis [7].The principle of PEM electrolysis together with half-cell and overall-cell reactions can be written as and are shown in Fig. 1a.The typical operating cell voltage for commercial electrolyzers (1.8-2.0V) is much larger than the theoretical minimum value (1.23 V), and they are typically operated at higher current densities to produce hydrogen at a high rate in order to offset the capital cost of the materials used.For water electrolysis, the fabrication of cost-effective non-noble metals has attracted significant interest as a replacement for expensive Pt-based catalysts.However, they would also need to operate at a relatively high hydrogen production rate to lower the total cost of hydrogen produced over the lifetime of the system.Transition metal oxides such as WO 3 can be promising alternatives and have previously mainly been used as photoanodic catalystst for oxygen evolution with modest solarto-hydrogen (STH) efficiency but has recently also been highlighted as a cathodic electrocatalys that show promising stability with HER stability maintained after 2000 cycles [8].
PV-technologies such as crystalline-and thin film-silicon, perovskites, and (Ag, Cu)InGaSe 2 (A-CIGS) can be used to supply the energy to drive an electrolyzer [6,[9][10][11][12] with theoretical STH efficiencies spanning from about 25% from serial interconnected solar cells to 32% for tandem approaches when using un-concentrated solar illumination [13].High optical absorption coefficient and adjustable band gap of A-CIGS solar cells make them attractive because of the possibility to achieve high efficiency (η) in the cells.The band gap of the A-CIGS cells can be changed by changing the Ga/(Ga+In) ratio [14][15][16].This property is helpful in finding an optimum condition for a highly efficient solar water splitting system [17].
Electrochromism on the other hand, is an energy-saving technology that allows control of the amount of solar heat and visible light transmitted by a thin film.Electrochromic thin films are integral parts of multi-layer devices (Fig. 1b) that are capable of varying their optical properties by application of an externally applied voltage [18,19].Electrochromic devices can be used as smart windows in buildings, and this technology can significantly reduce energy use for cooling, air conditioning, and illumination [20].
The history of electrochromism starts in the 1950s with the development of qualitative theory and quantitative prediction of the shift of the absorption and emission spectra of certain organic dyes by solvent and polarization effects [21,22].Based on these investigations, a possible change of color in dyes by applying an electric field was suggested by Platt in 1961 [23].It was the first time that the term electrochromism was used.In 1969, electrochromism was experimentally demonstrated in an inorganic compound, WO 3 , by Deb [24] and in 1985, Svensson and Granqvist introduced the potential application of electrochromic materials in smart windows [25].A typical smart window, as shown in Fig. 1b, contains five layers between two transparent substrates where a transparent ion conductive layer lies at the center of the structure.One side of the layer is in contact with a thin film of a cathodic electrochromic material.This layer is colored by the insertion of ions and charge-balancing electrons into the film.On the opposite side of the ion conductor, an ion storage layer is situated.Usually, anodic electrochromic thin films are used in this layer in order to get the highest possible modulation of the transmittance.The anodic layer is colored by the extraction of ions and electrons from the film.In current electrochromic technology, the most common devices include thin films based on cathodically coloring WO 3 and anodically coloring NiO.The coloration of WO 3 is attributed to intervalence electron transfer between W 6+ and W 5+ valence states [26].The most accepted model for the coloration is that polarons formed by the localization of the injected electrons at W 5+ sites are hopping between adjacent W i and W j sites when the light is absorbed [27].
The electrochromic phenomenon in WO 3 occurs with the intercalation of M + cations such as H + , Li + , and Na + from the electrolyte and the reaction mechanism is given as: The expression for the time-dependent intercalation level is [28].
where z is the valence of the intercalated ion, e is the elementary charge, d film thickness, A is the active area of the electrode, N W is the tungsten number density, and I is the current during the measurement.Electrochromic properties are assessed by optical modulation which is the difference between the bleached (T b ) and colored transmission (T C ) states; optical density is defined as OD(λ, x) = dα(λ, x) where α is the absorption coefficient at the specified optical wavelength λ and intercalation level x; and differential coloration efficiency K(λ, x) given by [28].
where q is the inserted charge per unit area.Considering the relation of OD with total transmittance (T t ) and reflectance (R t ), the differential coloration efficiency K(λ, x) can be written as Transition metal oxides are of interest for many applications other than electrochromic devices, for example, as catalysts, batteries, supercapacitors, and gas sensors.Therefore, there is a growing interest in multi-functional electrochromic devices, which can provide energy storage and energy production functionality together with control over light and/or heat [29,30].Photoelectrochromic devices combine electrochromism with photovoltaic electricity generation in an integrated device [31,32].The energy need of the electrochromic devices can be supplied from the photovoltaic part and directly used for the optical modulation without any demand for external power.Very recently, it was found that electrochromic devices can also exhibit tunable microwave dielectric properties [33].Electrochromic batteries and electrochromic supercapacitors integrate energy storage functions with optical modulation [34,35].In particular, the ability to monitor the level of energy stored in the devices by the color variation makes electrochromic energy storage devices attractive.
Electrolysis in electrochromic devices was investigated to examine the role of water on device performance [36,37].The optical modulation has been shown to increase and the coloration and bleaching times decreased with the presence of water in the ion-conductive layer [38].
In this study, we combine investigations of electrocatalytic and electrochromic properties of WO 3 thin films.The study consists of three main parts.First, cathodic electrocatalytic properties of WO 3 thin films deposited using a wide range of sputter conditions on Ni foam were investigated in the potential range of the hydrogen evolution reaction (HER) in an acidic electrolyte and compared with a well-known NiMo electrocatalyst for the overall water splitting reaction.Second, optimum STH efficiency was investigated for electrolyzers combining films produced using different sputtering conditions of the WO 3 (cathode) with Ni foam (anode) and A-CIGS solar cell modules with cell band gaps varying in-between 1.15 and 1.25 eV.In addition, electrochromic and electrocatalytic properties of WO 3 films coated on indium tin oxide (ITO) substrates were investigated in a potential range relevant for the HER, where the coloration efficiency was quantified.Finally, we discuss H + ion intercalation and the relation to the hydrogen evaluation process and possible applications of the reported effect.

Sample preparation 2.1.1. WO 3 film deposition
WO 3 films were deposited on 1.6 mm thick Ni foam (350 g/m 2 surface density and sheet resistance of 0.1 Ω/□) and ITO coated glass (sheet resistance of 60 Ω/□) substrates by reactive DC magnetron sputtering using a Balzers UTT 400 unit.The target consisted of a 5 cm diameter metallic W (99.99% purity, Plasmaterials) disc.The target to substrate distance was 13 cm, and the system was evacuated to ~8 × 10 -5 Pa.Pre-sputtering was performed for 5 min in argon plasma in order to remove surface impurities.The substrate holder was rotated at 3 rpm to improve the homogeneity of the films.Coating parameters are given in Table S1.The film thickness determined by a Veeco Dektak 150 surface profilometry instrument was 165 ± 10 and 300 ± 25 nm for films coated on Ni foam and ITO substrates, respectively.
Ni foam surface coverage of the WO 3 thin film and its homogeneity were analyzed by energy-dispersive X-ray spectroscopy (EDS).The uniform elemental coverage of the WO 3 on the Ni foam substrates was identified by EDS (Fig. 2).Due to the shadowing effect in EDS, one could not get the same quantification from the lower and the upper surfaces of the porous Ni foam substrate.However, since sputtering is a highly directional process, only the surface of Ni foam that was facing the sputter target should be expected to be coated.

A-CIGS solar cell module preparation
The A-CIGS material and subsequent solar cell modules were fabricated on soda-lime glass substrates cut to a 5 × 5 cm 2 sample size.The sample structure consists of a stack of thin layers of Mo, NaF, A-CIGS, CdS, ZnO, and ZnO:Al.The Mo layer was fabricated by DC magnetron sputtering to a thickness of 300-400 nm.Then, a 10 nm NaF layer was deposited by evaporation.The CdS layer was deposited by wet chemical bath deposition, using 1.2 M ammonia, 3.2 mM cadmium acetate, and 0.1 M Thiourea.The samples were immersed in a beaker with 175 mL solution, which was heated in a 60 • C water bath for 10 min, of which the first 6 min were without Thiourea.After the process, the samples were rinsed and dried with nitrogen.A double layer of undoped ZnO (70 nm) and a layer of aluminum-doped ZnO (210 nm) were deposited as a transparent conducting layer by rf magnetron sputtering.A Ni-Al-Ni grid was evaporated through a shadow mask by electron gun evaporation (about 2.5% area coverage) as the top electrode.
The A-CIGS layer was deposited by vacuum co-evaporation using a process dedicated to give high-quality material and with an in-depth grading of the Ga/(Ga+In) ratio through the film thickness obtained by using a high Ga/(Ga+In) evaporation rate in the beginning of the process and a lower one in the rest of the evaporation.The Ag/(Ag+Cu) ratio was 0.2.The Ga/(Ga+In) ratio and band gap energy (E g ) of the samples are listed in Table 1.After the A-CIGS deposition, the surface was modified by deposition of 5-10 nm of KF at a substrate temperature of 350 • C. Separation of cells was approached by mechanical scribing as described previously [9,17].

Characterization techniques
Scanning electron microscopy (SEM)-Energy-dispersive X-ray spectroscopy (EDS) measurements were performed on a Zeiss 1530 instrument using 10 kV electron accelerating voltage.X-ray mapping and elemental analysis were done within the Aztec software.X-ray photoelectron spectroscopy (XPS) studies were carried out on a PHI Quantera II spectrometer with Al Kα radiation with the energy of 1453 eV.All spectra were calibrated to the adventitious C 1s peak (284.8 eV).The films were not sputtered prior to measurement since preferential sputtering of elements can lead to deviation from real chemical composition and state [39].The XPS spectra were analyzed using the CasaXPS software.The analysis of the O 1s peak was not included in this work to avoid misinterpretation of the oxygen composition and state, which could be due to the exposure of the surface to the atmosphere, adventitious contamination, oxidation or water.X-ray diffraction (XRD) of the films deposited on glass substrates was recorded with 1 • of grazing angle using a grazing-incidence Siemens D5000 diffractometer with CuKα1 radiation at 1.5406 Å, 45 kV, and 40 mA.
Half-cell and full-cell reactions were characterized by linear sweep voltametry (LSV) measurements performed using a CH Instrument model 760C workstation with a scan rate of 5 mV s − 1 for hydrogen evolution of WO 3 thin films.The electrolyte was 0.5 M H 2 SO 4 (pH 0), which was purged by nitrogen flow for 5 min before the measurements.The half-cell measurements were performed with a three-electrode configuration at room temperature.An Ag/AgCl electrode (E 0 = 0.235 V at 25 • C) and a Pt wire were used as reference and counter electrode, respectively.
Electrochemical impedance spectroscopy (EIS) measurements of the WO 3 films were performed using a computer-controlled BioLogic Science Instruments SP-200 electrochemical interface.The measurements were done in the frequency range of 0.01 Hz-0.1 MHz using a 10 mV amplitude AC potential at − 0.4 V (where HER occurs) and 0 V (where no HER occurs) vs. RHE.
The optical transmittance of the WO 3 film was recorded in situ in the visible range with a setup that has a light-emitting diode (LED) and a photodiode sensor.The LED emitted light peaking at a wavelength of 528 nm, which is close to the middle of the luminous spectrum, and with a full width at half maximum of 33 nm for the intensity distribution.The 100-%-level for optical transmittance was taken as the value recorded with nothing but electrolyte in the cell.
The current density-potential (j-V) characteristics of the A-CIGS cells were recorded under simulated AM 1.5G sunlight in a set-up with a halogen lamp (ELH).The cells were kept at 25 • C during measurement.

Results
WO 3 thin films were deposited by sputtering onto Ni foam, and ITO coated glass, as detailed in the Experimental section.In order to assess the optimum sputtering parameters of WO 3 thin films for HER application, WO 3 thin films were sputtered using 32 different sputtering conditions (pressure, O 2 /Ar gas ratio, and power) in a relevant region of parameters that could show electrochromic activity.Sputter parameters of all the WO 3 thin films are shown in Table S1.A number of films deposited at low pressures and low O 2 /Ar ratios were dark as-deposited and displayed poor electrochromism (see Supplementary Information (SI)).These films exhibited a slightly lower HER performance, as observed by close inspection of Fig. S1 in SI.These films will not be discussed further below.To quantify the effects of the sputtering parameters on HER performance, LSV measurements were performed in the potential region from 0 to − 0.7 V vs. RHE.Current density (j) of WO 3 thin films coated on Ni foam substrate versus potential (V) is shown in Fig. 3 for selected films in order to show the effect of (a) different sputter pressures and (b) different O 2 /Ar ratios.All HER measurements that were performed for WO 3 thin films on Ni foam are shown in Fig. S2a in SI.The overpotential (OVP) varied between 182 and 322 mV vs. RHE at 10 mA cm − 2 where it was 610 mV for bare Ni foam (Fig. S2b).The effect of sputtering power was negligible, as depicted in Fig. S1a in SI.There was not any observable trend connecting the j − V characteristics of the thin films with the O 2 /Ar ratio and pressure (Fig. 3a and b) during sputtering, as well.The Tafel slope of the WO 3 thin film on Ni foam showed a considerable variation between 45 and 110 mV dec − 1 (Fig. 3c  and d).This indicates that the HER follows a rate-determining Volmer or Heyrovsky step for different sputtering conditions without any order [40,41].The values of the Tafel slope could not be correlated to the sputter parameters.
WO 3 -based electrolyzers were built using WO 3 thin films coated on Ni foam as cathodic and uncoated Ni foam as anodic parts of the system.The performance of the electrolyzers was tested in 0.5 M H 2 SO 4 from 0 to 2 V for the different WO 3 thin films.The load curves for the WO 3 (HER)-Ni foam (oxygen evolution reaction (OER)) electrolyzers are shown in Fig. 4 for some selected sputtering conditions.The load curves were similar for the electrolyzers with different WO 3 thin films and the lowest potential needed for 10 mA cm − 2 in the overall reaction was 1.77 V.In order to further quantify the performance of the WO 3 films as cathodic electrocatalytic materials under acidic conditions, the results were compared with a known electrocatalyst configuration for hydrogen production.j-V characteristics of NiMo coated on Ni foam (HER) combined with NiO on Ni foam (OER) electrolysis in 1 M KOH are shown in Fig. 4. The measurements were done in a 5 mL cylindrical electrochemical cell without membrane and with a 1 cm interdistance between the electrodes.It is seen that j-V characteristics of the different systems are very similar in the operation range of a solar cell.Therefore, it can be concluded that WO 3 under acidic conditions is able to compete with known alkaline electrocatalysts for hydrogen production, where higher current densities naturally can be obtained upon further nanostructuring and by increasing the total internal surface area for either WO 3 , NiMo, or NiO.
STH efficiency was investigated for a solar water splitting system consisting of a PV-electrolyzer with WO 3 as an electrocatalytic cathodic material.In order to find an optimum efficiency of the PV-electrolysis, different combinations of the electrolyzer with A-CIGS-based thin film solar cell modules with different band gaps of the cell were examined.j-V measurements were done for the electrolyzer and the cells, as well as extended to 3-cell A-CIGS modules to generate a photovoltage large enough to fulfill the energy need of the electrolyzer.STH was estimated from the intersection of the j-V curves for the same area of the modules and the electrodes of the electrolyzer.The electrolyzer load curve can be shifted to the left and provide higher catalytic current at a given potential up to the current plateau but is here kept at the same area as the PV-module to more easily compare with previously published PV/photoelectrochemical approaches.This also allows comparison with STH found from direct photocatalysis, naturally sharing the same area of the photoabsorber and the area available for catalysis.Current densities versus the potential for 3-cell A-CIGS modules with different band gap energies and an electrolyzer with a WO 3 on Ni foam (HER) and Ni foam (OER) are shown in Fig. 5.The effect of band gap changes are seen in the j-V characteristics of the A-CIGS module.An increase in the STH efficiency from 12% to 13% resulted from lowering the bandgap from 1.25 eV to lower values in order to increase the photocurrent while still having a sufficient photovoltage to perform the full water splitting reaction (Fig. S3).A similar situation could be obtained by using slightly concentrated solar light but with increase of the likely cost of the system quite substantially.However, in order to achieve a more robust system, with less sensitivity to the sputtering parameters of the WO 3 films and allowing for variation in the active area and solar illumination conditions, the A-CIGS module with 1.25 eV band gap a cell would be preferred.This module was chosen for further study in order to reduce the effect of the module voltage on the STH since the intersection is in the plateau region of the j-V curve of the module.Thus,   the variation of the STH came mainly from the j-V characteristic of the electrolyzer and varied between 12.35% and 12.49% for the different sputtering conditions used to deposit WO 3 films.The results showed that the sputtering parameters of WO 3 films are less significant for achieving a high STH efficiency of the A-CIGS-Ni foam-WO 3 system (1% unit variance in STH) if a suitable band gap matching is done.
The above reported results are from a Ni foam supported approach where the electrochromic properties of the catalyst can be used to study the state or the health of the cathodic catalyst by viewing the surface color.However, reflectance modulations are typically rather small, and a much better sensitivity would be achieved by using a transmitting substrate.Furthermore, in order to thoroughly analyze the coloration efficiency and the electrochromic effect, a transparent electrode was also applied.In addition, for design of a combined electrochromic and electrocatalytic hydrogen evolution system where the transmission properties are in focus, Ni foam has to be replaced by other alternatives.Here, we deposited a WO 3 film onto a transparent and electrically conducting ITO coated glass electrode and analyzed the charge incorporation, transmitted light, and coloration efficiency in a proof-ofconcept for electrochromic HER.
Before studying the electrochromic HER performance of WO 3 , we investigated the role of the substrate on the catalytic activity by EIS and electrochemical real surface area (ECSA) analysis.EIS measurements were done on WO 3 thin films deposited on Ni foam and ITO coated glass substrates.The measurements were performed at DC potentials of 0 and − 0.4 V vs. RHE which represented the non-reaction and reaction cases.The Nyquist plots of the experimental data are shown in Fig. 6a and b for the films on Ni foam and ITO, respectively.It should be noted that the highest frequencies are at the left-hand side of the Nyquist plots and that frequencies decrease as one follows the curves towards the right.For the WO 3 films on Ni foam, the impedance response with a DC potential of 0 V vs. RHE where there was not any HER showed two almost overlapping arcs, and was dominated by the second arc in the lower frequency range with higher impedance values.Under HER condition (− 0.4 V vs. RHE), only one arc was seen, probably due to the supremacy of the reaction impedance.For WO 3 on ITO, the high--frequency arc was not complete due to the limit of the instrument at high-frequencies but was large enough to be clearly distinguished from the low-frequency behavior.For both substrates, the lower absolute impedance in the low-frequency range was seen when HER occurred compared to the measurement at the non-reaction state.Impedance values were noticeably higher for the films coated on ITO and underline that the substrate-catalyst interface also is important for the overall catalytic performance as recently highlighted [42].
The EIS data were fitted with an equivalent circuit model, embedded in Fig. 6a [42].The model consists of a series connection of a resistance (Rs) with a parallel connection of a high frequency resistance (R1) and a high frequency constant phase element (CPE1), and a parallel connection of a low frequency resistance (R2) and a low frequency constant phase element (CPE2).The impedance function of a constant phase element is given as where ω is angular frequency, T is a parameter related to the electrode capacitance, and P is the constant phase exponent.The elements R s is related to the electrolyte and electrical connections, the high frequency part is related to the surface geometry where R1 is the charge transfer resistance, and the low-frequency part is associated with the reaction kinetics with the reaction resistance of R2 that depends on the applied overpotential.The resistance values of the WO 3 films on Ni foam and ITO substrates at potentials corresponding to non-HER and HER states are given in Table 2. Uniform deposition of WO 3 thin films on ITO at 30 mTorr, 200 W and 0.90 O 2 /Ar was proven by SEM-EDS analysis (Fig. 7a).The O/W ratio of the film deposited on ITO at 30 mTorr, 200 W and 0.90 O 2 /Ar was reported as 3.44 in a recent study, using Rutherford backscattering spectroscopy (RBS) techniques [43].Fig. 7b shows the high resolution tungsten (W 4f) peaks for the WO 3 films deposited under the same conditions as the ones for the SEM-EDS and RBS measurements.The doublet of W 4f was observed at the binding energies of 35.3 and 37.4 eV corresponding to W 4f 7/2 and W 4f 5/2 , respectively.The energy separation between the main peaks of W(4f 7/2 ) and W (4f 5/2 ) was 2.12 eV.The results are in good agreement with those of WO 3 and that the tungsten (W) in the WO 3 film was present in the six-valence state (W 6+ ) [44,45].From the XRD analysis of the WO 3 films on glass, the films were amorphous (Fig. 7c); which is preferable for electrochromic applications as crystalline electrochromic films have lower ionic mobility than amorphous films that lowers the switching speed.Cyclic voltammetry measurements were performed (Fig. S4) to determine the ECSA of WO 3 catalysts on Ni foam and ITO substrates.The ECSAs of the WO 3 catalyst on Ni foam was five times larger than that of ITO.In principle, with neglect of changed recombination behavior, a catalyst with a larger ECSA is expected to provide an enhanced catalytic activity in correspondence with the increased exposed surface area.The ratio of the current density of the WO 3 thin films on Ni foam (− 34 mA cm − 2 , Fig. 3) and on ITO (− 5 mA cm − 2 , Fig. 8) at − 0.4 V vs RHE was 6.8.Thus, the increased catalytic activity seen in the Ni-foam can be mainly (74%) be attributed to the surface area enhancement and a minor effect (26%) to the lower barrier formed in-between the catalyst and the conducting substrate [42].
The electrocatalytic hydrogen evolution activity of the WO 3 thin film coated on ITO was characterized by LSV at a scan rate of 5 mV s − 1 in aqueous 0.5 M H 2 SO 4 (Fig. 8).In-situ optical transmittance measurements were performed at a wavelength of 528 nm with in-situ   photographs of the colored (during HER) with evolving H 2 gas and bleached state of the WO 3 electrocatalyst shown in Fig. 8 (see Supplementary video 1 which shows coloration during LSV).The 100-%-level for the transmittance corresponded to the value recorded with nothing but electrolyte in the measurement cell.The WO 3 thin films were in the bleached state without HER and in the colored state during HER.In the initial state, transmittance was 78% and decreased to ~10%, while delivering a current density of − 5 mA cm − 2 at an OVP of 400 mV, which was 0.014 mA cm − 2 for uncoated ITO (Fig. S5).We limited the LSV measurement at this potential because it was in the range for the application in solar water splitting.This limit is very close to the saturation level of H + ion intercalation (Fig. S6).Integrating the current density from 0 to − 0.6 V vs Ag/AgCl and assuming that all electrons compensate the intercalated H + ions, one reaches an ion insertion ratio of 0.72 H + per W (Fig. 8).This value is of the same order and slightly higher than a saturated ion insertion ratio of 0.56 Li + versus W recently reported [28].In the present situation, the saturation plateau is at a potential slightly higher than the voltage used in the water splitting and also at a higher insertion ratio due to the simultaneous consumption of electrons in the HER reaction.Using the data for the wider voltage range (Fig. S6), the apparent saturation of the H + ion insertion occurs around a ratio of 0.8 H + /W, and assuming that H + would have the same saturation ratio as Li + /W (0.56), it implies that 0.24 H + /W of the apparent ion saturation is related to the consumption of electrons in the HER.With a fixed applied potential as in a normal application of alkaline electrolysis, however, there is creation of a steady-state situation for the coloration and the majority of the cathodic current will instead be consumed by the HER process with formation of a cathodic current plateau (Fig. 8).
Recalling Eq. ( 6), the time-dependent intercalation level can be quantified by integrating the current density participating in the coloration and obtain the number of charges inserted per unit area.For [1-R t (λ,x)] > T t (λ,x) and a low variation in the total reflection R t (λ,x) in Eq. ( 8), the differential coloration efficiency can be extracted with The corresponding coloration efficiency at a wavelength of 528 nm and the OVP of 400 mV was 40 cm 2 C − 1 .
A short demonstration of hydrogen evolution of the WO 3 thin film at 400 mV OVP can be found in the Supplementary video 2.

Discussion and concluding remarks
In this study, water electrolysis of WO 3 thin films for hydrogen production in the potential range pertinent to electrochromic coloring was examined.The results showed that WO 3 under acidic conditions (0.5 M H 2 SO 4 ) could compete with well-known non-precious catalysts for hydrogen production under alkaline conditions.Films of WO 3 deposited on transparent substrates showed combined electrocatalytic and electrochromic properties, which were quantified and showed a change in transmission of visible light from 78% as an inactive electrocatalyst to 10% under the HER.The increased absorption in the film would be beneficial for increased heating of the catalyst area during operation if accessible by light or could be used for monitoring the charge state of the catalyst for process control.Monitoring of the charge state and health of the cathodic WO 3 catalyst film could also be carried out in reflection mode either by a simple low angle diode monitoring the reflected light or by utilization of reflection from a partly reflecting OER electrode.In transmission mode, solar heat and visible light through the thin film catalyst is modulated under operation, which can provide electrochromic functionality to water splitting devices for monitoring the state of HER or even pointing the way to possible future building integration.However, the latter option requires transparent electrodes to be also employed for the OER.The highest STH efficiency obtained by combining A-CIGS materials with varying band gaps, with electrolyzer units using WO 3 on Ni foam was about 13% for the range of WO 3 thin films reported.The lowest HER overpotential we find for the deposited WO 3 films (182 mV@10 mA cm − 2 ) is in the same range or slightly higher compared to previously reported values for undoped WO 3 [8] while being in the higher range of HER overpotentials reported for high surface area catalysts of iron-based phosphides (30-235 mV@10 mA cm − 2 ) [46] and FeNi-layered double hydroxides (down to 59 mV@10 mA cm − 2 ) [47].Hetero-atom doping and strategies to increase the active surface area of WO 3 , however, have resulted in HER overpotentials down to 38 mV@10 mA cm − 2 versus RHE [8] and show the promise for WO 3 -based catalyst systems that additionally allows electrochromic functionality.Utilization of a combined electrochromic and water electrolytic effect can readily be utilized with monitoring capabilities, while a system for building integration and light flux control critically depends on transparent conducting electrodes for both HER and OER and requires more work on optimizing, for example, by increasing the surface area of transparent conducting electrodes, and at the same time minimize light scattering.This is indeed very challenging but would allow building-integrated hydrogen production with the added benefits of controlling the temperature or solar heat absorption by the tunable optical modulation of the electrochromic material.
CRediT authorship contribution statement I.B.P. and T.E.initialized the research project and conceptualized the approach.I.B.P. and G.A. performed the experimental work on the catalysts.L.S. and M.S. provided the (Ag,Cu)InGaSe 2 solar cell and module manufacturing facilities.L.S., M.E., I.B.P., and T.E.provided the initial assessment of Ga variation and suitable band gap matching with the catalyst system.I.B-P and TE wrote the original draft and all authors participated in reviewing and editing the final manuscript.G.A.N. and T. E. provided instrumentation and participated in the analysis of the electrochemical characterization.I.B.P, L.S., M.S., and T.E.provided materials, access to characterization tools, and acquisition of the financial support for the project.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig. 1 .
Fig. 1.Schematic picture of PEM electrolysis (a) and a window type electrochromic device (b), where M + denotes the inserting ion.

Fig. 2 .
Fig. 2. Electron image and elemental mapping of WO 3 thin film on Ni foam.

Fig. 3 .
Fig. 3. Linear sweep voltammetry recorded in 0.5 M H 2 SO 4 with 5 mV s − 1 scan rate, showing the effect of varying (a) O 2 /Ar ratio with 30 mTorr pressure and 200 W power, (b) sputter pressure with 0.90 O 2 /Ar ratio and 200 W power, and (c, d) Tafel plots of (a, b) for electrocatalytic HER of WO 3 thin films on Ni foam.

Fig. 4 .
Fig. 4. Potential dependence of current density of overall water splitting recorded with 5 mV s − 1 scan rate for the WO 3 (HER)-Ni foam (OER) electrolyzers in 0.5 M H 2 SO 4 and a NiMo (HER)-NiO (OER) electrolyzer in 1 M KOH.

Fig. 5 .Fig. 6 .
Fig. 5. Potential dependence of current density for 3-cell A-CIGS modules with different cell band gaps and the load curve for the Ni foam (OER)-WO 3 deposited at 30 mTorr, 200 W and 0.90 O 2 /Ar (HER) electrolyzer in 0.5 M H 2 SO 4 .

Fig. 7 .
Fig. 7. (a) SEM-EDS images with W and O distribution, (b) high resolution XPS spectra of the W 4f doublet core level, and (c) XRD pattern of a WO 3 film deposited at 30 mTorr, 200 W and 0.90 O 2 /Ar on ITO-coated glass substrate.

2 )Fig. 8 .
Fig. 8. Linear sweep voltammogram (scanned from positive to negative) at a rate of 5 mV s − 1 in 0.5 M H 2 SO 4 and in-situ optical transmittance at a wavelength of 528 nm and transmittance as a function of ion intercalation H + / W ratio together with photographs of the colored and bleached states of the WO 3 thin films.The WO 3 film was deposited at 30 mTorr, 200 W and 0.90 O 2 /Ar on ITO.

Table 1
Ga/(Ga+In) ratio and E g of A-CIGS.

Table 2
Resistance elements from the fits of the experimental impedance data for WO 3 thin films on Ni foam and ITO substrates at the shown potentials.