Final Report Summary - COCHALPEC (Development of electrodes based on copper chalcogenide nanocrystals for photoelectrochemical energy conversion)
Hydrogen has been touted as a promising energy vector to replace the fossil fuels—paving the way for a renewable and sustainable economy. However, to date, the production of hydrogen, which is also in high demand for the chemical industry, mostly relies on non-renewable routes, like the steam reforming of natural gas (methane). With the aim to address the plan envisaged for the European Union wherein by 2050 the CO2 emission should be cut drastically by 90% compared to the 1990 levels, the project COCHALPEC attempted to develop a cost-efficient and sustainable route to produce hydrogen by using photoelectrochemical cells that employed the solar energy to directly reduce the water into H2. More concretely, the project focused on the design of novel solution process routes to fabricate highly efficient photoelectrodes for water splitting starting with colloidal dispersions of Cu-based chalcogenides nanocrystals (NCs) that are employed as building blocks.
The project COCHALPEC was constructed on the following objectives:
- Demonstration of facile methods for the size and composition control of ternary and quaternary copper chalcogenides.
- Investigate and optimize the effects of size, composition and shape of the nanocrystals on the photoelectrochemical performance of 3D arrays and sensitized electrodes.
- Enhance the stability and catalytic performance of the electrodes by applying state of the art overlayers and catalysts.
- Validate the tandem cell concept for the overall solar water splitting using chalcogenide NCs based photoelectrodes.
Semiconductor nanocrystals of CdS, Cu2ZnSnS4 (CZTS), Cu2ZnSnSe4, CuInxGa1-xS2 (CIGS), CuInxGa1-xSe2, ZnCuInS2, CuFeS2 were synthesized using colloidal routes. The synthetic procedures were adapted and optimized according our needs (viz. tight control on the NC size and composition as well as high NC concentration –colloidal “inks”-). Given the lower stability of the selenides for the photoreduction of water the studies were focused in sulfides. The characterization by X-ray diffraction (XRD) and Raman spectroscopy confirmed the nature of the products, whereas the transmission electron microscopy evidence the morphology of the NCs and the UV-Vis-nIR spectra characterize the optical properties. Interestingly, whereas in the case of the CZTS large deviations in stoichiometry causes phase segregation in the chalcopyrite CIGS interesting changes in the optical properties were obtained by changing the composition, with special interest on the optical properties of copper-poor CIS and CGS materials.
In order to prepare the electrodes for implementation in photoelectrochemical cells, four different approaches were tested: (i) the assembly of 3D arrays of NCs, (ii) the sensitization of p-type scaffolds, (iii) mesostructuring the NCs thin films and (iv) the fabrication of polycrystalline electrodes using the NCs as raw materials.
The fabrication of 3D arrays of NCs of CZTS (i) was initially carried out by using the layer-by-layer method where molecules such as 1,2-ethanedithiol, 1,4-benzenedithiol and ethanediamine were used as cross-linking agents. However, this technique turned out to be too tedious, poorly reproducible and critically time-consuming, slowing down the progress into our investigations. Alternatively, a tape-casting approach followed by a mild annealing at 275°C to prepare films using highly concentrated NC inks resulted to be much more desirable. As a result, thin films with a compact packaging of CZTS NCs were obtained, as confirmed the Scanning Electron Microscopy (SEM) images. It is well known that the CZTS, like other ternary/quaternary chalcogenides required the deposition of n-type overlayers to promote the extraction of photogenerated carriers. Traditionally, the n-CdS is the ubiquitous choice to create this pn-junction, though identifying other alternative materials could potentially lead to a more efficient charge extraction, and accordingly, to higher solar-to-hydrogen (STH) efficiencies. Using the CZTS NCs photocathodes we studied the effect of different surface modifications to enhance the extraction of photogeneration carriers in the CZTS. By using a successive ionic layer adsorption and reaction (SILAR) method ultrathin and conformal overlayers of difference chalcogenide (CdS, CdSe and ZnSe) were deposited. We found that the overlayers of CdSe and ZnSe on CZTS indeed led to better performance than those obtained for the CdS-modified films. Additionally, the specific adsorption of methylviologen (MV) on the surface of the modified CZTS films demonstrated further improvement of the photoelectrochemical response in all the cases. Results suggested that these molecules play a double role: (1) causing a downward shift of the energy bands then shifting the onset of photocurrent to more positive potentials and (2) mediating in the electron transfer from the electrode to the electrolyte working as a reversible redox couple attached on the surface of the film. However, despite the enhanced photoelectrochemical response with the treatments, the performance was observed to be far below the theoretical maximum photocurrents which suggests a high degree of recombination of the photogenerated charges—likely due to the high density of grain boundaries in the compact film.
The reduced active surface area of the compact films (i.e. the electrode/electrolyte interfacial area) could potentially be addressed by mesostructuring the photoelectrode. Two different routes were envisaged to mesostructured the photoelectrodes, and hence, enable an easy infiltration of the electrolyte throughout the film increasing the active area. First, the sensitization of mesostructured p-type scaffolds with the NCs (ii) was supposed to enhance the overall performance by accelerating the charge separation at the scaffold/NC interface. As p-type scaffolds, materials like CuSCN and CsSnI3 were tested with poor results, achieving the best results with mesoporous films of NiO prepared by either sol-gel or commercial powders. The NCs (CdS and CZTS) were adsorbed directly or by using bridge molecules to the oxide. Unfortunately, the performance was even lower than the one for the compact film, likely because of the high degree of recombination that exist also in the oxide film and in the additional NC/NiO heterojunction. Alternatively, we developed a novel approach to directly mesostructured the nanocrystalline film (iii) in one step. One of the most extended approaches to mesostructured nanocrystalline film entails mixing the nanoparticles with a porogen that can be removed by annealing creating in such a way porosity in the film. This strategy is widely employed for oxide nanoparticles, which remain stable in the conditions required to remove the porogens (> 450°C, air atmosphere). However, most of chalcogenides are prone to oxidation and decomposition when annealing in air or at temperatures higher than 300°C. We found that the nitrocellulose can be used as a promising porogen taking advantage of its low auto-decomposition temperature (210°C) and the almost negligible carbon residue at this temperature even when annealing under inert atmosphere. The results obtained for CdS and CZTS thin films revealed that by adjusting the NC:nitrocellulose ratio it is possible to tightly control the porosity of the films and, in turn, the active surface area. Results for CdS clearly evidence that by mesostructuring the film and increasing the surface area an 8-fold enhancement in photocurrent was obtained. More importantly the incident-photon to current-efficiency (IPCE) values equalize the best performance of CdS-sensitized films, the most extended CdS-based electrodes nowadays, but offering an easier processability, one step fabrication.
Aside from mesostructuring, another route to reduce the high degree of recombination of the compact films would be to reduce the grain boundaries by inducing the coalescence of the nanocrystals. To date, no reports on the grain growth of the ternary/quaternary sulfides have been reported. We developed a technology wherein the compact film of nanocrystals of CZTS or CIGS deposited on Mo foils were treated with SbCl3, BiCl3, MgCl2, CdCl2, KCl among other potential grain growth promoters. We found that Sb, Bi and Mg induces a significant coalescence (500 nm – 5 μm) of the NCs (10-20 nm). Interestingly, Sb and Bi give rise to grains of about 500 nm yielding the best performing films, whereas Mg showed a very poor performance probably connected to the large grain growth obtained (> 3 μm) which could account for the charge transport issues. The optimized depositions of solution processable overlayers of CdS, TiO2 and Pt brought out record photocurrent (equivalent to STH of ~ 10% in a tandem cell) for the films of CuInS2 with a faradaic efficiency of 100% without noticing degradation for more than 4 hours in operation conditions. Likewise, fabrication of a tandem cell including this electrode and an n-type W:BiVO4/Co-Pi film delivers stable conversion efficiencies of ~1 %. Unfortunately, the efficiencies experimentally attained in tandem cell were far from the expected 10 %, and below the 5 % pursued in the project due to the mismatch between the onset of photocurrent of the photoanodes/photocathodes employed. Efforts to shift the onset of photocurrent of these Cu-based chalcogenides toward more positive potentials should be carried out to efficiently combine this chalcogenides with state-of-the-art photoanodes.
In the framework of the project COCHALPEC, new technologies to completely solution process highly efficient and stable photocathode for water splitting were developed. These results open a new doorway for the cost-efficient fabrication of photocathodes for water splitting and could potentially pave the way for the implementation of photoelectrochemical cells as feasible instruments for the production of H2 in a clean and renewable way.
The project COCHALPEC was constructed on the following objectives:
- Demonstration of facile methods for the size and composition control of ternary and quaternary copper chalcogenides.
- Investigate and optimize the effects of size, composition and shape of the nanocrystals on the photoelectrochemical performance of 3D arrays and sensitized electrodes.
- Enhance the stability and catalytic performance of the electrodes by applying state of the art overlayers and catalysts.
- Validate the tandem cell concept for the overall solar water splitting using chalcogenide NCs based photoelectrodes.
Semiconductor nanocrystals of CdS, Cu2ZnSnS4 (CZTS), Cu2ZnSnSe4, CuInxGa1-xS2 (CIGS), CuInxGa1-xSe2, ZnCuInS2, CuFeS2 were synthesized using colloidal routes. The synthetic procedures were adapted and optimized according our needs (viz. tight control on the NC size and composition as well as high NC concentration –colloidal “inks”-). Given the lower stability of the selenides for the photoreduction of water the studies were focused in sulfides. The characterization by X-ray diffraction (XRD) and Raman spectroscopy confirmed the nature of the products, whereas the transmission electron microscopy evidence the morphology of the NCs and the UV-Vis-nIR spectra characterize the optical properties. Interestingly, whereas in the case of the CZTS large deviations in stoichiometry causes phase segregation in the chalcopyrite CIGS interesting changes in the optical properties were obtained by changing the composition, with special interest on the optical properties of copper-poor CIS and CGS materials.
In order to prepare the electrodes for implementation in photoelectrochemical cells, four different approaches were tested: (i) the assembly of 3D arrays of NCs, (ii) the sensitization of p-type scaffolds, (iii) mesostructuring the NCs thin films and (iv) the fabrication of polycrystalline electrodes using the NCs as raw materials.
The fabrication of 3D arrays of NCs of CZTS (i) was initially carried out by using the layer-by-layer method where molecules such as 1,2-ethanedithiol, 1,4-benzenedithiol and ethanediamine were used as cross-linking agents. However, this technique turned out to be too tedious, poorly reproducible and critically time-consuming, slowing down the progress into our investigations. Alternatively, a tape-casting approach followed by a mild annealing at 275°C to prepare films using highly concentrated NC inks resulted to be much more desirable. As a result, thin films with a compact packaging of CZTS NCs were obtained, as confirmed the Scanning Electron Microscopy (SEM) images. It is well known that the CZTS, like other ternary/quaternary chalcogenides required the deposition of n-type overlayers to promote the extraction of photogenerated carriers. Traditionally, the n-CdS is the ubiquitous choice to create this pn-junction, though identifying other alternative materials could potentially lead to a more efficient charge extraction, and accordingly, to higher solar-to-hydrogen (STH) efficiencies. Using the CZTS NCs photocathodes we studied the effect of different surface modifications to enhance the extraction of photogeneration carriers in the CZTS. By using a successive ionic layer adsorption and reaction (SILAR) method ultrathin and conformal overlayers of difference chalcogenide (CdS, CdSe and ZnSe) were deposited. We found that the overlayers of CdSe and ZnSe on CZTS indeed led to better performance than those obtained for the CdS-modified films. Additionally, the specific adsorption of methylviologen (MV) on the surface of the modified CZTS films demonstrated further improvement of the photoelectrochemical response in all the cases. Results suggested that these molecules play a double role: (1) causing a downward shift of the energy bands then shifting the onset of photocurrent to more positive potentials and (2) mediating in the electron transfer from the electrode to the electrolyte working as a reversible redox couple attached on the surface of the film. However, despite the enhanced photoelectrochemical response with the treatments, the performance was observed to be far below the theoretical maximum photocurrents which suggests a high degree of recombination of the photogenerated charges—likely due to the high density of grain boundaries in the compact film.
The reduced active surface area of the compact films (i.e. the electrode/electrolyte interfacial area) could potentially be addressed by mesostructuring the photoelectrode. Two different routes were envisaged to mesostructured the photoelectrodes, and hence, enable an easy infiltration of the electrolyte throughout the film increasing the active area. First, the sensitization of mesostructured p-type scaffolds with the NCs (ii) was supposed to enhance the overall performance by accelerating the charge separation at the scaffold/NC interface. As p-type scaffolds, materials like CuSCN and CsSnI3 were tested with poor results, achieving the best results with mesoporous films of NiO prepared by either sol-gel or commercial powders. The NCs (CdS and CZTS) were adsorbed directly or by using bridge molecules to the oxide. Unfortunately, the performance was even lower than the one for the compact film, likely because of the high degree of recombination that exist also in the oxide film and in the additional NC/NiO heterojunction. Alternatively, we developed a novel approach to directly mesostructured the nanocrystalline film (iii) in one step. One of the most extended approaches to mesostructured nanocrystalline film entails mixing the nanoparticles with a porogen that can be removed by annealing creating in such a way porosity in the film. This strategy is widely employed for oxide nanoparticles, which remain stable in the conditions required to remove the porogens (> 450°C, air atmosphere). However, most of chalcogenides are prone to oxidation and decomposition when annealing in air or at temperatures higher than 300°C. We found that the nitrocellulose can be used as a promising porogen taking advantage of its low auto-decomposition temperature (210°C) and the almost negligible carbon residue at this temperature even when annealing under inert atmosphere. The results obtained for CdS and CZTS thin films revealed that by adjusting the NC:nitrocellulose ratio it is possible to tightly control the porosity of the films and, in turn, the active surface area. Results for CdS clearly evidence that by mesostructuring the film and increasing the surface area an 8-fold enhancement in photocurrent was obtained. More importantly the incident-photon to current-efficiency (IPCE) values equalize the best performance of CdS-sensitized films, the most extended CdS-based electrodes nowadays, but offering an easier processability, one step fabrication.
Aside from mesostructuring, another route to reduce the high degree of recombination of the compact films would be to reduce the grain boundaries by inducing the coalescence of the nanocrystals. To date, no reports on the grain growth of the ternary/quaternary sulfides have been reported. We developed a technology wherein the compact film of nanocrystals of CZTS or CIGS deposited on Mo foils were treated with SbCl3, BiCl3, MgCl2, CdCl2, KCl among other potential grain growth promoters. We found that Sb, Bi and Mg induces a significant coalescence (500 nm – 5 μm) of the NCs (10-20 nm). Interestingly, Sb and Bi give rise to grains of about 500 nm yielding the best performing films, whereas Mg showed a very poor performance probably connected to the large grain growth obtained (> 3 μm) which could account for the charge transport issues. The optimized depositions of solution processable overlayers of CdS, TiO2 and Pt brought out record photocurrent (equivalent to STH of ~ 10% in a tandem cell) for the films of CuInS2 with a faradaic efficiency of 100% without noticing degradation for more than 4 hours in operation conditions. Likewise, fabrication of a tandem cell including this electrode and an n-type W:BiVO4/Co-Pi film delivers stable conversion efficiencies of ~1 %. Unfortunately, the efficiencies experimentally attained in tandem cell were far from the expected 10 %, and below the 5 % pursued in the project due to the mismatch between the onset of photocurrent of the photoanodes/photocathodes employed. Efforts to shift the onset of photocurrent of these Cu-based chalcogenides toward more positive potentials should be carried out to efficiently combine this chalcogenides with state-of-the-art photoanodes.
In the framework of the project COCHALPEC, new technologies to completely solution process highly efficient and stable photocathode for water splitting were developed. These results open a new doorway for the cost-efficient fabrication of photocathodes for water splitting and could potentially pave the way for the implementation of photoelectrochemical cells as feasible instruments for the production of H2 in a clean and renewable way.