CO2 CONVERSION WITH NITRIDE SEMICONDUCTOR-METAL INTERFACE

Description

CROSS-REFERENCE TO RELATED APPLICATION

This application claims the benefit of U.S. provisional application entitled “CO₂ Conversion with Nitride Semiconductor-Metal Interface,” filed Aug. 20, 2024, and assigned Ser. No. 63/685,137, the entire disclosure of which is hereby expressly incorporated by reference.

STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT

This invention was made with government support under Contract No. W911NF-21-1-0337 awarded by the U.S. Army. The government has certain rights in the invention.

BACKGROUND OF THE DISCLOSURE

Field of the Disclosure

The disclosure relates generally to photoelectrochemical and other chemical conversion of carbon dioxide (CO₂).

Brief Description of Related Technology

Artificial photosynthesis which transforms CO₂, water, and sunlight to highly valuable chemicals and fuels, is a promising technology for energy and environmental sustainability. This process can be accelerated in a solar-powered photoelectrochemical CO₂ reduction reaction (PEC CO₂RR) system. To date, however, only small amounts of C₁ products, such as CO, CH₄, HCOOH, have been reported with desirable efficiency, while the more valuable multi-carbon (C₂₊) products (such as ethylene) are rarely achieved with high efficiency using PEC CO₂RR.

For example, the typically reported Faradic efficiency (FE), activity and stability for ethylene production by PEC CO₂RR in aqueous solution are lower than 30%, 2.5 mA/cm² and 20 h, respectively. Although the FE of ethylene can be improved by using an organic electrolyte with a proton donor, the complex conditions, low current density (0.6 mA/cm²) and stability (6 hours) are still limited for practical applications. Therefore, there is an urgent need to develop a new avenue to overcome the efficiency and stability bottleneck of PEC CO₂RR to multi-carbon products under mild conditions.

Copper (Cu) is one of the best catalysts for deep CO₂RR both in PEC and electrocatalysis (EC) systems, but the selectivity for multi-carbon or single multi-carbon products is still challenging. Partially oxidized Cu species and low-coordinated Cu sites (defects or nanograins) are typically considered as the active structure for C—C coupling to multi-carbon products. However, rationally constructing these catalytic structures is difficult both in electrocatalysis and photo-electrocatalysis, because it is challenging to stabilize low-coordinated Cu sites and to generate or stabilize the partially oxidized Cu species under strong and continuous reduction conditions. Furthermore, it is more complex when combining the active structure in the photocathode, due to the simultaneous considerations of photo-absorbers, chemical corrosion and photodegradation. It has remained challenging to identify a suitable platform and a desirable pathway to overcome these issues, in both electrochemical and photochemical processes.

SUMMARY OF THE DISCLOSURE

In accordance with one aspect of the disclosure, a catalytic device includes a substrate having a surface, an array of conductive projections supported by the substrate and extending outward from the surface of the substrate, each conductive projection of the array of conductive projections including a nitride semiconductor, and a plurality of nanoclusters disposed over the array of conductive projections, each nanocluster of the plurality of nanoclusters including a metal. Each nanocluster of the plurality of nanoclusters is coupled to a respective conductive projection of the array of conductive projections via an oxidized interface. The oxidized interface includes an oxide species of the metal. The oxidized interface further includes an oxynitride species based on the nitride semiconductor.

In accordance with another aspect of the disclosure, a method of fabricating a catalytic device includes forming an array of conductive projections on a semiconductor substrate, each conductive projection of the array of conductive projections including a nitride semiconductor, depositing a plurality of nanoclusters across the array of conductive projections, each nanocluster of the plurality of nanoclusters including a metal, and, after depositing the plurality of nanoclusters, implementing a procedure to form an oxidized interface between each nanocluster of the plurality of nanoclusters and a respective conductive projection of the array of conductive projections. The oxidized interface includes an oxide species of the metal. The oxidized interface further includes an oxynitride species based on the nitride semiconductor.

In connection with any one of the aforementioned aspects, the devices, systems, and/or methods described herein may alternatively or additionally include or involve any combination of one or more of the following aspects or features. The oxidized interface is configured for catalytic conversion of carbon dioxide (CO₂). Each nanocluster of the plurality of nanoclusters has a distorted lattice. Each nanocluster of the plurality of nanoclusters has a defective surface. The oxidized interface has a sub-nanometer thickness. Surfaces of each nanocluster of the plurality of nanoclusters not at the oxidized interface are not oxidized. Each conductive projection of the array of conductive projections is covered by the oxynitride species away from the nanoclusters. The metal is a d-block metal. The metal is copper. The nitride semiconductor is a III-nitride semiconductor. The nitride semiconductor is gallium nitride (GaN). The nitride semiconductor is doped n-type. The substrate includes a semiconductor material. The semiconductor material is doped to define a junction to generate charge carriers upon absorption of solar radiation. Each conductive projection of the array of conductive projections includes a nanowire configured to extract the charge carriers generated in the substrate. An electrochemical system includes a working electrode configured in accordance with the catalytic device disclosed or claimed herein, and further includes a counter electrode, an electrolyte in which the working and counter electrodes are immersed, and a voltage source that applies a bias voltage between the working and counter electrodes. The bias voltage is set to a level for conversion of CO₂ into ethylene at the working electrode. The electrolyte lacks an organic solvent. The oxidized interface is configured for catalytic conversion of carbon dioxide (CO₂). Implementing the procedure includes conducting a photoelectrochemical CO2 reduction reaction. The method further includes synthesizing the plurality of nanoclusters via a solution-based chemical reduction process. Depositing the plurality of nanoclusters includes implementing a drop-cast assembly process. Forming the array of conductive projections includes growing an array of nanowires on the semiconductor substrate.

BRIEF DESCRIPTION OF THE DRAWING FIGURES

For a more complete understanding of the disclosure, reference should be made to the following detailed description and accompanying drawing figures, in which like reference numerals identify like elements in the figures.

FIG. 1 is a schematic illustration of in-situ construction of a strongly coupled Cu cluster/GaN nanowire/Si photocathode during a PEC CO₂RR process in accordance with one example, in which low-coordinated copper clusters were initially loaded on n-type GaN nanowires grown on a n⁺-p Si wafer. During the PEC process, a Ga—N—O on GaN surface was in situ formed. The Ga—N—O interacted with the copper clusters, inducing an interfacial positive copper state of copper cluster at the sub-nanometer interface region. The interface maintains a stable oxidized copper species at the Cu cluster-GaN interface for efficient C—C coupling during CO₂RR. At the same time, the in situ formed Ga—N—O interface promotes the H spillover to CO hydrogenation process, establishing a CHO involved C—C coupling pathway toward ethylene formation.

FIG. 2 depicts the structural characterization of a strongly coupled Cu/GaN nanowire interface of a photocathode in accordance with one example, including (a) a SEM image with a STEM inset of a single GaN nanowire, (b) a HAADF-STEM image, (c) a high resolution HAADF-STEM image, (d) STEM-EELS SI element maps, (e) oxygen and Cu distribution plots encircled in FIG. 2, part (d), (f, g) graphical plots of the EELS profile spectra for Cu-L (f) and the enlarged spectra around the interface (g), with an inset in FIG. 2, part (f) showing the region and the direction for EELS profile spectra extraction, and (h) a graphical plot of XPS spectra of Cu 2p with different Ar⁺ etching times.

FIG. 3 depicts graphical plots of photoelectrochemical CO₂RR performance, including (a) LSV curves of a strongly coupled Cu cluster/GaN nanowire/Si photocathode example under chopped light (1 sun illumination, AM 1.5G, 100 mW cm⁻²), (b) photocurrent-potential LSV curves of three different Si photocathodes under simulated 1 sun illumination, (c) the highest FEs of gas products generated by different photocathodes, (d) FEs of the products at different applied potentials for an example Cu cluster/GaN nanowire/Si photocathode, in which error bars represent the standard deviation of three independent measurements, (f) a performance comparison of an example Cu cluster/GaN nanowire/Si photocathode and other photocathodes for reduction of CO₂ to ethylene, and (g) chronoamperometry and FE results of an example Cu cluster/GaN nanowire/Si photocathode at a potential of −0.74 V vs RHE.

FIG. 4 depicts graphical and other illustrations of an experimental mechanism investigation, including (a) kinetic isotopic effect (KIE) measurements of hydrogen/deuterium (H/D) for C₂H₄ in PEC CO₂RR to C₂H₄ performance of an example Cu cluster/GaN nanowire/Si photocathode and of a Cu/Si photocathode at −0.74 V vs RHE in CO₂ saturated 0.1 M KHCO₃ electrolyte, (b) Nyquist plots measured at various potentials on an example Cu cluster/GaN nanowire/Si photocathode in CO₂ saturated 0.1 M KHCO₃ electrolyte, with an inset depicting the equivalent circuit for the EIS simulation, and with R_s, R₁ and T representing the solution resistance, the catalytic charge transfer resistance and double layer capacitance respectively, and with C_φ and R₂ reflecting the pseudo-capacitance and resistance of hydrogen adsorption, respectively, (c) the EIS-derived Tafel plots of the *H adsorption resistance R₂ for the Cu cluster/GaN nanowire/Si photocathode and the Cu/Si photocathode, (d) the model of in situ ATR-FTIR installation, (e) in situ ATR-FTIR spectra in high wavenumber region measured on an example Cu cluster/GaN nanowire/Si photocathode during PEC CO₂RR, (f) the time-dependent O—H peak position, and (g, h) in situ ATR-FTIR 2D spectra in 1000-2000 nm⁻¹ measured on (g) a Cu cluster/GaN nanowire/Si photocathode example and (h) a Cu/Si photocathode during PEC CO₂RR.

FIG. 5 schematically and graphically depicts the theoretical mechanism for an optimized Cu-GaN interface for CO₂RR to C₂H₄ conversion, including (a, b) simulated structural models of (a) Cu/GaN and (b) Cu/GaNO/GaN, (c) PDOS plots of Cu 3d orbitals of the Cu/GaNO/GaN and Cu/GaN structures, (d) the free energy diagram for the CO₂RR to C₂H₄ on the Cu/GaNO/GaN and Cu/GaN structures, (e) the energy barriers in different steps from the free energy diagram, and (f) the reaction mechanism of the CO₂RR to C₂H₄ on the Cu/GaNO/GaN structure.

FIG. 6 is a schematic view and block diagram of an electrochemical system having a working electrode with a nitride semiconductor-metal interface for catalytic conversion of carbon dioxide (CO₂) in accordance with one example.

FIG. 7 is a flow diagram of a method of fabricating a device (e.g., a photocathode) for catalytic conversion of carbon dioxide (CO₂) via a nitride semiconductor-metal interface in accordance with one example.

The embodiments of the disclosed devices, systems, and methods may assume various forms. Specific embodiments are illustrated in the drawing and hereafter described with the understanding that the disclosure is intended to be illustrative. The disclosure is not intended to limit the invention to the specific embodiments described and illustrated herein.

DETAILED DESCRIPTION OF THE DISCLOSURE

Photoelectrodes and other photocatalytic and catalytic devices having metal clusters coupled to an array of nitride semiconductor projections via oxidized interfaces are described. The oxidized interfaces include both an oxide species of the metal and an oxynitride species based on the nitride semiconductor. In some cases, the oxidized interfaces and thus, the photocatalytic devices, are configured for catalytic conversion of carbon dioxide (CO₂) into multi-carbon products, such as ethylene (C₂H₄). Methods for fabricating such devices are also described.

The disclosed devices and methods address multiple, fundamental challenges facing photoelectrochemical CO2 reduction reaction (PEC CO₂RR) systems for the synthesis of multi-carbon products, including the creation of efficient C—C coupling sites and the stabilization of the sites under long-time photoelectrochemical reduction conditions. As described herein, the interface structure between a CO₂RR catalyst and a supporting photocathode provides a useful approach for controlling the electronic structure, selectivity, and stability of CO₂RR-catalysts. The oxidized interfaces of the disclosed photocatalytic devices provide efficient C—C coupling sites for producing ethylene and other multi-carbon compounds. The C—C coupling sites are also stable under extended photo/electrochemical reduction conditions.

Examples are described below that include a low-coordinated Cu cluster catalyst that is strongly interfacially coupled with a GaN nanowire/Si support structure (referred to herein as “s-Cu/GaN NW/Si”). The combination functions as an efficient photocathode for ethylene synthesis from CO₂RR. During the PEC process, the interfacial coupling between the Cu clusters and the GaN nanowires is formed in situ. As described herein, an in situ, self-optimized, sub-nanometer Ga—N—O interface is formed between each Cu cluster and a respective GaN nanowire. The interface was found to induce and stabilize the interfacial oxidized Cu species of the Cu cluster, which was confirmed to boost C—C coupling for ethylene production (see, e.g., FIG. 1). A stable cluster-scale dispersed Cu active structure with low-coordinated sites is also strongly anchored on the GaN nanowire against aggregation due to the strong interfacial interaction. Thus, the low-coordinated Cu structure with partial oxidized copper species at the Cu cluster-GaN interface is robust for efficient and stable C—C coupling in CO₂RR. Moreover, the self-optimized Ga—N—O interface also acts as a hydrogen feeder, providing the active H intermediate for the CO hydrogenation process, thereby guiding the CHO involved C—C coupling pathway toward ethylene formation. Examples with this structure exhibited record high ethylene FE of about 61% and partial current density of 14.2 mA cm⁻² in a PEC system, along with stable operation up to about 116 hours. The exhibited activity and stability is an approximately 6-fold improvement compared to the best reported values to date. The disclosed devices are thus useful for promoting C—C coupling in photoelectrochemical devices for artificial photosynthesis and other reactions and applications.

The copper nanoclusters are supported by an array of conductive projections (e.g., nanowires). One-dimensional (1-D) nanostructured metal nitrides, such as gallium nitride (GaN) nanowires (GaN nanowires), are useful in solar fuels production and capable of being grown via molecular beam epitaxy (MBE) defect-free on planar silicon. The heterostructure of the GaN nanowires presents a large surface-to-volume ratio, which is beneficial for sunlight harvesting and catalyst loading with a dramatically reduced amount, but high-density, of catalytic centers. Furthermore, the defect-free structure and high charge carrier mobility of GaN nanowires lead to charge carrier extraction from the silicon substrate. The electronic properties of gallium nitride are useful for activating the stable carbon dioxide molecule, thereby presenting a useful platform for supporting nanoclusters to construct an effective nanoarchitecture for solar-driven CO₂ conversion.

Although described herein in connection with electrodes having GaN-based nanowire arrays for PEC CO₂ reduction, the disclosed electrodes are not limited to PEC reduction or GaN-based or other nanowires. A wide variety of types of photocatalytic reactions and chemical cells may benefit from use of the conductive projection (e.g., nanowire)-nanocluster interface, including, for instance, electrochemical cells and thermochemical cells. Moreover, the nature, construction, configuration, characteristics, shape, and other aspects of the conductive projections (e.g., nanowires) may vary. The disclosed devices, systems, and methods may also be directed to CO₂ reduction products other than or in addition to ethylene, such as CO, CH₃OH, CH₄, C₂H₅OH, C₃H₇OH. The disclosed devices may also be used to catalyze other reactions, including, for instance, hydrogen evolution and nitrate reduction.

Although described herein in connection with Cu nanoclusters, the disclosed devices, systems, and methods may use other metals. For example, other d-block metals, such as gold (Au), silver (Ag), and zinc (Zn), may be used. D-block metals include the elements in groups 3-12 of the periodic table. The metal nanoclusters may alternatively or additionally include still other d-block metals. For instance, platinum (Pt) may be used for hydrogen evolution, and Co may be used for nitrate reduction.

Although described in connection with PEC-based in situ formation of the Cu cluster-GaN nanowire oxidized interfaces, the disclosed methods may use other in situ procedures, including bias-free procedures. For instance, high-concentration light (e.g., greater than 20 suns) may be used to form the oxidized interfaces without involving a bias input.

Details regarding the synthesis of Cu clusters and fabrication of a number of example devices are provided below in connection with FIGS. 1 and 2.

An example GaN nanowire (NW) array 100 was grown on planar n⁺-p silicon wafer 102 by molecular beam epitaxy. Each nanowire 104 of the array 100 had a length of about 300 nm and diameter of about 50 nm. The nanowires 104 were doped n-type with silicon. Low-coordinated Cu cluster catalysts were synthesized by a rapid chemical reduction method using sodium borohydride in an ethanol/water solution. The clusters mainly exhibited low-coordinated facets, such as (101), (200) and (hkl), and no obvious high-coordinated surface (111) was detected. Then Cu nanoclusters 106 were assembled onto the GaN nanowires 104 using a drop-cast-directed assembly process. After photoelectrochemical activation (e.g., of the interfaces between the nanoclusters 106 and the nanowires 104) during PEC CO₂RR as described herein, a stable and efficient Cu cluster/GaN nanowire/Si photocathode was obtained. As schematically shown in FIG. 1 and described herein, in the example photocathode, each nanocluster 106 is coupled to a respective one of the nanowires 104 via an oxidized interface 108. The oxidized interface 108 includes an oxide species of the metal and an oxynitride species based on the nitride semiconductor.

Scanning electron microscope (SEM) images show the well-kept nanowire array structure with the cluster deposition on the photocathode (FIG. 2, part a). High-angle annular dark-field images performed in scanning transmission electron microscopy mode (HAADF-STEM) show that the Cu clusters, with an average size around 1.7 nm (accordingly referred to herein as “nanoclusters”), are uniformly deposited on the surface of the GaN nanowires (FIG. 2, part b). Due to the ultrasmall size and low-coordinated structure of the copper clusters, the clusters may have a distorted lattice and are dispersed on the GaN nanowire surface (FIG. 2, part c). The GaN nanowires with the deposited clusters maintain high crystallinity after the copper cluster deposition, as shown in FIG. 2, part c.

FIG. 2, part c, also shows that there is a distinct sub-nanometer interface region between the Cu cluster and the GaN surface. This interface region is derived from oxygen insertion at the Cu cluster-GaN interface during the reaction (e.g., activation reaction), which induces the structure distortion, charge redistribution and altered chemical state of the Cu clusters, as explained herein. Based on electron energy loss spectroscopy (EELS) spectrum imaging performed in STEM mode (STEM-EELS), there is an oxygen layer between the Cu cluster and the GaN nanowire (FIG. 2, part d). In contrast, the original GaN nanowires show a very weak oxygen signal on the entire nanowire in energy-dispersive X-ray spectroscopy images. This analysis confirms that the partially oxidized interface region between the copper clusters and the GaN surface is in situ constructed, in this case, during the PEC process.

Furthermore, the interpenetrated copper and oxygen between the copper cluster and oxygen interlayer can also be observed in FIG. 2, part e, on the resulting photocathode, i.e., the s-Cu/GaN nanowire/Si photocathode. Although previous reports have discussed the easy oxidation of the surfaces of unsupported copper catalysts, here the oxygen species are not detected on the outer surface of the copper cluster, which can be attributed to the metal-support interactions. The above results indicate that the interfacial, partially oxidized Cu species interacted with the in situ formed Ga—N—O layer.

To further uncover the chemical state of interfacial Cu sites, EELS profile spectra were extracted from the Cu—GaN interface to the outside surface of the Cu cluster on a layer by layer basis (FIG. 2, part f). The oxidation state of the Cu species can be identified from the shapes of the Cu L_2,3 edges, by which the metallic Cu presents an upshift spectral position with a flat and broad L₃, while partially oxidized Cu is characterized by a sharp L₃ with a downshift. Accordingly, from the sub-nanometer region to the Cu cluster, the chemical state of Cu is changed from partially oxidized Cu (Cu^δ+) at the interface to the metallic Cu at the cluster (FIG. 2, parts f and g). The Ga and N signals of the GaN support are gradually decreased. This result indicates that the interfacial partially oxidized Cu species is induced by the in situ optimized Cu cluster-GaN interface.

X-ray photoelectron spectroscopy (XPS) spectra were then performed to gain an insight into the structural characteristics of the Cu clusters and the GaN nanowires in a s-Cu/GaN nanowire/Si photocathode. Cu 2p XPS spectra in the s-Cu/GaN nanowire/Si photocathode indicate that the valence of Cu is mainly zero (FIG. 2, part h), accompanied with a small number of Cu^δ+ species. The XPS spectra together with the Ar⁺ etching in different times were further determined, which indicates that Cu^δ+ species is still existent during the etching process, implying that the Cu^δ+ species does not originate from the surface oxidized copper. In addition, the original copper clusters were measured by XPS and HAADF-STEM, which verifies that there is no distinguished oxidized copper species and no obvious oxidized layer on the copper cluster outer surface. The above results are consistent with the interfacial oxidized Cu species as revealed from the above EELS analyses. For the O 1s XPS spectra, the Ga—N—O species at about 531.7 eV and O-Ga/O-Cu species at 530.3 eV are detected, further implying the formed oxynitride species on GaN surface and interfacial Cu—O species. Similar results can also be derived from Ga 3d XPS spectra. The s-Cu/GaN nanowire/Si example displays an additional Ga—N—O peak at 20.9 eV between the Ga—N and Ga—O peaks. The Raman spectra further show that the extra Cu—O and Ga—O species were detected on the s-Cu/GaN nanowire/Si example, consistent with the XPS results.

The photoelectrochemical CO₂ reduction performance of the example devices is now described in connection with FIGS. 3 and 4.

Photoelectrochemical CO₂RR was conducted with the example devices under 100 mW/cm² of AM 1.5 simulated sunlight, using a H-type cell with a three electrode system in a CO₂-saturated 0.1 M KHCO₃ electrolyte. The LSV curve of the example s-Cu/GaN nanowire/Si photocathode under chopped light indicates the current under illumination derives from photogenerated electrons (FIG. 3, part a). A Cu/Si photocathode for reference was fabricated using a similar process to that used for the fabrication of the example s-Cu/GaN nanowire/Si device, except for replacing the GaN nanowire/Si support structure with an n⁺-p Si wafer. It is observed that the example s-Cu/GaN nanowire/Si device achieved enhanced total current density (j) with the onset potential greater than 0 V (vs RHE), as compared with the reference photoelectrodes of GaN nanowires/Si and Cu/Si under illumination (FIG. 3, part b), indicating its higher intrinsic CO₂RR activity. The Cu catalyst loading (e.g., 20 μg cm⁻²) is lower than typical reported electrocatalytic systems to achieve higher or similar current densities under similar conditions.

The product selectivity of the PEC CO₂RR on different photocathodes was further analyzed and compared. As most of the products from CO₂RR were observed in the gas phase, the focus is on the CO₂RR performance with gas-phase products. A maximum ethylene FE of about 61% was achieved for the example s-Cu/GAN nanowire/Si photocathode at −0.74 V (vs RHE), while only CO, CH₄ and a small amount of multi-carbon products were detected by Cu/Si photocathode (FIG. 3, part c), indicating that the GaN nanowires serve a role for producing multi-carbon products during CO₂RR. The electrochemical CO₂RR performance of copper clusters loaded on carbon paper was also performed. It showed that the overpotential is notably increased and C₂H₄ selectivity is largely reduced in electrochemical CO₂RR compared with the PEC condition, further confirming the coupled Cu-cluster/GaN photocathode is useful for low-energy and highly efficient CO₂ to C₂H₄ conversion.

Different electrolyte conditions were also analyzed to detect the local pH effect. The analysis showed that an appropriately increased local pH may promote the C₂H₄ production in the disclosed systems. In addition, the Nafion loading effect was explored, and the results show that Nafion addition leads to little effect for enhancing C₂H₄ production. Therefore, the enhanced selectivity for C₂H₄ mainly originates from the intrinsic activity of Cu/GaN interfacial active sites.

The potential-dependent activity and selectivity toward different products for the disclosed devices and systems are shown in FIG. 3, parts d and e. With an increased overpotential, the selectivity of C₂H₄ initially increased and then slightly decreased, while the CH₄ selectivity initially decreased and then increased (FIG. 3, part d). These results indicate that the CO₂RR to C₂H₄ in the disclosed systems is highly correlated to the competing reaction of CH₄ formation, which can be effectively controlled by the applied potential for steering the shared key intermediates (such as *CHO) for the two reactions. For instance, at a potential of −0.74 V vs RHE, the highest C₂H₄ FE value of 61±3% was obtained. The applied potential for C₂H₄ may vary in other cases. The partial current density of C₂H₄ was higher compared with other gas products of CH₄, CO, and H₂, achieving the highest value of 14.2 mA cm⁻² at −0.74 V vs RHE. No hydrocarbon product was detected from the Ar-purged 0.1 M KHCO₃ electrolyte, and ¹³CO₂ isotope labeling experiments further confirmed that the CO₂RR products originated from the CO₂ feed gas. The C₂H₄ selectivity for CO₂RR at different Cu cluster loadings shows that the optimal loading of 20 μg cm⁻² is sufficient for favored C₂H₄ production. Notably, the selectivity and activity of the disclosed s-Cu/GaN nanowire/Si photocathode are the highest among the recently reported PEC CO₂RR systems (FIG. 3, part f), around 6-fold that of the best result reported for partial current density of ethylene.

The durability of an example s-Cu/GAN NWs/Si photocathode was also examined at −0.74 V (vs RHE) for about 116 h. The example showed a stable ethylene FE of about 60% with a current density of about 20 mA cm⁻² over the span of the stability test (FIG. 3, part g). Additionally, after the long-term stability tests, the SEM image and EELS element mapping confirm that the nanowire array morphology and local Cu cluster structure on the GaN surface was well retained. As compared with previously reported PEC systems, the example s-Cu/GaN NWs/Si photocathode displayed a much higher stability (about 6-fold the best one reported) and simultaneously maintained the highest FE and current density of ethylene under comparable conditions (FIG. 3, part f).

To elucidate the underlying mechanism(s) of an example s-Cu/GaN NWs/Si photocathode for PEC CO₂RR to ethylene, in situ photoelectrochemical and spectroscopy experiments were performed to reveal the intrinsic functions of the Cu cluster-GaN interface during the reaction. Given that water is the primary proton source of CO₂ reduction, the effect of H₂O dissociation on CO₂RR to C₂ products was considered by the kinetic isotopic effect (KIE) of hydrogen/deuterium (H/D), which are defined as the ratios of FE_C2H4 in H₂O and D₂O on the s-Cu/GaN NWs/Si and Cu/Si photocathodes (FIG. 4, part a). When the H/D KIE value exceeds 1.5, the kinetics of H₂O dissociation can significantly impact the related reaction rate. The Cu/Si photocathode exhibited a significant decrease in the FE value of C₂H₄ in D₂O compared to that in H₂O, resulting in a KIE value of 2.14. In contrast, the example s-Cu/GaN NWs/Si photocathode exhibited a minimal change in KIE value (1.03), suggesting that the hydrogen supply to CO₂ deep reduction can be effectively mitigated by the hydrogen feeding function of Ga—N—O surface. In situ electrochemical impedance spectroscopy (EIS) was further measured to investigate the *H adsorption behavior and *H coverage on the catalyst during the CO₂RR. The Nyquist plots are fitted by a double-parallel equivalent circuit (FIG. 4, part b). The first parallel component (R₁) reflects the catalytic charge transfer resistance, implying the faster charge transfer of s-Cu/GaN NWs/Si photocathode for higher photocurrents. The second parallel component (R₂) represents the *H adsorption behavior on the catalyst, and the *H adsorption kinetics can be quantified by plotting the potential-dependent log (R₂). The slope of s-Cu/GaN NWs/Si photocathode is about half of that on the Cu/Si photocathode, indicating that the increased *H coverage for CO₂ activation and hydrogenation during the reaction by the hydrogen feeding function of Ga—N—O surface (FIG. 4, part c).

To further confirm these findings, in situ attenuated total reflection fourier transform infrared (ATR-FTIR) spectra under PEC CO₂RR conditions were examined to detect the adsorbed intermediates and identify the reaction pathway (FIG. 4, part d). The enhanced water dissociation on the gradually optimized Cu cluster-GaN nanowires can be observed by the in situ ATR-SEIRAS spectra. The peak position corresponding to the O—H stretching vibration on the s-Cu/GaN NWs/Si photocathode shows a much lower wavenumber than that on Cu/Si photocathode (FIG. 4, parts e and f), with a downshift during the reaction, further indicating the optimized water dissociation for hydrogen donation on the s-Cu/GaN NWs/Si photocathode during the reaction. This hydrogen feeding effect of the Ga—N—O surface was further confirmed by the in situ optimized HER performance with the prolonged HER chronoamperometric testing, during which the Ga—N—O surface was in situ formed. The above results together confirm the role of the Ga—N—O surface for acting as the hydrogen feeder to provide hydrogen for CO₂ deep reduction.

The function of the in situ formed interfacial oxidized Cu species on the Cu—GaN interface can be uncovered from analysis of the PEC process. With increased reaction time, the FE_C2H4/FE_CH4 ratio gradually increased, during which the Cu—O species also gradually increased. This result indicates that the improved Cu cluster-GaN interface was in situ formed during the reaction process, with interfacial oxidized Cu species being formed for C₂ products production. In addition, the ATR-FTIR spectra of the s-Cu/GaN NWs/Si photocathode displayed the absorption peaks of *CHO (or *COH) at 1471 cm⁻¹ (FIG. 4, part g), indicating increased *H migration to the adjacent Cu clusters on the s-Cu/GaN NWs/Si photocathode, where the adsorbed *CO is then hydrogenated to form *CHO. The peak at 1218 cm⁻¹ can be well assigned to the *CHOCO species absorbed on s-Cu/GaN NWs/Si photocathode. These intermediates are significantly increased with the reaction time, corresponding to the optimized Cu cluster-GaN nanowire interface and interfacial Cu^δ+ sites formation, which act as the C—C coupling sites of *CO and *CHO to generate *CHOCO for C₂H₄ formation. The weak CO absorption peak at 2050 cm⁻¹ is indicative of the rapid C—C coupling and the pathway of *CO—*CHO coupling for C₂H₄ formation. In contrast, the C₂ intermediates with the peaks of *CHO and the *CHOCO signal were not obviously observed for the Cu/Si photocathode (FIG. 4, part h), and only very weak signals for C₁ intermediate of *COOH were detected. These results are consistent with the very low efficiency for C₂ products formation on the Cu/Si photocathode.

Based on the results above, it is evident that the self-optimized interface with partially oxidized Cu species of the example s-Cu/GaN NWs/Si photocathode stabilizes the C₂ intermediates (*CHO and *CHOCO) during the PEC process, which cannot be achieved by utilizing a Cu/Si catalyst alone.

Density functional theory (DFT) calculations were further performed to uncover the intrinsic mechanism of this process. A Cu/GaN model was constructed by building a Cu cluster (about 1 nm) having 31 Cu atoms on GaN surface (FIG. 5, part a). An example s-Cu/GaN NWs/Si photocathode was constructed by partial oxygen substitution (⅓) of the surface N atoms of Cu/GaN to form a stable Ga—N—O layer (denoted as Cu/GaNO/GaN, FIG. 5, part b). A projected density of states (PDOS) of Cu 3d was performed to investigate the electronic properties of optimized Cu species at different interfaces. When the Ga—N—O layer is formed between the Cu cluster and GaN, the d-band center and band edge of the interfacial Cu atom presents a downward shift from the Fermi level compared with the Cu/GaN model (FIG. 5, part c). This indicates that Cu atoms adjacent to the interface of Ga—N—O are more positively charged, consistent with the characterization for the interface Cu species (FIG. 2), which improves both stability and C₂ product selectivity. The results reveal that the strongly coupled Cu cluster-GaN interface in the example s-Cu/GaN NWs/Si photocathode induces a unique charge redistribution and altered chemical state of Cu cluster as well as the geometrical distortions (highlighted in the dashed boxes of FIG. 5, parts a and b).

The free energy diagram of the reaction pathway on the two models is shown in FIG. 5, part d. The simplified interface models without adding more variables (such as electrolyte ions), adopted in previous theoretical studies, were employed to reduce the complexity in assessing the interfacial active sites for CO₂RR, given that it has remained challenging to fully address all reaction conditions and sites under actual operational conditions. CO₂ is initially activated to form *CO and then hydrogenated into *CHO, which will be further coupled with *CO to generate the key *CHOCO intermediate, and finally undergoes a series of energy downhill processes for C₂H₄ production. It is evident that the energy barrier of the key elementary steps of CO₂ adsorption and activation to *COOH is lowered on Cu/GaNO/GaN (FIG. 5, part e). Thus, the CO₂ reduction activity can be increased by steering the selectivity from the hydrogen evolution reaction to the CO₂ reduction pathway. The energy barrier of *CO to *CHO, a conjoint step for both CH₄ and C₂H₄ formation, is similar in the two systems. This explains why CH₄ is the major competing product in the potential-altered process (FIG. 3, part d) and activation process (FIG. 4, part h). Thus, the C—C coupling step (*CO+*CHO to *CHOCO) determines CH₄ or C₂H₄ as the major products. The rate-determining step for C₂H₄ products formation on Cu/GaN is the C—C coupling step, while the interfacial Ga—N—O layer significantly reduces the energy barrier from 1.39 eV on Cu/GaN to 0.25 eV for the Cu/GaNO/GaN structure (FIG. 5, part e). The competitive CO₂ to CH₄ pathway was also calculated for the Cu/GaNO/GaN structure, indicating that the energy barrier at *CH₂OH formation in the CH₄ production pathway (0.68 eV) is much higher than that of the *CO—*CHO coupling (0.25 eV). It validates that the more feasible C₂H₄ formation pathway of the Cu/GaNO/GaN structure suppresses the competitive CH₄ formation. In addition, this C—C coupling step on the Cu cluster surface sites of the Cu/GaNO/GaN structure was also calculated. It shows a much higher energy barrier of 1.13 eV than that on the interfacial Cu site (0.25 eV), indicating that the interfacial Cu site is more active for C—C coupling than the surface Cu sites in CO₂RR. Thus, the DFT calculations demonstrate that the interfacial oxidized Cu species induced by the in situ generated Ga—N—O layer stabilizes the C₂ intermediate of *CHOCO, and thus promotes the C—C coupling step for C₂H₄ production, which is consistent with the experimental results (FIG. 3, part d; FIG. 4, part g).

The possible *CHO—*CHO coupling pathway for the Cu/GaNO/GaN structure was also calculated, which shows a higher barrier of 0.41 eV. Thus, the *CO—*CHO coupling pathway is more favorable, aligning with the ATR-FTIR observations (FIG. 4, part g). Then, the *CO to *COH pathway for C₂H₄ formation was calculated. Compared with the very high energy barrier for *CO to *COH for the Cu/GaN model, the Cu/GaNO/GaN structure significantly lowers the energy barrier of *CO hydrogenation to *COH for subsequent *CO—*COH coupling (FIG. 5, part e), which also promotes the C₂H₄ production. In addition, the *CO hydrogenation to *CHO or *COH can be enhanced by the improved hydrogen donation from the GaNO as illustrated in FIG. 4, parts a-f, which is also confirmed by the lower water dissociation energy of the Cu/GaNO/GaN structure due to the improved interfacial sites for breaking H—O bond. This effect helps steer the reaction to the facile *CHO involved C—C coupling pathway. These results confirm that the strongly coupled Cu cluster-GaN interface with the interfacial partially oxidized Cu species promotes CO hydrogenation and stabilizes C₂ intermediates for C—C coupling, both of which support stable and efficient C₂H₄ formation (FIG. 5, part f).

In addition to providing interfacial active sites, the GaN nanowires enhance both optical and electronic properties of the disclosed devices during PEC operation. The one-dimensional structure of the GaN nanowire arrays allows for spatial separation of catalysis from sunlight absorption, maximizing the synergistic effect between the Cu clusters and the GaN nanowires for efficient C₂H₄ synthesis. The corresponding energy band diagram illustrates that the GaN nanowire and the n⁺-p silicon junction possess conduction band edges that are nearly aligned. Moreover, the GaN nanowires grown by MBE exhibit minimal defects and high electron mobility. Thus, the photogenerated electrons can be easily extracted from the n⁺-p silicon junction to the GaN nanowire surfaces under illumination. The electrons can then quickly migrate to the deposited Cu cluster catalyst for CO₂RR. This is consistent with the incident photon to current efficiency (IPCE) measurement and electrochemical impedance spectroscopy findings. Additionally, the IPCE measurement reveals that the Cu—N—O interface of the Cu/GaNO/GaN structure can further improve the photoelectric properties of Cu/GaN alone, such that the photogenerated electrons can be more efficiently extracted to the copper clusters for the CO₂RR.

Described above are examples of catalytic devices having a strongly coupled sub-nanometer interface between a GaN nanowire/Si photocathode and low-coordinated Cu clusters. The interface efficiently stabilizes the partial oxidized Cu species on the Cu clusters for breaking the bottleneck of C—C coupling pathway and CO₂ activation during PEC CO₂ reduction. An ethylene Faradic efficiency of about 61% and stability of about 116 hours in photoelectrochemical systems were achieved. Compared with the conventional strategies of interface engineering for photocathode design, the in situ self-optimized interface engineering of the disclosed devices is more active and selective for C₂ hydrocarbon production, which is useful for greatly boosting the efficiency and stability of CO₂RR in electrolyzer, artificial photosynthesis, and other systems.

A number of examples of the disclosed devices, systems, and methods are now described in connection with the schematic diagram of FIG. 6 and the flow diagram of FIG. 7.

FIG. 6 depicts a system 600 for reduction of CO₂ into multi-carbon products such as ethylene in accordance with one example. The system 600 may also be configured for alternative or additional reactions, including, for instance, the evolution of H₂. The system 600 may be configured as an electrochemical system. In this example, the electrochemical system 600 is a photoelectrochemical (PEC) system in which solar or other radiation is used to facilitate the CO₂ reduction. The manner in which the PEC system 600 is illuminated may vary. In thermochemical examples, the source of radiation may be replaced by a heat source.

The electrochemical system 600 includes one or more electrochemical cells 602. A single electrochemical cell 602 is shown for ease in illustration and description. The electrochemical cell 602 and other components of the electrochemical system 600 are depicted schematically in FIG. 6 also for ease in illustration. The cell 602 contains an electrolyte solution 604 to which a source 606 of CO₂ is applied. Potassium bicarbonate KHCO₃ may be used as an electrolyte. The electrolyte may thus be free of an organic solvent. Additional or alternative electrolytes may be used. Further details regarding an example of the electrochemical system 600 are provided hereinabove.

The electrochemical cell 602 includes a working electrode 608, a counter electrode 610, and a reference electrode 612, each of which is immersed in the electrolyte 604. The counter electrode 610 may be or include a metal wire or mesh, such as a platinum wire or mesh. The reference electrode 612 may be configured as a reversible hydrogen electrode (RHE). The configuration of the counter and reference electrodes 610, 612 may vary. For example, the counter electrode 610 may be configured as, or otherwise include, a photoanode at which water oxidation (2H₂OO2+4e⁻+4H⁺) occurs.

Both reduction of CO₂ and evolution of H₂ may occur at the working electrode 612 as follows:

To that end, electrons flow from the counter electrode 610 through a circuit path external to the electrochemical cell 602 to reach the working electrode 608. The working and counter electrodes 608, 610 may thus be considered a cathode and an anode, respectively. The competition between reduction of CO₂ and evolution of H₂ may be managed or controlled (e.g., to favor CO₂ reduction) via the composition or other characteristics of the working electrode 608 and/or the applied voltage, as described herein.

In the example of FIG. 6, the working and counter electrodes may be separated from one another by a membrane 614, e.g., a proton-exchange membrane. In some cases, the membrane 614 is configured as, or otherwise includes, a Nafion membrane. The construction, composition, configuration and other characteristics of the membrane 614 may vary.

In the example of FIG. 6, the circuit path includes a voltage source 616 of the electrochemical system 600. The voltage source 616 is configured to apply a bias voltage between the working and counter electrodes 608, 610. The bias voltage may be used to promote the production of ethylene at the working electrode, as described herein. The bias voltage may alternatively or additionally be used to establish other production characteristics, such as a ratio of CO₂ reduction to hydrogen (H₂) evolution. The circuit path may include additional or alternative components. For example, the circuit path may include a potentiometer in some cases.

In the examples described above, an H-type cell separated by a Nafion membrane using an electrochemical station with a three-electrode system was used. Ag/AgCl (3.5 M KCl) and Pt gauze were used as the reference electrode and the counter electrode, respectively. All potentials were converted into the RHE scale after iR correction. Both compartments were filled with 25 mL of 0.1 M KHCO₃ solution, in which the catholyte was saturated by CO₂ for 30 min and continuously bubbled with CO₂ (20 sccm) throughout the measurement.

In some cases, the working electrode 608 is configured as a photocathode. Light 618, such as solar radiation, may be incident upon the working electrode 608 as shown. The electrochemical cell 602 may thus be considered and configured as a photoelectrochemical cell. In such cases, illumination of the working electrode 608 may cause charge carriers to be generated in the working electrode 608. Electrons that reach the surface of the working electrode 608 may then be used in the CO₂ reduction and/or the H₂ evolution. The photogenerated electrons augment the electrons provided via the current path. The photogenerated holes may move to the counter electrode for the water oxidation. A number of examples of, and further details regarding, photocathodes are provided hereinabove in connection with, for instance, FIGS. 1-5. In those examples, the light source had an AM 1.5G filter and a light intensity of 100 mW cm⁻².

The working electrode 608 includes a substrate 620. The substrate 620 of the working electrode 608 may constitute a part of an architecture, or a support structure, of the working electrode 608. The substrate 620 may be uniform or composite. For example, the substrate 620 may include any number of layers or other components. The substrate 620 thus may or may not be monolithic. The shape of the substrate 620 may also vary. For instance, the substrate 620 may or may not be planar or flat.

The substrate 620 of the working electrode 608 may be active (functional) and/or passive (e.g., structural). In the latter case, the substrate 620 may be configured and act solely as a support structure for a catalyst arrangement formed along an exterior surface of the working electrode 608, as described below. Alternatively or additionally, the substrate 620 may be composed of, or otherwise include, a material suitable for the growth or other deposition of the catalyst arrangement of the working electrode 608.

The substrate 620 may include a light absorbing material. The light absorbing material is configured to generate charge carriers upon solar or other illumination. The light absorbing material has a bandgap such that incident light generates charge carriers (electron-hole pairs) within the substrate. Some or all of the substrate 620 may be configured for photogeneration of electron-hole pairs. To that end, the substrate 620 may be composed of, or otherwise include, a semiconductor material. In some cases, the substrate 620 is composed of, or otherwise includes, silicon. For instance, the substrate 620 may be provided as a silicon wafer. The silicon may be doped. In some cases, the substrate 620 is heavily n-type doped, and moderately or lightly p-type doped, to form a junction. The doping arrangement may vary. For example, one or more components of the substrate 620 may be non-doped (intrinsic), or effectively non-doped. The substrate 620 may include alternative or additional layers, including, for instance, support or other structural layers. In other cases, the substrate 620 is not light absorbing. In these and other cases, one or more other components of the photocathode (e.g., nanowires) may be composed of, or otherwise include, a semiconductor material configured to act as a light absorber. Thus, in photoelectrochemical cases, the semiconductor material of the substrate and/or other components supported by the substrate may be configured to generate charge carriers upon absorption of solar (or other) radiation, such that the chemical cell is configured as a photoelectrochemical system.

The substrate 620 of the working electrode 608 establishes a surface at which a catalyst arrangement of the electrode 608 is provided for catalytic conversion of carbon dioxide (CO₂) in the chemical cell 602 into, e.g., ethylene. The catalyst arrangement includes a conductive projection (e.g., nanowire)-nanocluster architecture as described herein.

The electrode 608 includes an array of nanowires 622 and/or other conductive projections supported by the substrate 620. Each nanowire 622 extends outward from the surface of the substrate 620. The nanowires 622 may thus be oriented in parallel with one another. Each nanowire 622 is composed of, or otherwise includes, a nitride semiconductor. In some cases, the nitride semiconductor is gallium nitride (GaN). Additional or alternative nitride semiconductor materials may be used, including, for instance, indium nitride, indium gallium nitride, aluminum nitride, boron nitride, aluminum oxide, silicon, and/or their alloys.

The nanowires 622 may facilitate the conversion in one or more ways. For instance, each nanowire 622 may be configured to extract the charge carriers (e.g., electrons) generated in the substrate 620. The nanowires 622 may be doped (e.g., n-type doped) to facilitate the extraction and movement of the charge carriers. The extraction brings the electrons to external sites along the nanowires 622 for use in the CO₂ reduction. The composition of the nanowires 622 may also form an interface well-suited for reduction of CO₂, as explained herein.

Each nanowire 622 may be or include a columnar, post-shaped, or other elongated structure that extends outward (e.g., upward) from the plane of the substrate 620. The nanowires 622 may be grown or formed as described in U.S. Pat. No. 8,563,395, the entire disclosure of which is hereby incorporated by reference. The dimensions, size, shape, composition, and other characteristics of the nanowires 622 (and/or other conductive projections) may vary. For instance, each nanowire 622 may or may not be elongated like a nanowire. Thus, other types and shapes of nanostructures or other conductive projections from the substrate 620, such as various shaped nanocrystals, may be used.

In some cases, one or more of the nanowires 622 is configured to generate electron-hole pairs upon illumination. For instance, the nanowires 622 may be configured to absorb light at frequencies different than other light absorbing components of the electrode 608. For example, one light absorbing component, such as the substrate 620, may be configured for absorption in the visible or infrared wavelength ranges, while another component may be configured to absorb light at ultraviolet wavelengths. In other cases, the nanowires 622 are the only light absorbing component of the electrode 608.

The electrode 608 further includes nanoclusters 624 disposed over the array of nanowires 622. Each nanocluster 624 is configured for the catalytic conversion of carbon dioxide (CO₂) in the chemical cell 602. A plurality of the nanoclusters 624 are disposed on each nanowire 622, as schematically shown in FIG. 6. The nanoclusters 624 are distributed across the outer surface of each nanowire 622. For example, each nanowire 622 has a plurality of the nanoclusters 624 distributed across or along sidewalls of the nanowire 622. The nanoclusters 624 may also be disposed on a top or upper surface of each nanowire 622. The distribution may or may not be uniform or symmetric as shown. As described herein, each nanocluster 624 may be composed of, or otherwise include copper for the reduction of carbon dioxide (CO₂) in the chemical cell 602.

Alternative or additional metals may be used. For instance, other d-block metal elements may be used, including, for instance, silver, gold, platinum, and zinc. The use of alternative or additional metals may lead to alternative or additional reduction products of the CO₂ conversion. In some cases, additional nanoclusters and/or nanoparticles may be used in combination with the copper (or other metal) nanoclusters described herein, including nanoparticles composed of, or otherwise including one or more noble metals, such as gold.

The nanoclusters 624 may be sized in a manner to facilitate the CO₂ reduction. The size of the nanoclusters 624 may be useful in catalyzing the reaction, as described herein. The size of the nanoclusters 624 may be promote the CO₂ reduction in additional or alternative ways. For instance, the nanoclusters 624 may also be sized to avoid inhibiting the illumination of the light absorber (e.g., the substrate 620).

The manner in, or extent to, which the array of nanowires 622 is ordered may vary. In some cases, the nanowires 622 may be arranged laterally in a regular or semi-regular pattern. In other cases, the lateral arrangement of the nanowires 622 is irregular. In such cases, the ordered nature of the nanowires 622 is instead limited to the parallel orientation of the nanowires 622.

In some cases, each nanowire 622 is coated with the nanoclusters 624. The extent of the coating may vary. For instance, a top surface of each nanowire 622 may be entirely coated with the nanoclusters 624, while one or more portions of the sidewalls of the nanowires 622 may be partially coated. The distribution of the nanoclusters 624 may accordingly be uniform or non-uniform. The nanoclusters 624 may thus be distributed randomly across each nanowire 622. The schematic arrangement of FIG. 6 is shown for ease in illustration.

The nanowires 622 and the nanoclusters 624 are not shown to scale in the schematic depiction of FIG. 6. The shape of the nanowires 622 and the nanoclusters 624 may also vary from the examples shown and described herein.

As described herein, each nanocluster 624 is coupled to a respective nanowire 622 via an oxidized interface. The oxidized interface may have a sub-nanometer thickness. The oxidized interface includes an oxide species of the metal (e.g., Cu^δ+). The oxidized interface further includes an oxynitride species based on the nitride semiconductor. For instance, in GaN cases, the oxynitride species may be a Ga—N—O compound. As also described herein, each nanowire 622 may be covered by the oxynitride species away from the nanoclusters 624.

These and other aspects of the oxidized interface are configured for catalytic conversion of carbon dioxide (CO₂).

In some cases, each nanocluster 624 has a distorted lattice. For instance, each nanocluster 624 has one or more defective surfaces. The nanoclusters 624 may thus present low-coordination sites to promote C—C coupling to multi-carbon products. The surfaces (e.g., outer surfaces) of each nanocluster 624 not at the oxidized interface may not be oxidized, as described herein.

FIG. 7 depicts a method 700 of fabricating an electrode of an electrochemical system in accordance with one example. The method 700 may be used to manufacture any of the working electrodes described herein or another electrode or device. The method 700 may include additional, fewer, or alternative acts. For instance, the method 700 may or may not include one or more acts directed to synthesizing nanoclusters (act 708). The nanoclusters may thus be previously prepared via another process.

The method 700 may begin with an act 702 in which a substrate is prepared. The substrate may be or be formed from a p-n Si wafer. In one example, a 2-inch Si wafer was used, but other (e.g., larger) size wafers may be used. Other semiconductors and substrates may be used. Preparation of the substrate may include one or more cleaning procedures.

The act 702 may include an act 704 in which the substrate is doped. Thermal diffusion and/or other procedures may be used. The doping may be directed to forming a junction. The substrate may accordingly be doped with p-type dopant(s) and n-type dopant(s). An act 706 may then be implemented to anneal the substrate.

In some cases, an n⁺-p silicon junction of the substrate is formed through a standard thermal diffusion process using, e.g., a (100) silicon wafer. For instance, phosphorus and boron as n-type and p-type dopants, respectively, may be deposited on the front and back sides of the polished p-Si (100) wafer by spin-coating, but other dopants may be used. The wafer may then be annealed, e.g., at 950 degrees Celsius under nitrogen atmosphere for four hours. The process parameters may vary in other cases. For instance, the wafer may be annealed at 900 degrees Celsius under argon atmosphere.

In the example of FIG. 7, the method 700 includes an act 708 in which metal nanoclusters are synthesized. The act 708 may include implementation of a solution-based chemical reduction process. The synthesis procedure may vary, e.g., in accordance with the metal. In the copper-based example described herein, 1.0 mL CuCl₂·2H₂O aqueous solution (0.2 M) is mixed with 0.5 mL ethanol to form a homogeneous solution. Then, 1.0 mL sodium borohydride aqueous solution (5.0 M) is added to above solution and reacted for 30 minutes. The resulting products are centrifuged, washed by degassed distilled water and ethanol several times, and finally dispersed in ethanol for further use.

The method 700 includes an act 710 in which GaN or other nanowire arrays (or other conductive projections) are grown or otherwise formed on the substrate. Each nanowire (or other conductive projection) is composed of, or otherwise includes, a nitride semiconductor as described herein. The nanowire growth may be achieved in an act 712 in which plasma-assisted molecular beam epitaxy (MBE) is implemented. The growth may be implemented under nitrogen-rich conditions in accordance with an act 714.

In the GaN examples described herein, plasma-assisted MBE was used for growing GaN nanowires on silicon wafer under nitrogen-rich conditions. The substrate temperature was 790° C. and the growth duration was about 2 hours. The forward plasma power was 350 W with a Ga flux beam equivalent pressure (BEP) of 5×10⁻⁸ Torr.

The growth parameters may vary in other cases. For instance, in another example, the growth conditions were as follows: a growth temperature of 790° C. for 1.5 hours, a Ga beam equivalent pressure of about 6×10⁻⁸ Torr, a nitrogen flow rate of 1 standard cubic centimeter per minute (sccm), and a plasma power of 350 Watts. The substrate and the nanowires provide or act as scaffolding for the catalysts deposited in the following steps.

In an act 716, the nanoclusters are deposited across each nanowire or other conductive projection. The act 716 may include an act 718 in which a drop-cast assembly process is implemented. In the Cu examples described above, Cu catalyst ink was diluted to a desired concentration using an ethanol/Nafion mixed solution and drop cast onto the GaN nanowire/Si support structure to realize the desired mass loading. The structure was then dried under N₂ atmosphere for 30 min before use. Additional or alternative types of deposition procedures may be used.

At this stage of the method 700, each nanocluster has a metallic composition. For instance, the nanoparticles may be composed of, or otherwise include, Cu. Additional or alternative d-block metals may be included in the metallic composition, including, for instance, Ag, Au, Zn, and combinations thereof.

The method 700 includes an act 720 in which an electrochemical procedure is implemented after the deposition of the nanoclusters to form an oxidized interface between each nanocluster and a respective nanowire. The act 720 may thus be considered to activate the catalyst arrangement of the device as described herein.

As described herein, the oxidized interface includes an oxide of the metal of the nanoclusters as well as an oxynitride species based on the nitride semiconductor of the nanowires. The oxidized interface may be configured for catalytic conversion of carbon dioxide (CO₂) and/or another reaction.

In some cases, the act 720 includes an act 722 in which a photochemical reduction reaction is conducted as the procedure. The activation procedure may or may not be similar or identical to the CO₂ reduction reaction conditions under which the device is operated.

In one example, a photoelectrochemical CO₂ reduction reaction was conducted in a CO₂ saturated electrolyte at reductive potential under light illumination (100 mW/cm²). The electrolyte used for the electrochemical measurements was an aqueous solution of 0.1 M KHCO₃ (Sigma-Aldrich, 99.95%) prepared by dissolving the solid salt in deionized water.

The nature of the electrochemical procedure may vary from the examples described above in one or more ways. For instance, the composition of the electrolyte and/or the gas mixture may vary.

The method 700 may include further acts. For instance, the device may be further processed to define contacts for photoelectrochemical operation. In the examples described herein, GaIn eutectic alloy droplets were spread on the back surface of the GaN nanowire/Si support structure. Then, Cu wires were connected to the backside and the edges of the electrodes were covered by epoxy.

Described above are examples of photoelectrochemical synthesis of valuable multi-carbon products from carbon dioxide, sunlight and water. The disclosed devices overcome previous challenges in creating and stabilizing efficient C—C coupling sites to achieve multi-carbon products with high selectivity, yield, and stability. Examples included a low-coordinated copper cluster catalyst in-situ interfacially coupled with a GaN nanowire photocathode, achieving a high ethylene Faradic efficiency of about 61% and partial current density of 14.2 mA cm⁻², with a robust stability of about 116 hours. The in situ self-optimized Ga—N—O interface was confirmed to facilitate and stabilize the interfacially oxidized copper species of copper clusters, which function as efficient C—C coupling sites for ethylene production. Furthermore, the hydrogen feeding effect of GaN for promoting CO hydrogenation also guides the CHO involved C—C coupling pathway. The disclosed devices thus provide a useful interface design and approach to efficient and stable (photo)electrosynthesis of highly valuable fuels and other products from CO₂.

The term “about” is used herein in a manner to include deviations from a specified value that would be understood by one of ordinary skill in the art to effectively be the same as the specified value due to, for instance, the absence of appreciable, detectable, or otherwise effective difference in operation, outcome, characteristic, or other aspect of the disclosed methods and devices.

The present disclosure has been described with reference to specific examples that are intended to be illustrative only and not to be limiting of the disclosure. Changes, additions and/or deletions may be made to the examples without departing from the spirit and scope of the disclosure.

The foregoing description is given for clearness of understanding only, and no unnecessary limitations should be understood therefrom.