Silicon Nanowire-Supported Catalysts for the Photocatalytic Reduction of Carbon Dioxide ^†

Abstract

The conversion of carbon dioxide, a greenhouse gas, into valuable chemicals using sunlight is highly significant technologically and holds the promise of providing a more sustainable alternative to fossil fuels. To effectively utilize the abundant solar irradiation, it is essential to develop catalysts that can absorb a significant portion of the solar spectrum, particularly in the UV, visible, and infrared regions. Silicon nanowire arrays grown on silicon substrates meet this criterion, as they can absorb over 85% of solar irradiation and show minimal reflective losses across the UV, visible, and infrared portions of the solar spectrum. Herein, we report the deposition of various catalysts, including iron oxyhydroxides, copper, nickel, and ruthenium, on silicon nanowires using different catalyst deposition techniques. The photocatalytic reduction of carbon dioxide was evaluated using these catalysts. The results show that silicon nanowires coated with nickel and ruthenium oxide had the highest activity towards the photocatalytic reduction of carbon dioxide, with photomethanation rates reaching 546 μmolg_cat⁻¹h⁻¹ for RuO₂@SiNWs and 278 μmolg_cat⁻¹h⁻¹ for Ni/NiO@SiNWs. Continued improvement of photocatalysts using nanostructured silicon supports could enable the development of solar refineries for converting gas-phase CO₂ into value-added chemicals and fuels.

1. Introduction

The concentration of carbon dioxide (CO₂) in the atmosphere, mainly due to the combustion of fossil fuels, is rapidly rising and has reached unprecedented levels over the last decade. As global fossil fuel consumption increases, CO₂ emissions are expected to continue climbing to potentially hazardous levels for life on Earth. To address this, researchers have explored the production of solar fuels, a process that uses sunlight to drive the chemical reaction between CO₂ and water to produce useful fuels such as methanol, dimethyl ether, carbon monoxide, or methane [1,2,3]. If this process can be carried out with high technological efficiency, on a global scale, and at a cost that competes with fossil fuels, achieving a sustainable future becomes a realistic objective. This process requires the development of photoactive materials that can harness the abundant solar energy from the sun. This means the catalyst should be capable of absorbing a significant portion of the solar spectrum, particularly visible light [4]. One such material is silicon, particularly in its nanostructured form as silicon nanowires (SiNWs). Due to their unique structural, optical, and electronic properties, SiNWs have garnered significant interest as the material of choice in energy conversion and storage applications, including photovoltaics, thermoelectrics, and photocatalysis. Additionally, silicon is earth-abundant, low-cost, and scalable. Furthermore, with a band gap of 1.1 eV, SiNWs have broad-band and efficient solar spectral harvesting properties, traversing the near-infrared, visible, and ultraviolet wavelength range, and can be tailored to absorb over 85% of the solar irradiance [5]. In particular, when silicon nanowire arrays are produced from vertically etched silicon wafers, they allow for very low reflective losses across the solar spectral wavelength range [6]. However, SiNWs on their own have been shown not to catalyze CO₂ reduction reactions [6]. As a result, different catalysts are deposited onto the SiNWs that can efficiently transform CO₂ into useful products [7,8]. In this architecture, the SiNW arrays serve two purposes: to absorb most of the visible spectral range and to transport electronic charge to co-catalysts deposited on them to catalyze the chemical transformation of CO₂. Herein, we describe the fabrication of vertically aligned SiNW arrays, which is accomplished by the use of a fast, cheap, and room-temperature synthetic route known as metal-assisted chemical etching (MaCE) of silicon [9]. We then use these SiNW arrays to deposit a variety of photocatalysts employed for the photocatalytic conversion of CO₂ into useful chemical fuels and feedstocks.

2. Experimental

2.1. Materials

Silicon wafers (University wafers, 1–100 Ω/cm), ethanol (99.8%), acetone (99.5%), deionized water, sulphuric acid (99.99%), hydrofluoric acid (48%HF), silver nitrate (AgNO₃) (99.0%), nitric acid (HNO₃) (60%), Fe₂(SO₄)₃ (97%), and FeCl₃ (99.9%).

2.2. Silicon Nanowire Fabrication

Silicon nanowire arrays were fabricated using a top-down fabrication technique known as metal-assisted chemical etching (MaCE), as previously reported [10,11]. Briefly, p-type silicon wafers were cut into 1-inch squares and cleaned with ethanol, acetone, and deionized water. The wafers were further cleaned by soaking in piranha solution (H₂SO₄:H₂O₂ = 3:1 by volume) for 3 h and then rinsed with deionized water. To fabricate the silicon nanowires, the cleaned wafers were immersed in 23 mL of an etching solution consisting of 5 M HF and 0.02 M AgNO₃ (VWR redi-pak) and allowed to etch for 1 h at room temperature. After etching, the sample was removed from the etching solution, and the silver dendrites covering the silicon nanowires were washed off with deionized water. To ensure complete removal of the silver nanoparticles and dendrites, the etched wafers were placed in concentrated nitric acid (18 M HNO₃) for 1 h. The wafers were then washed, dried, and cut into 1 cm² pieces. The different catalysts were deposited onto the silicon nanowire substrates through different techniques, as discussed in Section 3.

2.3. Catalyst Characterizations

Scanning and transmission electron microscopy were used to study the morphology and structure of the samples. A Hitachi S-5200 SEM was employed to examine the surface topography of the samples, while a Hitachi H-3300 TEM operating at a voltage of 300 kV was utilized for further characterization. X-ray photoelectron spectroscopy (XPS) measurements were conducted to analyze the surface and electronic structure of the samples. XPS was performed in an ultrahigh vacuum chamber with a base pressure of 10⁻⁹ mTorr. The system used a Thermo Scientific K_α XPS spectrometer (Thermo Fisher Scientific, Waltham, Massachusetts, USA), equipped with an Al K_α X-ray source operating at 12 kV and 6 A, with an X-ray energy of 1486.7 eV. The spectra were acquired with an analyzer pass energy of 50 eV and an energy spacing of 0.1 eV. All data analysis was carried out using Thermo Scientific Avantage^TM V6 software.

2.4. Photocatalytic Measurements

The performance of the catalysts for the photocatalytic reduction of CO₂ was evaluated using a custom-built photoreactor system. The reactor used is a 6.5 mL stainless steel vessel designed to hold a 1” by 1” (6.45 cm²) sample. The samples are placed in the photoreactor and the reactor, which operates at a temperature of 80 °C. Initially, the reactor is evacuated to remove any pre-existing gasses, then carbon dioxide and hydrogen gasses are injected. During the photocatalytic reaction, the samples are illuminated with a 400 W metal halide bulb. The concentrations of the products are determined by sampling the product gasses using either gas chromatography (GC) or combined gas chromatography-mass spectrometry (GC/MS).

3. Results and Discussion

3.1. Synthesis and the Characterizations of the Silicon Nanowire-Supported Catalysts

Various catalysts have been deposited onto silicon nanowires to create hybrid catalysts capable of driving the photocatalytic conversion of carbon dioxide into useful solar fuels. These catalysts include iron oxyhydroxide (FeOOH), copper, nickel, and ruthenium, as discussed in the subsequent sections.

3.1.1. Iron Oxyhydroxide/Silicon Nanowire Catalysts

Both α-FeOOH and β-FeOOH polymorphs were successfully grown onto the silicon nanowire substrate through a hydrothermal synthetic technique to produce α-FeOOH@SiNWs and β-FeOOH@SiNWs, respectively. An Fe₂(SO₄)₃ precursor was used to grow the α-FeOOH on the SiNWs, while an FeCl₃ precursor was used to grow the β-FeOOH on the SiNWs. To synthesize the α-FeOOH, 350 mg of Fe₂(SO₄)₃ was placed in 20 mL of water, stirred, and sonicated until completely dissolved. The solution was poured into an autoclave containing a silicon nanowire substrate at the bottom. It was then heated for 6 h at 125 °C to produce α-FeOOH@SiNWs. The same procedure was used to grow the β-FeOOH onto the SiNWs, except an FeCl₃ precursor was used instead of Fe₂(SO₄)₃, and the solution was heated for 2 h at 125 °C instead of 6 h as the hydrothermal synthesis of β-FeOOH is a lot faster than that of α-FeOOH. The α-FeOOH@SiNWs and β-FeOOH@SiNWs catalysts are shown in Figure 1. While the morphology of the α-FeOOH of the α-FeOOH@SiNWs catalyst consists of a mixture of nanoparticles and nanorods (Figure 1a,b), that of the β-FeOOH of the β-FeOOH@SiNWs catalyst consists of rice-like nanoparticles (Figure 1c,d).

3.1.2. Copper/Silicon Nanowire Catalysts

Copper nanoparticles were deposited onto the SiNWs using an electroless deposition technique. This process involves the chemical reduction of copper ions from a solution onto the surface of SiNWs in the presence of a reducing agent. Specifically, a 1” by 1” (6.45 cm²) silicon wafer, etched at the top to form SiNWs, was immersed in a solution containing copper chloride (0.15 M) and HF (0.05 M) for 30 s. Finally, the plate was rinsed with deionized water and dried in atmospheric air. Figure 2a shows the top view, while Figure 2b shows the side view of the Cu@SiNWs samples produced. Various copper deposition times were tested, but a 30 s deposition provided sufficient and uniform copper coverage, resulting in an optimal CO₂ photoreduction rate.

3.1.3. Nickel/Silicon Nanowire Catalysts

For the synthesis of Ni@SiNWs catalysts, nickel nanoparticles were deposited onto SiNWs using a solution of NiCl₂ precursor. Briefly, 20 mg of NiCl₂ was dissolved in 20 mL of water, and then 0.2 mL (0.015 M) of hydrazine was added. The solution was stirred and heated to 50 °C. SiNW substrates were placed in the solution for 15 min, resulting in the formation of Ni@SiNWs catalysts, as shown in Figure 3a,b.

The samples were then calcined at 300 °C for an hour, leading to the oxidation of some nickel nanoparticles to nickel oxide. This process resulted in the production of a Ni/NiO@SiNWs hybrid catalyst, as shown in Figure 3c,d. This heterojunction architecture of Ni/NiO@SiNWs was found to facilitate efficient CO₂ photoreduction due to favorable electron flow facilitated by the positions of the silicon and NiO band gaps relative to the fermi energy of the metallic nickel illustrated in Figure 4. In a heterojunction comprising a metal and a p-type semiconductor—nickel and nickel oxide in this case—at the interface, the Fermi energy level of the nickel will align with that of the nickel oxide. This alignment causes band bending between the nickel and the nickel oxide, facilitating electron transfer from the nickel oxide to the nickel metal, as depicted in Figure 4a. Figure 4b illustrates the band structure of silicon and nickel oxide relative to the Fermi energy position of metallic nickel. In the Ni/NiO/Si heterostructure, such as in Ni/NiO@SiNWs, the alignment of the Fermi levels of the three components enables a cascade of electrons from the nickel oxide to the silicon and finally to the metallic nickel, where the carbon dioxide reduction reaction occurs (Figure 4c). Heterostructures not only facilitate electron transfers but also create charge separation junctions, allowing for the effective separation of electron-hole pairs, thereby providing sufficient time for surface reactions to occur.

3.1.4. Ruthenium Oxide/Silicon Nanowire Catalysts

The deposition of RuO₂ on the silicon nanowires was achieved using a wet chemical deposition method. Initially, the SiNWs sample was immersed in a 48% hydrogen fluoride solution for 1–2 min to remove any SiO₂ from the surface and to terminate the surface with hydrogen. The hydrogen-terminated samples were placed in a solution containing 30 mg of ruthenium (III) nitrosyl chloride hydrate (RuCl₃NO.H₂O) dissolved in 40 mL of water. This solution was heated in a water bath at 40–45 °C, and three to five drops of hydrazine were added to facilitate the reduction of ruthenium and the formation of RuO₂ on the SiNWs substrate. The resulting RuO₂@SiNWs catalyst is shown in Figure 5a. The RuO₂ nanoparticles coated on the SiNWs were extremely small, ranging from 2 to 5 nm, making them unobservable under SEM. However, they are visible in the TEM image shown in Figure 5b.

Furthermore, we utilized XPS to analyze the electronic structure and determine the oxidation state of the ruthenium coated onto the silicon nanowires. Signals from 3d photoelectrons of ruthenium typically provide reliable information about the oxidation states of ruthenium species. As shown in Figure 5c, Ru 3d_5/2 and Ru 3d_3/2 peaks appeared at approximately 281.5 and 285.7 eV, respectively, which is consistent with the peak positions for RuO₂[12]. Additionally, the O1s peak in Figure 5d reveals two distinct peaks. The peak around 530 eV is assigned to the lattice oxygen of RuO₂, while the peak around 532 eV is attributed to oxygen from SiO₂. The presence of SiO₂ is due to the surface oxidation of the SiNWs, resulting in the formation of a silicon dioxide layer.

3.2. CO₂ Photocatalysis Performance Comparisons of the Different Catalysts

The performance of the catalysts (α-FeOOH@SiNWs, β-FeOOH@SiNWs, Cu@SiNWs, Ni/NiO@SiNWs, and RuO₂@SiNWs) for the photocatalytic reduction of CO₂ was evaluated. The results, displayed in Figure 6, show the rates of product formation for these catalysts at 80 °C under illumination with a 400 W metal halide bulb. The only products obtained from the photocatalytic reduction of CO₂ in a hydrogen environment are CO and CH₄, as displayed in Figure 6.

Both α-FeOOH@SiNWs and β-FeOOH@SiNWs produced more CO than CH₄, while the Cu@SiNWs, Ni/NiO@SiNWs, and RuO₂@SiNWs exhibited higher CH₄ selectivities. Although the silicon nanowires of Cu@SiNWs and Ni/NiO@SiNWs catalysts had similar nanoparticle coverages, the CH₄ formation rate of the Ni/NiO@SiNWs catalyst was significantly higher (278 μmolg_cat⁻¹h⁻¹) than that of Cu@SiNWs (124 μmolg_cat⁻¹h⁻¹). This enhanced performance is attributed to the efficient electron separation facilitated by the alignment of the silicon and NiO band gaps relative to the fermi energy of the metallic nickel in Ni/NiO@SiNWs, which provides electrons for the efficient photoreduction of CO₂. Of all the catalysts tested, RuO₂@SiNWs had the highest CH₄ production rates, reaching up to 546 μmolg_cat⁻¹h⁻¹. Despite ruthenium being a very expensive rare earth metal, only a minimal amount of RuO₂ was required to coat the SiNWs compared to other catalysts.

4. Conclusions

In summary, we have deposited several nanocrystal catalysts onto silicon nanowire arrays grown on silicon wafers. These catalysts have been characterized using SEM, TEM, and XPS techniques. Their performance for the photocatalytic reduction of carbon dioxide to useful fuels was then evaluated. The highest photomethanation rate was 546 μmolg_cat⁻¹h⁻¹ for RuO₂@SiNWs, followed by 278 μmolg_cat⁻¹h⁻¹ for Ni/NiO@SiNWs. Utilizing silicon nanowire arrays, which can harvest nearly all incident photons from solar radiation as a photocatalytic support, has the potential to be an efficient method for converting carbon dioxide, a major greenhouse gas, into solar fuels.