Electrocatalytic CO₂ Reduction over Cu₃P Nanoparticles Generated via a Molecular Precursor Route

Abstract

The design of nanoparticles (NPs) with tailored morphologies and finely tuned electronic and physical properties has become a key strategy for controlling selectivity and improving conversion efficiency in a variety of important electrocatalytic transformations. Transition metal phosphide NPs, in particular, have emerged as a versatile class of catalytic materials due to their multifunctional active sites and composition- and phase-dependent properties. Access to targeted transition metal phosphide NPs with controlled features is necessary to tune the catalytic activity. To this end, we have established a solution-synthesis route utilizing a molecular precursor containing M–P bonds to generate solid metal phosphide NPs with controlled stoichiometry and morphology. We expand here the application of molecular precursors in metal phosphide NP synthesis to include the preparation of phase-pure Cu₃P NPs from the thermal decomposition of [Cu(H)(PPh₃)]₆. The mechanism of [Cu(H)(PPh₃)]₆ decomposition and subsequent formation of Cu₃P was investigated through modification of the reaction parameters. Identification and optimization of the critical reaction parameters (i.e., time, temperature, and oleylamine concentration) enabled the synthesis of phase-pure 9–11 nm Cu₃P NPs. To probe the multifunctionality of this materials system, Cu₃P NPs were investigated as an electrocatalyst for CO₂ reduction. At low overpotential (−0.30 V versus RHE) in 0.1 M KHCO₃ electrolyte, Cu₃P-modified carbon paper electrodes produced formate (HCOO−) at a maximum Faradaic efficiency of 8%.

Introduction

The projected availability of abundant renewable electrons presents an opportunity to sustainably produce valuable chemical and fuel products from waste streams and renewable carbon sources such as CO₂ and biomass.¹⁻⁴ These underutilized feedstocks can serve as not only precursors for the synthesis of fuels and chemicals but also liquid energy carriers to improve power grid management by mitigating the impact of the intermittency associated with renewable energy generation. However, the marriage of these renewable electrons with underutilized renewable feedstocks to improve grid stability and displace fossil-derived chemical production requires the development of economical routes for the electrochemical transformations with high selectivity and Faradaic efficiency (FE). These critically important transformations are challenging because of the multiple proton and electron transfers and bond breaking and bond forming steps needed to generate valuable products.¹⁻³ Additionally, undesirable reactions such as H₂ evolution are thermodynamically competitive under the reaction conditions utilized for these transformations, which can lead to reduced efficiencies toward valuable products. Therefore, the development of catalytic systems to suppress competitive reactions is necessary to improve the overall energy efficiency and viability of the electrochemical conversion of waste streams and renewable carbon sources to valuable fuels and chemicals.

Electrocatalysts with sophisticated features, such as multifunctional active sites and tailored electronic properties, have been explored to overcome the difficulties associated with these transformations.⁵⁻⁸ Transition metal phosphides have emerged as a versatile materials platform with interesting structural and electronic properties that have led to exemplary catalytic performance for important energy conversions^9,10 such as electrocatalytic hydrogen evolution.^11,12 The incorporation of phosphorus into the metal structure imparts unique electronic properties, such as extensive hybridization of the M and P orbitals and charge transfer between the M and P sites.¹³ Surface phosphorus sites can also actively participate in the catalytic transformation through direct coordination with or stabilization of reaction intermediates.^11,14 Because of this colocalized multifunctionality, tuning the stoichiometry and corresponding phase of metal phosphides confers a high degree of control over the catalytic selectivity, efficiency, and activity.

Despite the interesting properties of transition metal phosphides, the favorable thermodynamics and kinetics of H₂ evolution in aqueous solution has traditionally limited their application as electrocatalysts for other important energy conversions such as the electrochemical reduction of CO₂.¹⁵ However, there are several recent examples of transition metal phosphides electrochemically reducing CO₂ selectively to C₁ and even more complex C₃ and C₄ oxygenates.¹⁶⁻¹⁸ Bulk nickel phosphides (Ni₃P, Ni₂P, Ni₁₂P₅, Ni₅P₄, and NiP₂) were observed to electrochemically reduce CO₂ to methylglyoxal and 2,3-furandiol via formate as an intermediate reaching a maximum FE of 71% for 2,3-furandiol at 0.00 V versus RHE over Ni₂P.¹⁶ Higher conversion of CO₂ to C₃ and C₄ oxygenates, rather than formate and H₂, was observed for the more phosphorus-rich nickel phosphides, implicating phosphorus in the enhancement of CO₂ reduction efficiency through the formation of reactive hydrides and nucleophilic CO₂ binding sites that facilitate the generation of formate and C₃₊ oxygenates. The phosphorus-dependent CO₂ reduction activity observed for nickel phosphides highlights the advantage of colocalized multifunctionality.¹⁹ A Cu₃P/C composite was also shown to electrochemically reduce CO₂ to CO at −0.30 V versus RHE with an average FE of 47%.¹⁸ These results suggest the possibility of transition metal phosphides serving as a unique materials platform whereby CO₂ reduction can outperform H₂ evolution within a certain potential range.

Due to the unique CO₂ reduction capabilities of copper-based electrocatalysts²⁰⁻²² and the copper-based phosphide material,¹⁸ we identified nanostructured Cu₃P as a promising synthetic target. Nanostructuring offers the benefits of high surface area and efficient metal utilization, and provides a tunable platform whereby the structural, physical, and electronic properties can be controlled to achieve the desired catalytic activity.^5,7 Solution-based synthesis methods have emerged as a viable strategy to produce nanostructured transition metal phosphides;^13,14 however, solution synthesis routes to Cu₃P nanoparticles (NPs) are limited²³⁻²⁷ in comparison to other metal phosphides. A solution synthesis route to NP nickel, rhodium, and palladium phosphides (Ni₂P, Rh₂P, and Pd₃P) was recently reported using commercially available molecular precursors, which afforded control over the morphology and phase of the resultant metal phosphide NPs.²⁸ The metal phosphide NPs synthesized from the molecular precursor route were shown to catalyze a variety of thermochemical and electrochemical transformations.²⁸⁻³⁰ Herein, we present a synthetic route to phase-pure Cu₃P NPs via thermal decomposition of the molecular precursor [Cu(H)(PPh₃)]₆ and elucidate the mechanism of NP formation through an investigation of key reaction parameters. The resulting Cu₃P NPs were evaluated as electrocatalysts for CO₂ reduction in CO₂-saturated KHCO₃ aqueous solutions.

Experimental Section

General

Synthetic manipulations to prepare Cu₃P NPs were conducted in a N₂ atmosphere using standard Schlenk techniques or in a N₂-filled Vacuum Atmospheres glovebox. Oleylamine (OAm, 70% technical grade) and 1-octadecene (ODE, 90%) were purchased from Sigma-Aldrich and dried prior to use by heating to 120 and 150 °C, respectively, under vacuum for 5 h and stored in a N₂-filled glovebox. Triphenylphosphine (PPh₃, 99%) was purchased from Sigma-Aldrich, and [Cu(H)(PPh₃)]₆ (96%) was purchased from Acros Organics and stored in a N₂-filled glovebox. Water used for electrochemistry experiments was purified by a Milli-Q Water Purification System with a specific resistance of 18.2 MΩ·cm at 25 °C. Standards for high-performance liquid chromatography (HPLC) experiments were purchased from AccuStandards. Deuterium oxide (D 99.9%) containing 0.05 wt % 3-(trimethylsilyl)propionic-2,2,3,3-d₄ acid (Na salt), carbon tetrachloride (99.99% ACS reagent), tetrachloroethylene, potassium bicarbonate (99.95%), sodium chloride (99.999%), and Chelex 100 (Na form) were purchased from Sigma-Aldrich. Fluorine doped tin oxide (FTO) coated glass was purchased from Techinstro.

Synthesis of Cu₃P Nanoparticles

In a three-neck round-bottom flask fitted with a condenser and two septa, [Cu(H)(PPh₃)]₆ (0.326 g, 1 mmol Cu) was combined with dried OAm (4.9 mL, 15 mmol) and ODE (8.0 mL, 20 mmol) and heated to 250 °C under N₂ with rapid stirring. The mixture was held at 250 °C for 30 min and then heated to 320 °C (ca. 3.5 °C/min). The reaction was maintained at 320 °C for 15 min, followed by the removal of the heat source and ambient cooling. Approximately 5 mL of CHCl₃ was added to the reaction mixture in air followed by the addition of 30 mL of ethanol to flocculate the particles, which were then separated by centrifugation at 10000 rpm for 5 min. This washing procedure was performed an additional 4 times to remove impurities and OAm was added (ca. 0.5–1.0 mL) to each wash to ensure retention of the dispersibility of the NPs. The washed NPs were dried and stored in a N₂-filled glovebox prior to evaluation. Following the complete washing procedure, a 60% yield of Cu₃P NPs was calculated from elemental analysis of the Cu concentration of a representative sample. Cu₃P NPs were supported on carbon powder by adding a CHCl₃ suspension of NPs dropwise to a rapidly stirring suspension of Vulcan XC 72R (Cabot) in CHCl₃ (ca. 100 mL). The mixture was sonicated for 5 min, stirred overnight, and recovered by centrifugation at 8000 rpm for 10 min. The resultant carbon-supported Cu₃P NPs were dried under vacuum overnight and stored in an N₂-filled glovebox.

Modification of Cu₃P Reaction Parameters

To understand the mechanism of Cu₃P formation, the following reaction parameters were modified: time, temperature, OAm concentration, and phosphorus concentration. A 30 min hold at 250 °C was employed for all reactions. Following the hold at 250 °C, the reaction mixture was heated to the target temperature and held for the designated time. To assess the role of phosphorus concentration, PPh₃ (2–4 equiv relative to Cu) was added to the reaction mixture containing [Cu(H)(PPh₃)]₆, OAm (15 mmol), and ODE (25 mmol). Cu₃P NPs synthesized in the presence of added PPh₃ are denoted generally as P–Cu₃P with OAm–P–Cu₃P and R–P–Cu₃P referring to the as-synthesized and reduced derivatives, respectively. When the OAm concentration was changed, ODE was used to keep the total solvent volume at 12.9 mL. Aliquots (0.5 mL) of the reaction mixture were extracted via syringe over the duration of the experiment to investigate the properties of the NPs at specific temperatures and times. The aliquots were purified by addition of ethanol followed by centrifugation to separate the particles.

Characterization of Cu₃P

Powder X-ray diffraction (XRD) data on the Cu₃P materials were collected using a Rigaku Ultima IV diffractometer with a Cu Kα source (40 kV, 44 mA). Diffraction patterns were collected in the 2θ range of 20–80° at a scan rate of 4°/min. The as-prepared NPs were drop-cast onto a glass slide from a chloroform suspension. The resulting patterns were compared to powder diffraction files (PDF) from the International Centre for Diffraction Data (ICDD). NIST Si standard was used to calibrate the XRD peak positions. For transmission electron microscopy (TEM) analysis, the unsupported NPs were drop-cast onto continuous carbon-coated copper grids (Ted Pella part no. 01824) from chloroform suspensions. Imaging was performed using an FEI G² T20 Tecnai TEM operated at 200 kV and an FEI Tecnai G² ST30 TEM operated at 300 kV, and all image analysis was conducted with ImageJ software.³¹ Lattice spacings were measured from the fast-Fourier transforms (FFTs) of high-resolution TEM (HRTEM) images. Size distributions were determined from a manual diameter measurement of >100 particles. The metal and phosphorus loadings of the as-prepared NPs and the reaction yield were determined by inductively coupled plasma optical emission spectroscopy (ICP-OES) performed by Galbraith Laboratories Inc., (Knoxville, TN). For FTIR spectroscopy analysis, the as-prepared NPs were drop-cast onto a silicon wafer and analyzed using a Thermo Scientific Nicolet 6700 FT-IR Spectrometer (4000 to 1500 cm^–1, 32 scans, 4 cm^–1 resolution). UV–vis-NIR spectroscopy analysis of as-prepared Cu₃P NPs suspended in CCl₄ and C₂Cl₄ were conducted using an Agilent Technologies Cary Series UV–vis-NIR Spectrophotometer (2500 to 500 nm, 600 nm/min). In situ X-ray absorption spectroscopy (XAS) measurements were performed at the 10-BM beamline at the Advanced Photon Source (APS) at Argonne National Laboratory. All measurements were performed at the Cu K edge (8.979 keV) in transmission mode in fast scan from 250 eV below the edge to 800 eV above the edge, which took approximately 10 min per scan. At the Cu K edge, the Cu–P (CN = 1, R = 2.34 Å), Cu–Cu (CN = 1, R = 2.56 Å), and Cu–O (CN = 1, R = 1.95 Å) scattering pairs were simulated. S_o² was calibrated by fitting the Cu foil, which gave a value of 0.67. A least squared fit the first shell of r-space and isolated q-space were performed on the k² weighted Fourier transform data over the range 2.7 to 11 Å^–1 in each spectrum to fit the magnitude and imaginary components. XAS measurements were performed on Cu₃P NPs synthesized in the presence of 2 equiv of PPh₃ that were supported on carbon to achieve a nominal loading of 5 wt % (P–Cu₃P/C). P–Cu₃P/C was pressed into a stainless-steel sample holder and placed in a sample cell. The cell was sealed and transferred to the beamline for measurement. XAS measurements were first collected on the as-synthesized carbon-supported Cu₃P NPs (OAm–P–Cu₃P/C) at ambient temperature. OAm–P–Cu₃P/C was then treated in flowing 5% H₂/He (100 sccm) at 450 °C for 2 h to generate reduced carbon-supported Cu₃P NPs (R–P–Cu₃P/C), cooled to ambient temperature in He (100 sccm), sealed, and transferred to the beamline for a second measurement. Without exposure to air, R–P–Cu₃P/C was then passivated by treating in flowing 1% O₂/He (100 sccm) at ambient temperature for 1 h, sealed in 1% O₂, and transferred to the beamline for a third measurement.

Preparation of Cu₃P-Modified Working Electrodes

Coiled copper wire was affixed to carbon paper (CP) (Fuel Cell Store, Spectracarb 2050A 0850) using silver epoxy and allowed to dry. The copper wire was fitted through a glass tube and two-part epoxy was used to cover the exposed copper wire and silver epoxy, and to define a 1 cm × 1 cm working area. Once the epoxy was dried, as-synthesized Cu₃P-modified CP (OAm–Cu₃P/CP) were prepared by drop-casting Cu₃P NPs suspended in chloroform onto CP to produce loadings of 1–1.5 mg/cm². The loadings were calculated based on the total mass of NPs. As-synthesized Cu₃P-modified FTO electrodes were prepared in an analogous fashion to OAm–Cu₃P/CP. To prepare reduced Cu₃P-modified CP electrodes (R–Cu₃P/CP), Cu₃P NPs were suspended in chloroform and drop cast onto CP to produce loadings of 1 mg/cm². After drying, the Cu₃P-modified carbon papers were reduced in a tube furnace in flowing (500 sccm) 5% H₂/N₂. The temperature was increased at 5 °C/min to 450 °C and held for 2 h. R–Cu₃P/CP were cooled to below 50 °C in flowing 5% H₂/N₂. Due to the air-sensitive nature of Cu₃P and the inability to completely eliminate air exposure when preparing and executing the electrochemistry experiments, the samples were passivated in flowing (500 sccm) 1% O₂/N₂ for 1 h at ambient temperature. This passivation step allows for a controlled oxidation of the catalyst surface in contrast to the rapid oxidation expected upon directly exposing the reduced catalyst to air. Coiled copper wire was affixed to R–Cu₃P/CP using silver epoxy and allowed to dry. The copper wire was fitted through a glass tube, and two-part epoxy was used to cover the exposed copper wire and silver epoxy and to define a 1 cm × 1 cm working area. The Cu₃P-modified working electrodes were stored in a N₂-filled glovebox prior to use.

Electrochemical Methods

Electrochemical measurements were performed in 0.1 and 0.5 M KHCO₃ aqueous electrolyte in a two-compartment, three-electrode cell utilizing a Metrohm Autolab potentiostat. Metallic impurities in the as-prepared electrolyte were removed before electrolysis by chelating the solution with Chelex 100. Potassium bicarbonate electrolytes were sparged with CO₂ for 30 min prior to electrochemical analysis. The two compartments were separated by an anion-exchange membrane (Selemion AMV AGC Inc.). The working chamber of the cell was continuously purged with CO₂. A graphite rod was used as the counter electrode and the reference electrode was an Ag/AgCl (3 M NaCl) electrode separated from the cell by a Vycor frit. Electrochemical impedance spectroscopy was employed to measure the uncompensated resistance (R_u) of the electrochemical cell. The potentiostat’s IR compensation function was used to compensate 85% R_u. The final 15% of R_u was mathematically corrected for after the electrochemical data was collected.

CO₂ Reduction Product Analysis

Gas phase products were quantified with gas chromatography (Agilent Technologies 7890A), equipped with a thermal conductivity detector and a flame ionization detector. The catholyte and anolyte from each electrolysis experiment were analyzed by nuclear magnetic resonance (NMR) spectroscopy to quantify liquid products, and the NMR results from select electrolysis experiments were confirmed by high-performance liquid chromatography (HPLC). A Bruker Avance Nanobay spectrometer at 9.4 T (400 MHz) equipped with a Bruker 5 mm BBO probe was used to collect ¹H spectra of the electrolyte samples. Both catholyte and anolyte samples were analyzed for liquid products to account for product crossover. Suppression of the water peak was achieved using WATERGATE,³² a recycle delay of 1 s, and a total of 64 scans. Pulse calibration was performed on each electrolyte sample prior to beginning the NMR spectroscopy experiment to account for the KHCO₃ concentration differences between samples. Electrolyte samples were added to an NMR tube containing a capillary filled with 0.05 wt % TSP in D₂O, which was used as the internal standard and reference for the NMR spectroscopy experiments. A calibration curve for formate in KHCO₃ solutions was generated for product quantification. HPLC analysis of the catholyte and anolyte samples was performed using Agilent 1260 LC with a refractive index detector (RID) and 0.02 N H₂SO₄ mobile phase. Samples were prepared using neat reaction media (0.1 or 0.5 M KHCO₃) and adjusted to pH 4 using H₂SO₄. An HPX-87H Aminex column was used with 20 uL injection volume with 0.5 mL/min mobile phase flow rate for 19 min, followed by 0.6 mL/min until 35 min, and final flow rate of 0.5 mL for 15 min, giving a method run time of 50 min. The RID was 55 °C, and the column temperature was programmed at 30 °C. The concentrations of analytes in standards were calculated relative to external calibrations using various standard solutions. The standards consisted of mix 1: formaldehyde and acetaldehyde (catalogue number M-8315); mix 2: hydroxyacetone, methanol, ethanol, acetone and 1-propanol (custom order); mix 3: glyoxal, methylgyoxal, propanal (custom order); and mix 4: formic acid, acetic acid, propionic acid, isobutyric acid, butyric acid, isovaleric acid, valeric acid, isocaproic acid, caproic acid, and heptanoic acid (catalogue number FAMQ-004) in water ranging in stock concentrations of 0.5–1.3 mg/mL. Stock standards were further diluted in 0.5 M KHCO₃ previously adjusted to pH = 4 using H₂SO₄ to generate calibration concentrations ranging from 0.01 to 1 mg/mL.

Results and Discussion

Synthesis and Characterization of Cu₃P Nanoparticles

The commercially available copper-phosphine hexamer, [Cu(H)(PPh₃)]₆, was identified as an attractive molecular precursor for the preparation of phase-pure Cu₃P NPs because of the low valence state of copper and the presence of preformed Cu–P bonds. Thermal decomposition of [Cu(H)(PPh₃)]₆ (1 mmol Cu) in 15 mmol OAm and 25 mmol ODE was performed at 250 °C for 30 min to promote uniform decomposition of the precursor and homogeneous particle growth. The further heating of the reaction mixture to 320 °C and holding at this temperature for 15 min yielded phase-pure Cu₃P NPs. The resultant NPs were extensively washed with ethanol and isolated in 60% yield. The ethanol wash was necessary to remove soluble impurities such as clusters that are commonly formed from polyhydrido-copper complexes like [Cu(H)(PPh₃)]₆.³³ Without the extensive washing procedure with ethanol, impurities were observed by XRD. In comparison to the reported methods to access phase-pure Cu₃P NPs, the procedure established here has the combined advantages of being a single-pot synthesis that requires only a short-reaction time (<1 h) and does not rely on highly reactive phosphine sources. The size and morphology of Cu₃P NPs isolated by the synthetic procedure developed here are similar to those previously reported.

Analysis by XRD indicated the formation of hexagonal Cu₃P (P6₃cm) with good agreement to the reference pattern (PDF 01–071–2261) without any other observed crystalline phases, as shown in Figure 1a. The Cu₃P NPs are single-crystalline and solid with a size distribution of 8.7 ± 1.6 nm as determined by analysis of TEM images (Figure 1b, Figure S1a). The high-resolution TEM (HRTEM) image of hexagonal Cu₃P presented in Figure 1c and the corresponding FFT (Figure 1d) indexed to the [0001] zone axis have measured lattice spacings of 0.21 and 0.36 nm corresponding to the (112̅0) and (011̅0) crystal planes, respectively. Increasing the hold time at 320 °C to 30 min led to further growth of the NPs with analysis of TEM images indicating a size distribution of 10.3 ± 2.3 nm (Figure S1b). Longer reaction times of 1 h or more at 320 °C resulted in precipitation of nondispersible Cu₃P aggregates. The optical properties of Cu₃P were probed using UV–vis-NIR spectroscopy, and a plasmonic absorbance in the NIR region was observed (Figure S2). This plasmonic absorbance has been attributed to Cu vacancies in the NPs, and the presence and prevalence of these vacancies in copper phosphide NPs has led to the use of the formula Cu_3-xP.^25,27,34

Figure 1.

(a) XRD pattern of Cu₃P NPs from 15 min reaction at 320 °C with corresponding crystal structure model (Cu: blue; P: pink) and PDF 01–071–2261 reference pattern (P6₃cm), below. (b) TEM image of the Cu₃P NPs, (c) HRTEM image with outlined region used for fast Fourier transform (FFT) analysis, and (d) FFT pattern indexed to the [0001] zone axis.

Effect of Modulating Reaction Conditions on Cu₃P Formation

Previous reports on the synthesis of Cu₃P NPs suggested two primary mechanisms of NP formation: (1) generation of metallic Cu NPs followed by phosphidation or (2) direct, homogeneous nucleation of small Cu₃P nuclei.²³⁻²⁷ These synthetic methods employed copper chloride as the Cu precursor and trioctylphosphine (TOP), triphenyl phosphite, phosphine gas, or tris(trimethylsilyl)phosphine as the phosphorus source. We sought to understand the mechanism of Cu₃P formation when a single-source precursor that contains preformed Cu–P bonds was used in the absence of an external phosphorus source. To this end, a parametric exploration of the key reaction variables (i.e., time, temperature, and oleylamine concentration) was conducted to understand the transformation of [Cu(H)(PPh₃)]₆ into Cu₃P NPs (Scheme 1).

Scheme 1. Mechanism of Cu₃P Formation As a Function of Reaction Temperature, Time, and Oleylamine Concentration.

Influence of Reaction Temperature and Time

The reaction conditions established for the synthesis of phase pure Cu₃P involved heating [Cu(H)(PPh₃)]₆ in the presence of 15 mmol OAm and 25 mmol ODE at 250 °C for 30 min, followed by heating at 320 °C for 15 min. Although Cu₃P NPs were also formed by directly heating [Cu(H)(PPh₃)]₆ to 320 °C without the 30 min hold at 250 °C, we employed this intermediate heating step to promote decomposition of [Cu(H)(PPh₃)]₆ and homogeneous formation of Cu₃P from the NP precursors.²⁸ XRD analysis of an aliquot of the reaction mixture removed after heating for 30 min at 250 °C indicated the presence of crystalline face-centered cubic Cu metal (Figure S3), revealing that Cu₃P formation proceeded through a Cu intermediate (Scheme 1). TEM analysis of aliquots removed after 30 min at 250 °C and immediately upon reaching 280 °C revealed particles with average size distributions of 5.0 ± 1.4 nm and 4.8 ± 1.3 nm, respectively (Figure S4a,b). The resultant particles were prone to oxidation upon air exposure, as evidenced by conversion of the brown suspension to a green solid. To further understand the conversion of Cu to Cu₃P, reaction mixtures were heated to a final temperature of 300 °C, 310 °C, and 315 °C and held at those temperatures for 30 min. This resulted in the formation of Cu, Cu/Cu₃P, and Cu₃P NPs (Figure S5a), respectively, indicating that increasing the temperature enhanced P-incorporation (Scheme 1).

To explore how phosphorus incorporation changed with reaction time, the reaction mixture was heated at 300 °C with reaction times varied from 5 min to 2 h (Figure 2a). Cu NPs were readily formed at 300 °C after only 5 min with an average size distribution of 6.3 ± 1.2 nm (Figure S4c); however, the resultant particles rapidly turned green upon air exposure, indicating the formation of oxidized Cu species. As the reaction time was increased to 1 h, a broad peak in the XRD pattern between 45° and 50° emerged that is assigned to the formation of Cu₃P (Figure 2a), and the particle size determined by TEM analysis increased to 7.2 ± 1.6 nm (Figure S 4d). The major Cu peak at 43° continued to diminish over 2 h as Cu was converted to Cu₃P (Scheme 1) and this behavior was also observed at 310 °C (Figure S5b). These experiments reveal that phosphorus incorporation increases with reaction time and temperature and the transformation of [Cu(H)(PPh₃)]₆ to Cu₃P proceeds through a Cu intermediate. It has been established that Ni NPs synthesized in the presence of alkylphosphines are phosphorus doped.^35,36 Due to this literature precedent, the Cu intermediate generated here could contain phosphorus. Although we do not have conclusive evidence of phosphorus doping, the major XRD peak for the Cu synthesized from [Cu(H)(PPh₃)]₆ is shifted to higher 2θ (ca. 0.11 degrees) in comparison to the Cu reference, which could suggest a lattice contraction upon introduction of the smaller P into the Cu structure (Figure S6). Further work is needed to assess the viability of a P-doped Cu intermediate.

Figure 2.

(a) XRD patterns after 5 min, 15 min, 30 min, 1 h, and 2 h reaction at 300 °C with an oleylamine concentration of 15 mmol. (b) XRD patterns of reaction products formed after 15 min at 320 °C with varying concentrations of OAm. Reference patterns for Cu and Cu₃P are shown below, and the dotted lines on the experimental patterns indicate the highest intensity peak for Cu and Cu₃P.

Influence of Oleylamine Concentration

The impact of OAm concentration on Cu₃P formation was investigated by heating the reaction mixture at 320 °C for 15 min and varying the concentration of OAm while maintaining a constant reaction volume (12.9 mL total with balance of ODE) (Figure 2b). In the presence of only OAm or ODE, [Cu(H)(PPh₃)]₆ decomposed to metallic Cu, which precipitated rapidly from solution. Varying the OAm concentration from 5–20 mmol revealed that a balanced mixture of OAm (15–20 mmol) and ODE (20–25 mmol) was necessary to promote phosphidation and formation of phase-pure Cu₃P (Figure 2b). TEM image analysis revealed the average particle size distribution increased from 8.7 ± 1.6 nm to 9.3 ± 3.3 nm (Figure S7) when the amount of OAm was increased from 15 to 20 mmol OAm. When 5 and 10 mmol of OAm was used, a mixture of Cu/Cu₃P and Cu₃P with trace Cu was produced, respectively (Figure 2b, Scheme 1). We also sought to expand our understanding of the role of OAm concentration on the rate of phosphorus incorporation at lower temperatures. To this end, 1–2 h reactions at 300 °C using 15 and 20 mmol of OAm were performed and the ratio of Cu to Cu₃P was evaluated. At the higher OAm concentration, significant Cu metal was observed even after 2 h (Figure S8), indicating that high OAm concentrations reduce the phosphidation rate. Investigation into the formation of nickel phosphide NPs has also identified excess OAm as hindering phosphorus incorporation and favoring metal-rich phases.^37,38

Influence of Phosphorus Concentration

Through modulation of synthetic parameters including temperature, time, and OAm concentration, it was determined that Cu₃P formation proceeded through a metallic copper intermediate that underwent phosphidation in the presence of sufficient OAm (15–20 mmol) at temperatures at or above 300 °C (Scheme 1). During modification of the reaction parameters, the phosphorus concentration was kept constant as the precursor, [Cu(H)(PPh₃)]₆, was the only source of phosphorus. Our previous report on the solution synthesis of metal phosphide NPs using single-source precursors indicated that increasing the P:M ratio through addition of excess PPh₃ facilitated phosphidation.²⁸ In a previous report of Cu₃P NP synthesis, increasing the P:Cu ratio changed the mechanism of formation from phosphidation of a Cu intermediate to direct growth of Cu₃P.²⁴ To explore the effect of P:Cu ratio on Cu₃P formation from [Cu(H)(PPh₃)]₆, the same reaction procedure was used, as previously discussed, but with the addition of 2 to 4 equiv of PPh₃.

At 320 °C, the addition of 2 equiv of PPh₃ resulted in Cu₃P after 15 and 30 min reaction times (Figure S9) with average size distributions from TEM analysis of 8.6 ± 2.3 nm and 10.6 ± 2.3 nm, respectively (Figure S10). To differentiate between Cu₃P synthesized with and without PPh₃, P–Cu₃P will be used to denote the NPs generated with added phosphorus. In a comparison of Cu₃P NPs synthesized after heating for 15 min at 320 °C with and without PPh₃, the addition of 2 equiv of PPh₃ did not result in particle growth. Both conditions generated NPs with similar average size distributions of 8.7 ± 1.6 nm for Cu₃P (Figure S1a) compared to 8.6 ± 2.3 nm for P–Cu₃P. Elemental analysis revealed increased phosphorus concentrations at higher P:Cu ratios. The NPs synthesized with and without 2 equiv PPh₃ contained 27 mol % P and 25 mol % P, respectively. The elemental analysis only provides the overall composition of Cu_3-xP, and the P-concentration could contain contributions from unincorporated precursor decomposition products or surface-bound PPh₃. Additionally, Cu vacancies, which will lead to a higher P-content than expected, are common for Cu_3-xP, and the plasmonic behavior of Cu_3-xP observed here and in previous work has been attributed to the presence of Cu vacancies (Figure S2).

To assess if Cu₃P forms directly rather than through a metallic Cu NP intermediate at lower temperatures with increased P:Cu ratio, 4 equiv of PPh₃ were added to the reaction and the mixture was held at 300 °C for 1–2 h (Figure S11). After 1 h, Cu₃P was the major product; however, Cu was also observed, indicating that even in the presence of excess phosphine, Cu₃P formation still proceeded through a Cu intermediate under these reaction conditions. After 2 h at 300 °C, XRD analysis indicated the formation of phase-pure Cu₃P; however, the resultant product was a mixture of polydisperse aggregates analyzed by TEM (Figure S11b) and nondispersible material. Although Cu₃P was accessed at the lower temperature of 300 °C with higher P:Cu ratios, the particles displayed significant aggregation and limited dispersibility. Poor dispersibility also resulted from reaction times greater than 1 h under most of the conditions utilized here. Understanding the influence of reaction parameters on phase-pure Cu₃P formation, homogeneous particle morphology, and particle dispersibility, led to the selection of conditions of 30 and 15 min holds at 250 and 320 °C, respectively, in the presence of 15 mmol OAm with or without 2 equiv of PPh₃ to generate Cu₃P NPs for electrocatalytic studies.

Electrocatalytic CO₂ Reduction

FTIR spectroscopy of as-synthesized Cu₃P particles (Figure S12a) revealed the distinct bands in the range of 3000–2800 cm^–1 that can be assigned to C–H stretches that indicate the presence of residual organic species (e.g., OAm). Surface ligands, such as oleylamine, have been utilized to direct and improve CO₂ reduction selectivity;³⁹ however, previous electrochemical analysis on metal phosphide NPs have indicated that the presence of surface ligands reduces the electrochemically accessible surface area, thus inhibiting the catalytic activity.^40,41 For electrochemical analysis, electrodes were prepared by drop-casting as-synthesized Cu₃P NPs suspended in CHCl₃ onto carbon paper (CP) to form OAm–Cu₃P/CP. In order to evaluate the impact of surface bound oleylamine on the electrochemical performance, the OAm–Cu₃P/CP was reduced at 450 °C in the presence of H₂, a method that has been established to remove organic surface ligands from metal phosphide NPs.^28,42 Retention of the Cu₃P hexagonal phase following thermal treatment at 450 °C in the presence of H₂ was confirmed by XRD analysis (Figure S13). Following thermal reduction, the reduced Cu₃P-modified CP (R–Cu₃P/CP) was passivated with 1% O₂ to protect the surface from rapid oxidation upon air exposure.

The Cu₃P/CP electrodes were electrochemically investigated in CO₂-saturated 0.1 M KHCO₃ aqueous electrolyte. To evaluate the influence of surface bound OAm on electrocatalytic CO₂ reduction activity, electrochemical analysis was performed on both OAm–Cu₃P/CP and R–Cu₃P/CP. Evaluation of the capacitance current at the open circuit potential, which is directly related to the electrochemically accessible surface area, revealed increased capacitance for R–Cu₃P/CP (Figure S14). Cyclic voltammograms indicated that R–Cu₃P/CP had a more positive catalytic onset than OAm–Cu₃P/CP (Figure 3a). The increased surface area and more positive onset potential following thermal reduction to remove surface ligands has been previously observed for metal phosphide NPs.^11,43

Figure 3.

(a) Polarization curves of OAm–Cu₃P/CP, R–Cu₃P/CP, and carbon paper in CO₂-saturated 0.1 M KHCO₃ at a scan rate of 50 mV/s. (b) Chronoamperometry measurements of R–Cu₃P/CP at different applied potentials (−0.25 to −0.50 V versus RHE) in CO₂-saturated 0.1 M KHCO₃. (c) Partial current densities for HCOO– (inset) and H₂ for R–Cu₃P/CP in CO₂-saturated 0.1 M KHCO₃.

The large current increase between −0.30 and −0.40 V versus RHE in the cyclic voltammograms is primarily attributed to electrocatalytic H₂ evolution (Figure 3a). Electrolysis with OAm–Cu₃P/CP and R–Cu₃P/CP was performed for 3 h, at potentials ranging from −0.25 to −0.50 V versus RHE, to quantify gaseous and liquid products (Figure 3b, S15). At all potentials analyzed, H₂ was the only gaseous product as well as the major product observed. At potentials more negative than −0.45 V versus RHE, H₂ was generated with FE > 80%. Because of the dominance of H₂ evolution at applied potentials more negative than −0.45 V, subsequent electrolysis experiments focused on potentials between −0.25 to −0.45 V versus RHE to assess if CO₂ reduction could occur at potentials where the kinetics of H₂ generation are diminished.

Electrolysis performed between −0.25 to −0.45 V (Figure 3b, S15) revealed production of formate as the only C-containing product with maximum FE of 8% and 6% reached at −0.30 V and −0.35 V for R–Cu₃P/CP and OAm–Cu₃P/CP, respectively (Tables 1, S1). Utilization of unmodified CP as an electrode for CO₂ reduction under the same conditions resulted in minimal current generation and no formate production. The FE for H₂ did not exceed 50% within the analyzed potential window and the rest of the current is tentatively attributed to charging of the interface, a nonfaradaic process,^44,45 and reduction of the catalyst, the necessity of which is discussed in more detail below. Faradaic efficiencies closer to unity were obtained at applied potentials more negative than −0.45 V versus RHE, where the kinetics of H₂ evolution dominated and formate production was diminished. Analysis of the partial current densities (Figure 3c) reveal formate production accounts for a minimal amount of the generated current during electrolysis especially at more negative potentials where the FE for H₂ reaches unity. Nickel phosphide electrocatalysts have exhibited analogous CO₂ reduction behavior, where CO₂ reduction occurred at less negative applied potentials, and H₂ production occurred at more negative potentials.¹⁶ Although CO₂ reduction has been observed within a narrow potential window over nickel phosphide, silver phosphide, and copper phosphide catalysts, H₂ evolution, which is thermodynamically and kinetically favored, is never fully suppressed. The exclusive reduction of CO₂ to formate demonstrated here has also been observed over sulfur-modified copper. Sulfur-modified copper developed by Pérez-Ramírez et al. generated formate with FE of 80% at −0.80 V versus RHE in CO₂-saturated 0.1 M KHCO₃.⁴⁶ The inclusion of sulfur modulated the adsorption of key intermediates thus switching the selectivity from the C–C coupled products typically produced over copper to formate.⁴⁶⁻⁵¹ Following this concept, we propose a similar affect here, where the presence of P in the Cu₃P/CP electrodes changes the intermediates versus metallic Cu, leading to formate production rather than C–C coupling.

Table 1. Faradaic Efficiency (FE) of Formate Production for R–Cu₃P/CP in CO₂-Saturated 0.1 M KHCO₃.

Potential (V) versus RHE	Formate FE (%)
–0.25	0.0
–0.30	8.0
–0.35	4.0
–0.40	4.0
–0.45	3.3
–0.50	1.9

The effect of increasing the potassium bicarbonate concentration from 0.1 to 0.5 M on the efficiency of formate generation was investigated for R–Cu₃P/CP. As has been reported previously,⁵² the onset of catalysis was shifted to less negative potentials and larger current densities were generated as the electrolyte conductivity increased due to the higher bicarbonate concentration (Figure S16). Formate production began at −0.20 V versus RHE in 0.5 M KHCO₃ in comparison to −0.30 V versus RHE in 0.1 M KHCO₃; however, the maximum FE remained at 8% (Table S3). The reduction in overpotential for formate production in 0.5 M KHCO₃ also corresponded to a reduction in overpotential for H₂ production, resulting in increased FE for H₂ at less negative applied potentials. The impact of phosphorus on the CO₂ reduction activity was also explored. Cu₃P NPs synthesized in the presence of 2 equiv of PPh₃ are denoted as follows: OAm–P–Cu₃P (as-synthesized) and R–P–Cu₃P (reduced). OAm–P–Cu₃P and R–P–Cu₃P electrochemically reduced CO₂ to formate at comparable FE as the Cu₃P NPs synthesized without PPh₃ (Figure S17).

Previous reports have identified that NPs commonly undergo transformations during electrocatalysis;⁵³ therefore, it was necessary to evaluate the stability of Cu₃P NPs during electroreduction in the aqueous conditions utilized here. For OAm–Cu₃P/CP, a gradual increase in current density over time was observed (Figure S15) for electrolysis experiments performed from −0.25 to −0.50 V versus RHE in CO₂-saturated 0.1 M KHCO₃. To assess if the current instability was related to the surface OAm ligands, OAm–Cu₃P NPs were drop-cast on FTO electrodes and FTIR spectroscopy was performed on the electrodes before and after 3 h electrolysis experiments at −0.40 V and −0.50 V versus RHE. Analysis of the FTIR spectra (Figure S12b) indicated that the C–H stretches associated with OAm persisted. However, the utilization of the transparent FTO electrode allowed for visual observation of electrode fouling (green discoloration) upon air exposure following the electrolysis experiments. The OAm–Cu₃P/CP electrode was collected following 3 h electrolysis at −0.45 V versus RHE, and XRD analysis of the electrode was performed in air. The XRD pattern (Figure S18) showed diminished peaks associated with Cu₃P revealing the electrochemical instability of this material. The peaks at 36° and 42.5° indicate the formation of copper oxide upon air exposure of the electrochemically polarized OAm–Cu₃P/CP, which is consistent with the observed electrode discoloration. Therefore, the increase in current density observed during electrolysis of OAm–Cu₃P/CP is due to the reductive transformation of the catalyst. The origin of the reduction-induced structural evolution of OAm–Cu₃P is further discussed below. In contrast to the OAm–Cu₃P/CP electrodes, which demonstrated an increase in current density over time, electrolysis with R–Cu₃P/CP revealed a reduction in the generated current over 3 h (Figure 3b) and 6 h (Figure S19a) and a lower FE for formate after 6 h. The reduction in current and FE could be attributed to deactivation of R–Cu₃P/CP over time. XRD of the R–Cu₃P/CP electrodes performed following 3 and 6 h of electrolysis showed retention of the bulk structure with no evidence of crystalline copper oxides (Figure 4a, S19b). However, this analysis does not preclude the presence of other deactivation pathways such as transformation of the catalyst morphology and particle size or formation of amorphous degradation species. The impact of phosphorus on the electrochemical stability was also investigated. The XRD patterns of OAm–P–Cu₃P/CP prepared from Cu₃P NPs synthesized with 2 equiv of PPh₃, following 3 h of electrolysis at −0.45 V versus RHE (Figure S18), also did not show crystalline copper oxide formation.

Figure 4.

(a) XRD patterns of R–Cu₃P/CP following 3 h electrolysis at potentials ranging from −0.25 to −0.50 V versus RHE in CO₂-saturated 0.1 M KHCO₃. (b) Normalized Cu K edge XANES for P-Cu₃P/C before and after reduction at 450 °C in H₂ and subsequent room temperature passivation in O₂. (c) k²-weighted Cu K edge EXAFS of P-Cu₃P/C.

X-ray absorption spectroscopy (XAS) was performed to determine if differences in the geometric and electronic properties of the OAm–Cu₃P, R–Cu₃P, P–Cu₃P NPs could explain the improved electrochemical stability for R–Cu₃P and P–Cu₃P. XAS was performed at the Cu K edge on carbon black-supported Cu₃P NPs (Cu₃P/C). The crystal structure of Cu₃P contains four distinct Cu sites with different local arrangements of atoms, and XAS analysis provides an average of the Cu environments in the structure. Cu₃P was analyzed in the following states: as-synthesized, thermally reduced at 450 °C in the presence of H₂, and subsequently passivated with 1% O_2. The instability observed electrochemically for OAm–Cu₃P extended to the XAS experiments. OAm–Cu₃P/C consistently degraded despite attempts to minimize air exposure during preparation and collection of the XAS data. The accelerated degradation could be attributed to the higher surface area carbon (Vulcan XC 72R) used for XAS analysis in comparison to the carbon paper employed for the electrochemistry experiments. Due to the inability to limit degradation of OAm–Cu₃P and the similar electrocatalytic performance, only the OAm–P–Cu₃P NPs were utilized for XAS experiments. Figure 4b shows the Cu K edge X-ray absorption near-edge structure (XANES) spectra of Cu₃P NPs and Cu foil. The Cu₃P NPs had a higher XANES energy (8.9805 keV) compared to the Cu foil (8.979 keV). The white line intensity of the as-synthesized Cu₃P was higher than the Cu foil followed by dampened extended X-ray absorption fine structure (EXAFS) oscillations due to scattering from a lighter element (i.e., O versus Cu), likely due to the presence of small amounts of oxidized Cu. The K edge XANES represents an electron transition of the 1s to 3p electron and is sensitive to small changes in oxidation state. Following thermal reduction in H₂, the white line intensity decreased relative to the as-synthesized Cu₃P and the shape of the XANES spectrum broadened slightly. These changes correspond to a decreased energy of the unfilled states of Cu and can be due to both the loss of Cu oxides and the incorporation of P. The XANES energy increased slightly to 8.9806 keV for the H₂ treated Cu₃P NPs, likely because Cu sites bear a partial positive charge due to the electron density shift from the metal to the more electronegative phosphorus sites.¹³ No change in the XANES spectrum was observed following treatment of the thermally reduced Cu₃P with 1% O₂, indicating retention of the metallic character after surface passivation.

The EXAFS spectra were fit to determine the local coordination of Cu (Figure 4c, Figure S20). Because XAS is a bulk-averaged technique, the EXAFS fits are determined for an average Cu atom. In a pure phase Cu₃P structure, there are four distinct Cu sites that contain an average Cu–P coordination number of 2.7 at 2.37 Å and average Cu–Cu coordination number of 3 at 2.68 Å (CN_Cu–P/CN_Cu–Cu = 0.9). The OAm–P–Cu₃P NPs were fit with Cu–O, in addition to Cu–P and Cu–Cu scattering pairs. The optimized EXAFS fit contained 1.6 Cu–O bonds at 1.92 Å, in addition to 1.5 Cu–P and 1.2 Cu–Cu bonds at 2.34 and 2.65 Å, respectively. The addition of P to form Cu₃P elongates the Cu–Cu bond (2.65 Å) compared to what is expected in Cu NPs (2.51 Å).⁵⁴ Here, the total coordination number was 4.3. This suggests that approximately 37% (1.6/4.3) of the nanoparticle was oxidized prior to reduction, though Cu₃P was present. However, the absence of Cu-oxide reflections in the XRD pattern suggests the oxide species present are amorphous, perhaps on the nanoparticle surface, as opposed to crystalline, bulk copper oxide. After reduction, the Cu–O bonds were lost to the increased coordination of Cu–Cu and Cu–P with only Cu–Cu and Cu–P scattering pairs observed. The results suggest that prior to reduction, the Cu₃P NPs were partially oxidized. The R–P–Cu₃P NPs contained approximately 2.5 Cu–P bonds at 2.34 Å and 2.0 Cu–Cu bonds at 2.65 Å. The CN_Cu–P/CN_Cu–Cu = 1.3 is higher than the expected CN_Cu–P/CN_Cu–Cu = 0.9 for the ideal Cu₃P structure, thus R–P–Cu₃P NPs may contain Cu vacancies and/or excess or residual phosphorus species, which could result in a P-rich material (Table S4).⁵⁵ Additional characterization is needed to determine the exact chemistries (Cu vacancies versus excess P-species) that lead to the diminished Cu–Cu contributions. There is no change to the EXAFS data following passivation of the reduced Cu₃P NPs with 1% O₂. Although bulk Cu–O contributions were not observed after passivation, XAS is not an appropriate technique to evaluate surface oxidation and thus, due to the air-sensitive nature of Cu₃P, the presence of surface oxides following exposure to 1% O₂, while not directly observed, is likely.

The XAS results provide some insight into the influence of thermal reduction and addition of PPh₃ to the synthetic procedure on the stability of Cu₃P NPs. The inability to collect XAS data on OAm–Cu₃P confirms the material is unstable prior to electrochemical testing and degradation is accelerated during electrolysis. In comparison to OAm–Cu₃P, the addition of 2 equiv of PPh₃ during the synthesis to generate P–Cu₃P helps promote phosphidation and possibly generate more robust and oxidatively stable NPs. Without additional PPh₃, OAm–Cu₃P is unstable, and although we were unable to collect XAS directly on OAm–Cu₃P, we suggest this instability is due to a greater number of Cu–O bonds in the structure of OAm–Cu₃P than OAm–P–Cu₃P. We hypothesize that the addition of PPh₃ during the synthesis facilitates isolation of Cu₃P NPs more closely resembling the ideal structure and with a reduced number of Cu–O scattering pairs. Therefore, the contribution of Cu–O to the Cu₃P structure is a key determinant of stability.

The extensive Cu–O bonds in OAm–P–Cu₃P, revealed by XAS analysis and that we hypothesize are present in OAm–Cu₃P, may provide an unstable catalytic surface under the electrochemical reduction conditions utilized. It is established that copper oxide is quickly reduced to metallic copper during CO₂ electroreduction in aqueous media.^56,57 Although XRD of OAm–P–Cu₃P/CP following electrolysis did not show the same crystalline copper oxide formation as OAm–Cu₃P/CP, the presence of Cu–O bonds in OAm–P–Cu₃P as evidenced by EXAFS necessitates more extensive investigation into the material’s stability during electroreduction. XAS of Cu₃P NPs reduced in H₂ at 450 °C revealed the loss of the Cu–O bonds and an increase in the number of Cu–Cu and Cu–P bonds (Figure 4c, Table S1). Reductive pretreatment removes the unstable Cu–O contributions thus providing a more stable material for electrocatalysis. In-situ characterization is being pursued to provide more detailed insight into the impact of surface structure (as-synthesized versus reduced) and phosphorus stoichiometry on the electrocatalytic stability and performance of Cu₃P NPs.

In addition to understanding the electrochemical stability of Cu₃P NPs to facilitate the design of copper phosphide catalysts with improved electrocatalytic CO₂ reduction activity, identification of the mechanism of formate production over Cu₃P is also important. Because Cu₃P can produce formate at small overpotentials, albeit at low FE, direct reduction of CO₂ through a one-electron transfer to generate the radical anion or via a proton-coupled electron transfer (PCET) to produce adsorbed formate on the electrocatalyst surface is an unlikely pathway.⁵⁸ Rather, the mechanism could proceed via an electrohydrogenation process whereby a proton and electron transfer occur simultaneously, coupling a surface hydride with CO₂ in solution to generate formate. This electrohydrogenation pathway for the generation of formate from CO₂ has been proposed for nickel phosphide catalysts. Nickel phosphide catalysts have been shown to produce C₃ and C₄ oxygenates through generation of formate via electrohydrogenation of CO₂.¹⁶ It should be noted that CO production was not observed for the previously reported nickel phosphide catalysts or the copper phosphide catalysts evaluated here. CO has been identified as the key CO₂ reduction intermediate to generate C₂₊ products,^20,22 however, the proposed mechanism of CO₂ reduction to C₃ and C₄ oxygenates over nickel phosphides does not rely on the formation of CO to facilitate C–C bond coupling. The presence of both nickel and phosphorus surface sites is proposed to preferentially stabilize oxygen-bound intermediates (HCOO*) versus the traditional carbon-bound intermediates (COOH*) on metals during CO₂ reduction, which facilitates the production of the unique oxygenated molecules, methylglyoxal and 2,3-furandiol. Rigorous experimental and computational exploration into the mechanism of formate production over Cu₃P is necessary to determine the viability of the electrohydrogenation pathway. Such insights have the potential to guide the synthesis of Cu phosphide-based materials with tailored composition or structure to control selectivity and promote higher FE to formate, and possibly carbon-coupled products.

Conclusions

A facile molecular precursor route has been established for the preparation of solid, phase-pure Cu₃P NPs. The decomposition of [Cu(H)(PPh₃)]₆, a commercially available precursor, at high temperature in a hydrocarbon/amine solvent mixture results in the generation of Cu₃P NPs in a single-pot reaction. Evaluation of the mechanism of Cu₃P formation revealed that the reaction proceeds via a Cu intermediate that undergoes increased rates of phosphidation at higher temperatures, longer reaction times, and in the presence of excess PPh₃. Cu₃P NPs prepared via the molecular precursor route were evaluated as electrocatalysts for the reduction of CO₂ in CO₂-saturated KHCO₃ aqueous solutions. The highly dispersible nature of the Cu₃P NPs allowed for facile deposition of the NPs from organic solvent onto carbon paper electrodes. The electrochemical behavior was investigated for the as-synthesized Cu₃P that retained OAm surface ligands and Cu₃P that was reduced in 5% H₂ to remove the surface ligands. Unstable current response and electrode fouling during electrolysis was observed for the as-synthesized Cu₃P, highlighting the instability of the material under the electrocatalytic conditions tested. The electrochemical stability of as-synthesized Cu₃P NPs was improved with increasing phosphorus content. In contrast, the reduced Cu₃P was markedly more stable than the as-synthesized material, and further electrochemical investigation revealed the production of formate beginning at −0.30 V versus RHE in CO₂-saturated 0.1 M KHCO₃ with a maximum FE of 8%.

Metal phosphides have recently been demonstrated to be a promising platform for electrochemical CO₂ reduction, and this application of Cu₃P NPs for the reduction of CO₂ to formate builds on this pioneering work. Specifically, the utilization of nickel phosphides for the conversion of CO₂ to C₃ and C₄ oxygenates has prompted a reevaluation of the application of this highly tunable and synthetically versatile materials class in the field of CO₂ reduction. Continued investigations into the mechanism of formate production and the transformation of the catalyst surface during electrolysis are underway to compare the reaction pathways over Cu phosphides and Ni phosphides, and to determine the factors that are vital for facilitating CO₂ conversion, such as phosphorus stoichiometry. These insights can assist in the design of advanced catalyst architectures, and the molecular precursor method we have developed for a wide-array of metal phosphide NPs provides a unique and advantageous pathway for the preparation of well-defined and highly controlled NPs to meet these designs. The utilization of the solution-based, molecular precursor route to highly processable Cu₃P provides opportunities for tailoring the composition, morphology, and surface chemistry of Cu₃P to achieve higher CO₂ conversion efficiencies and potentially access more reduced and economically valuable oxygenates.

Glossary

Abbreviations

NPs

nanoparticles

CP

carbon paper

RHE

reversible hydrogen electrode

FE

Faradaic efficiency

NMR

nuclear magnetic resonance spectroscopy

HPLC

high-performance liquid chromatography

TEM

transmission electron microscopy

FTIR

Fourier transform infrared spectroscopy

XAS

X-ray absorption spectroscopy

XANES

X-ray absorption near edge structure

EXAFS

extended X-ray absorption fine structure

PXRD

powder X-ray diffraction.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsaem.0c01360.

• TEM images of Cu and Cu₃P NPs, XRD patterns of Cu and Cu₃P NPs formed by modulating reaction parameters, XRD patterns of Cu₃P before and after electrolysis, Faradaic efficiency data, chronoamperometry data for OAm–Cu₃P/CP, chronoamperometry data for R–Cu₃P/CP in 0.5 M KHCO₃, FTIR of Cu₃P NPs before and after electrolysis, and capacitance data (PDF)

The authors declare no competing financial interest.