Competitive hydrogen evolution and multiple proton-coupled electron transfer reactions limit photoelectrochemical CO2 reduction in aqueous electrolyte. Here, oxygen-terminated lightly boron-doped diamond (BDDL) thin films were synthesized as a semiconductor electron source to accelerate CO2 reduction. However, BDDL alone could not stabilize the intermediates of CO2 reduction, yielding a negligible amount of reduction products. Silver nanoparticles were then deposited on BDDL because of their selective electrochemical CO2 reduction ability. Excellent selectivity (estimated CO:H2 mass ratio of 318:1) and recyclability (stable for five cycles of 3 h each) for photoelectrochemical CO2 reduction were obtained for the optimum silver nanoparticle-modified BDDL electrode at −1.1 V vs. RHE under 222-nm irradiation. The high efficiency and stability of this catalyst are ascribed to the in situ photoactivation of the BDDL surface during the photoelectrochemical reaction. The present work reveals the potential of BDDL as a high-energy electron source for use with co-catalysts in photochemical conversion.
Artificial photosynthesis is an emerging process for sustainable conversion of solar energy to chemical energy, and is important because of the increasing atmospheric CO2 concentration and global energy demand1,2,3. Artificial photosynthesis generally uses a semiconductor that absorbs solar radiation and generates electron/hole pairs1,4,5. These electron/hole pairs produce active radical species on the semiconductor surface prior to the adsorption of water or reacting molecules, or electron/hole pairs directly transfer to guest molecules through space6,7. The high energy of these active radicals and electron/hole pairs means they readily take part in certain chemical conversions under suitable conditions. For example, TiO2 is a well-known photocatalyst for water splitting and pollutant degradation6,7. TiO2 splits water into hydrogen and/or oxygen in the presence of a Pt counter electrode or hole scavenger using a photon energy greater than its bandgap. Other chemical conversions have also been achieved through semiconductor/molecular photocatalysis8,9,10,11. However, in multistep chemical conversion, which involve many intermediates to obtain a desired product, each intermediate has the possibility to divert the reaction course, resulting in poor product selectivity12,13,14. Although product selectivity in multistep reactions can be increased by choosing a suitable photocatalyst or controlling the reaction conditions, the reliability and efficiency of such reactions are still far from those required for commercial use13. Therefore, efficient conversion through multistep or multi-intermediate photochemical reaction remains challenging. Chemists typically attempt to divert such conversions into new mechanistic pathways with the fewest possible intermediates.
Photocatalytic or photoelectrochemical CO2 reduction in aqueous solution is one example where product selectivity is limited by both thermodynamics and kinetics. Commercial conversion of CO2 into high-value chemicals is generally inefficient, laborious, and expensive12,13,14. Efficient photocatalytic or photoelectrochemical CO2 reduction with excellent product selectivity is a research goal in the field of sustainable energy conversion, and represents an alternative to natural photosynthesis.
The atmospheric concentration of CO2 has increased tremendously over the last 50 years. Selective, efficient conversion of CO2 into a valuable chemical fuel is not only important for sustainability but also economically favourable15,16,17. Natural photosynthesis is an indispensable tool to capture CO2, but gross destruction of green plants and rapid civilization have forced researchers to search for alternatives to natural photosynthesis to efficiently and selectively convert CO2 into high-value chemicals. The selectivity of chlorophyll in natural photosynthesis is unprecedented. In artificial photocatalytic or photoelectrochemical CO2 reduction, conversion of CO2 into useful chemicals requires suitable catalysts2,4. Photocatalytic or photoelectrochemical CO2 reduction generally involves multiple proton-coupled electron transfer reactions in aqueous media. These reactions provide numerous intermediates, and thereby produce multiple CO2 reduction products4. Electrochemical or photoelectrochemical CO2 reduction has been explored using various catalysts with the aim of identifying mechanistic pathways that finally could lead to efficient conversion of CO2 into a particular product12,18,19,20,21,22.
Halmann et al.22. first reported photoelectrochemical CO2 reduction to produce formic acid, methanol and formaldehyde using a p-type GaP semiconductor in 1978. Since then, many semiconductor photocatalysts have been used for CO2 reduction23,24,25,26. Metal nanoparticles deposition on the semiconductors and fabrication of metal-semiconductor junction have successfully studied for improvement of photocatalytic CO2 reduction activity27,28,29. Recently, Hamers and co-workers found that hydrogen-terminated boron-doped diamond (H-BDD) can produce solvated electrons through injection of photoexcited electrons into aqueous electrolytes30. The solvated electrons produced by H-BDD can reduce CO2 into carbon monoxide (CO) alone via a one-electron transfer reduction mechanism under sufficiently high pressure. However, the extremely high pressure used in this approach is undesirable. In addition, H-BDD is easily oxidized under illumination30,31.
In this work, we study an alternative solution to overcome the challenges of one-electron transfer under normal pressure and temperature by using oxygen-terminated lightly boron-doped (1,000 ppm) diamond (BDDL) incorporated with co-catalysts. The surface of oxygen-terminated BDDL is reduced during photoelectrochemical reaction so it acts as a semiconducting high-energy electron source30,31. Silver (Ag) co-catalyst deposited on BDDL stabilizes the intermediates of CO2 reduction reactions under normal pressure to realize selective photoelectrochemical reduction. The semiconducting nature of BDDL is investigated by changing the light sources used in photoelectrochemical studies.
Surface morphology and photoelectrochemical study
Field-emission scanning electron microscopy (FESEM) and elemental analysis were used to investigate the deposition of Ag nanoparticles on BDDL. First, Ag nanoparticles were deposited on BDDL by a chronoamperometric method in 0.1 M AgNO3 at −0.5 V for 60 s to produce a sample denoted 0.1 Ag-BDDL. Low-magnification FESEM analysis of this sample revealed that Ag nanoparticles with a diameter of ~300 nm were deposited on BDDL (Fig. 1a). However, a high-magnification FESEM image indicated that primarily smaller Ag nanoparticles (average size ~20 nm) were deposited on BDDL (Fig. 1b). The size distributions of smaller Ag nanoparticles are shown in Supplementary Fig. S1 by analysing several high magnified FESEM images. Average size of the smaller Ag nanoparticles was found to be around ~20 nm (Supplementary Fig. S1) with the 0.1 Ag-BDDL. The 0.1 Ag-BDDL sample was then used for electrochemical/photoelectrochemical CO2 reduction. Figure 1c displays cyclic voltammograms (CVs) of a 0.1 Ag-BDDL electrode in 25 mM Na2SO4. In N2, no cathodic peak was obtained up to −1.0 V vs. RHE, suggesting no reduction of Ag-BDDL or the electrolyte. The low cathodic current density (onset potential −1.0 V vs. RHE) may indicate a poor hydrogen evolution reaction32. Indeed, under CO2-saturated conditions, 0.1 Ag-BDDL displayed a strong cathodic peak at −1.1 V vs. RHE in 25 mM Na2SO4, clearly indicating the cathodic electrochemical reduction of CO2. The cathodic current density gradually decreased with increasing cycle number (Supplementary Fig. S2a). Careful observation revealed that the difference of cathodic current density between two consecutive cycles decreased as cycling progressed. After several cycles, the cathodic current density remained almost constant at −1.1 V (Supplementary Fig. S2a). The opposite behaviour was observed under irradiation at 222 nm with an excimer lamp (7 W). Under irradiation, the cathodic current density increased slowly and reached its maximum value after several cycles (Supplementary Fig. S2b). The increase in cathodic current density is attributed to the photoexcitation and photoactivation of 0.1 Ag-BDDL by high-energy photons at −1.1 V vs. RHE. Therefore, under irradiation and applied bias potential, charge transport and cathodic CO2 reduction are facilitated, increasing cathodic current density. The final cycles in both the dark and light-irradiation conditions show quite different cathodic current densities (Fig. 1c). This large difference clearly reveals the excellent photoelectrochemical activity of the 0.1 Ag-BDDL electrode for CO2 reduction.
To further confirm the photoelectrochemical behaviour of the 0.1 Ag-BDDL electrode, partial current densities were measured in both the dark and light in 25 mM Na2SO4 at −1.1 V vs. RHE. In the first few minutes in the dark, the partial current density changed sharply from −0.3 to −0.2 V and then remained constant (Fig. 1d). This clearly agrees with the CV characteristics of Ag-BDDL under CO2-saturated conditions. However, under photoexcitation at 222 nm, the partial current density at −1.1 V began to increase over time (Fig. 1d). This increase in partial current density is caused by the photoexcitation and photoactivation of 0.1 Ag-BDDL and the subsequent participation of photoexcited electrons in the cathodic CO2 reduction reaction on the Ag surface, as revealed by CV analysis.
CV and chronoamperometric analysis clearly suggest the electrochemical and photoelectrochemical reduction of CO2 by the 0.1 Ag-BDDL electrode at −1.1 V vs. RHE in 25 mM Na2SO4. Therefore, the CO2 reduction products obtained under similar conditions at −1.1 V vs. RHE with different amounts of Ag loaded on the BDDL electrodes were detected. Gas chromatography (GC) and ion chromatography (IC) were used to investigate the reduction products. Figure 2 shows the amount of CO and H2 obtained using Ag-BDDL electrodes with different Ag loading during irradiation (222 nm). The concentration of AgNO3 used during deposition was slowly increased from 0 to 0.1 M (0 indicates a bare BDDL electrode, while 0.025 Ag-BDDL, 0.05 Ag-BDDL and 0.1 Ag-BDDL denote deposition of Ag from 0.025, 0.05 and 0.1 M AgNO3 aqueous solutions on BDDL at −0.5 V for 60 s, respectively). The detailed of surface morphology with 0.025, and 0.05 Ag-BDDL is shown in Supplement Fig. S1. Average sizes of the smaller Ag nanoparticles were found to be ~10 and ~15 nm with the samples 0.025, and 0.05 Ag-BDDL electrodes, respectively. The photoelectrochemical CO2 reduction products obtained over these electrodes were analysed (Fig. 2a). The bare electrode (0 Ag-BDDL) produced a negligible amount of CO. The amount of CO produced increased markedly with Ag loading on BDDL. Although similar amounts of CO were obtained over 0.1 Ag-BDDL (590 μg in 3 h) and 0.05 Ag-BDDL (577 μg in 3 h), less H2 was formed by 0.1 Ag-BDDL (1.85 μg in 3 h) than 0.05 Ag-BDDL (2.59 μg in 3 h). Minor amount of H2 could be produced from competitive proton reduction in aqueous electrolyte. The weight ratios of CO to H2 were 29:1, 222:1 and 318:1 over 0.025 Ag-BDDL, 0.05 Ag-BDDL and 0.1 Ag-BDDL, respectively. Therefore, the 0.1 Ag-BDDL electrode produces CO with minimal H2 at −1.1 V vs. RHE under 222 nm irradiation.
The role of photoelectrochemical CO2 reduction at 0.1 Ag-BDDL is emphasized by comparing the products obtained with and without light irradiation and under N2-saturated conditions at −1.1 V. GC analysis reveals that in the dark, the amount of CO produced (88 μg in 3 h) decreased considerably compared with that under irradiation, while the amount of H2 (4.35 μg in 3 h) remained similar (Fig. 2c,d). Similarly, under N2-saturated conditions, a negligible amount of CO (15 μg in 3 h) was detected, while the amount of H2 remained similar. Therefore, the greater amount of CO produced under CO2-saturated conditions and 222 nm irradiation results from the photoelectrochemical effect of 0.1 Ag-BDDL.
Ag-BDDL could also produce liquid products because CO2 reduction often follows a multiple proton-coupled electron transfer reaction mechanism in aqueous electrolyte, providing numerous products. Therefore, IC was used to detect possible liquid products including formate, acetate and oxalate. No liquid products were detected from the Ag-BDDL electrodes under CO2-saturated conditions in the light (222 nm) or dark at −1.1 V.
Selectivity of different Ag-modified carbon electrodes
We found that 0.1 Ag-BDDL displayed more efficient and selective photoelectrochemical CO2 reduction than 0.05 and 0.025 Ag-BDDL electrodes. To confirm the superior photoelectrochemical activity and selectivity of 0.1 Ag-BDDL, Ag nanoparticles were deposited on heavily boron-doped (10,000 ppm) diamond (denoted BDDH) and glassy carbon electrodes (GCEs) at −0.5 V for 60 s in 0.1 M AgNO3. The photoelectrochemical performance of the resulting electrodes was studied. Figure 3 shows the total Faradaic efficiency of these electrodes in 25 mM Na2SO4 at −1.1 V vs. RHE after 3 h of photoelectrolysis under 222 nm irradiation. The total Faradaic efficiency of 0.1 Ag-BDDL was 81.5%. The Faradaic efficiency for CO with this electrode was 72.5%, and only 9.08% for H2, thereby providing very high selectivity for CO. Conversely, the total Faradaic efficiency of 0.1 Ag-BDDH was 75.1%. Although this value is close to that of 0.1 Ag-BDDL, the Faradaic efficiency for CO of the 0.1 Ag-BDDH electrode is only 36.8% and the CO:H ratio is ~1:1 indicating facile proton reduction at 0.1 Ag-BDDH. Meanwhile, 0.1 Ag-GCE exhibited a very poor total Faradaic efficiency (39.8%) with a Faradaic efficiency for CO of just 8.7%. Therefore, Ag-BDDL displays increased conversion and selectivity in the photoelectrochemical reduction of CO2 in 25 mM Na2SO4 at −1.1 V vs. RHE compared with those of 0.1 Ag-BDDH and 0.1 Ag-GCE.
Stability and recyclability of the optimal photocathode
Electrode stability and recyclability are important factors that must be considered for their reliable use in photoelectrochemical solar energy conversion devices. Therefore, we studied the stability and recyclability of the optimized 0.1 Ag-BDDL photocathode by performing an isotopic experiment because very high energy photons (222 nm, ~5 eV) could degrade the BDDL surface or the resin used in the electrode contacts. The isotopic experiment was carried out by purging the 25 mM Na2SO4 electrolyte for 10 min with 13CO2 at a flow rate of 3 mL/min. Photoelectrochemical reduction was then carried out at −1.1 V vs. RHE under 222 nm irradiation and GC–mass spectrometry (GC-MS) was used to analyse the isotopic yield. Figure 4a shows the amounts of isotopic 13CO and normal 12CO after different periods. The amount of 13CO produced (267.56 μg after 5 h of illumination at −1.1 V vs. RHE) is lower than that of 12CO obtained previously (590 μg after 3 h; see Fig. 2c). This is because CO2 purging was carried out for 1 h in the previous experiment, so more CO2 was dissolved in the electrolyte compared with that in the case of isotopic 13CO2 (Fig. 1d, Supplementary Fig. S3). After illumination for 1 and 5 h, the amounts of isotopic 13CO were 86.42% and 94.17%, respectively. The increase in 13CO content is because only a little non-isotopic 12CO (16.8 μg after 5 h) could enter the system from the resin used to fabricate the electrode contact, which might be degraded by high-energy photons. The amount of 12CO produced by resin degradation is very small compared with that of 13CO produced through photoelectrochemical reduction. Therefore, the relative amount of isotopic 13CO increased over time. This isotopic experiment clearly reveals that CO originates from the reduction of CO2, not from degradation of resin, other organics or the BDD surface by the high-energy irradiation. Even after 5 h of 222 nm irradiation, 0.1 Ag-BDDL was stable and produced 13CO with an excellent yield.
X-ray photoelectron spectroscopy (XPS) was performed before and after photoelectrocatalysis to investigate the elemental states of the 0.1 Ag-BDDL electrode. Figure 4b displays the metallic Ag 3d photoelectron peaks33,34. The relative amount of Ag and nature of Ag 3d photoelectron peaks remained similar before and after photoelectrolysis at −1.1 V under 222 nm irradiation, indicating the oxidation states of the Ag nanoparticles did not change during photoelectrochemical operation; i.e., the metallic state of Ag is stable.
The recyclability of the 0.1 Ag-BDDL electrode was examined by performing five runs for 3 h at −1.1 V in 25 mM Na2SO4. After each run, the cathodic peak position shifted to lower overpotential with slightly increased hydrogen evolution current compared with that at higher overpotential. After five runs, the cathodic peak for CO2 reduction moved to lower overpotential by 0.15 V with a current density of 0.2 mA/cm2 (Supplementary Fig. S4). The amounts of CO and H2 produced were measured (Fig. 4c,d). After five runs, the amount of CO produced in 3 h was 260 μg (44%), comparatively lower than that obtained for the first cycle (590 μg in 3 h). Conversely, the amount of H2 produced in the fifth cycle (10.51 μg) was greater than that obtained in the first (1.85 μg). The decrease in CO and increase in H2 produced with cycle number are attributed to the shift of the onset potential of CO2 reduction to lower overpotential as photoelectrolysis was conducted at −1.1 V (Supplementary Fig. S4).
Efficient photoelectrochemical CO2 reduction is very important for use of CO2 as a chemical feed stock and from an environmental viewpoint. Diamond is a highly stable material with a very wide bandgap (5 eV), possessing a wide electrochemical potential window, mechanical stability and biocompatibility19,30,35. Another advantage of diamond is that its conduction band (CB) position is very high, and hydrogen termination can shift its CB above the vacuum level, providing negative electron affinity towards photoexcited electrons31. Despite the very high energy of the photoexcited electrons of diamond, it is difficult to use under normal conditions. This is because pure diamond has very poor active sites for adsorption of foreign molecules. At sufficiently high pressure, the photoexcited electrons of hydrogen-terminated diamond or H-BDD can be used for photoelectrochemical conversion30,31. The photoexcited electrons in hydrogen-terminated diamond or H-BDD are capable of reducing CO2 via a one-electron transfer mechanism in 0.1 M Na2SO4 at sufficiently high pressure (2.5 MPa)30. Both the requirements of high pressure and hydrogen termination of the catalyst limit the application of diamond as a catalyst. Therefore, here we modified oxygen-terminated BDDL with Ag nanoparticles to realize photoelectrochemical CO2 reduction under normal pressure and temperature. Ag nanoparticles were deposited on BDDL because Ag is known for its selective electrochemical CO2 reduction36,37. The selectivity of Ag originates from its ability to form strong chemical bonds with CO2 under a certain applied potential, which is very important for product selectivity36,37,38. Strong chemical bonding of a catalyst surface with CO2 reduction intermediates can terminate the reduction reaction at a particular point without further propagation, unlike Cu surfaces, which are known to further propagate reduction reactions to produce multiple CO2 reduction products12,36.
Photoelectrochemical studies and product analysis revealed that the optimized 0.1 Ag-BDDL electrode exhibited high efficiency and selectivity in the formation of CO in aqueous electrolyte at −1.1 V vs. RHE with a Faradaic efficiency of 72.5% after 3 h under 222 nm irradiation (Fig. 3). Photoelectrochemical studies using this optimized 0.1 Ag-BDDL electrode were also carried out under similar conditions with 172- and 308-nm irradiation. Irradiation at 308 nm induced a small increase of CO production (235 μg in 3 h). Conversely, 172-nm irradiation yielded less CO (97 μg in 3 h), similar to -electrochemical CO production alone over the same electrode (Fig. 2c). The marked photoelectrochemical effect of 222 nm irradiation on the performance of the optimized 0.1 Ag-BDDL electrode reveals the role of excitation of valence band (VB) electrons to the CB (Figs 1c,d and 2, Supplementary Fig. S5) in CO production. This is because the band gap of BDDL (~5 eV) matches well with a 222 nm light source.
Bare BDDL irradiated with 222 nm light did not efficiently produce CO2 reduction products (~19 μg CO that originated from the degradation of impurities; see Fig. 4a) under similar experimental conditions, as shown in Fig. 2a and Supplementary Fig. S6. This is consistent with the finding of Hamers et al.30. that photoexcited electrons produced in a BDD semiconductor could not reduce CO2 into CO under normal pressure (Supplementary Fig. S6). Thus, the electrochemical CO2 reduction activity of Ag-BDDL is related to the presence of Ag nanoparticles36. However, electrochemical reduction of CO2 by Ag-BDDL in aqueous electrolyte at −1.1 V vs. RHE decreased markedly after a few minutes and then became poor (Fig. 1d). This behaviour is also evident from the consecutive CV runs conducted under CO2-saturated conditions in the dark (Supplementary Fig. S2a). In the dark, the cathodic peak current obtained at −1.1 V vs. RHE in 25 mM Na2SO4 decreased until it reached a constant value (Supplementary Fig. S2a). Conversely, the cathodic peak current increased until it reached its maximum value after a certain irradiation time under 222 nm irradiation at −1.1 V vs. RHE. This increase in photocurrent (Fig. 1d, Supplementary Fig. S3) is thought to be caused by two factors: first, photoreduction of the BDDL surface produces photoactive sites at high negative applied potential, and second, simultaneous enhanced photoelectrochemical reduction of CO2 on the Ag surface. This situation was confirmed by XPS analysis. The C 1 s peaks of 0.1 Ag-BDDL before and after photoelectrolysis at −1.1 V vs. RHE in 25 mM Na2SO4 for 5 h are presented in Supplementary Figs S7 and S8. A broader C 1 s peak [full width at half maximum (FWHM) of 2.73 eV] is obtained before photoelectrochemical reaction, and becomes narrower (FWHM=1.51 eV) after photoelectrolysis for 5 h at −1.1 V in the presence of 222 nm irradiation (Supplementary Fig. S8). This is because BDDL is oxygen terminated, so a small amount of C-O-C and C=O/C-OH bonds are present in addition to C-C bonds at its surface. The C=O/C-OH bonds might be reduced at −1.1 V under 222 nm irradiation. Therefore, the C 1 s peak displays greater C-C sp3 character after photoelectrolysis because of the decreased amount of C=O/C-OH bonds, causing it to narrow (Supplementary Fig. S8). Note that electrochemical bias potential alone cannot reduce the diamond surface; 222 nm irradiation is needed for BDD surface reduction, as evident from XPS, CV, and chronoamperometric analyses (Fig. 1c,d; Supplementary Figs S7 and S8). Our results reveal that 0.1 Ag-BDDL is activated under 222 nm irradiation at −1.1 V vs. RHE to provide very high photoelectrochemical CO2 reduction activity.
In aqueous electrolyte, CO is mostly formed from CO2 via reaction (1), where 2 H+ and 2e− participate. In organic solvent/aprotic electrolyte, reaction (2) is favoured at sufficiently high overpotential or high pressure39,40. Generally, reaction (3) is rare for photoelectrochemical CO2 reduction because reduction is usually carried out at negative bias potential, so such photodissociation on the electrode surface is difficult30. Note that in aqueous media, hydrogen is often produced through proton reduction41, e.g., 2H+ + 2H+ + 2e− = H22e− in addition to CO2 reduction at the electrode surface. A negligible amount of hydrogen found with the Ag-BDDL electrode could produce via proton reduction. However, the experimental results reveal that this smaller amount of hydrogen is produced through electrochemical reduction of proton at the BDD surface (Figs 2 and 3). Hori et al.36. reported that Ag, Au and Zn can electrochemically reduce CO2 to CO with high selectivity. This is possibly because of the formation of strong chemical bonds between and the catalyst surface. Formation of the anion radical involves one-electron transfer (), and this anion species is very unstable. The high photoelectrochemical activity of 0.1 Ag-BDDL in aqueous electrolyte at −1.1 V vs. RHE is attributed to the transfer of very high energy photoexcited electrons in the CB of BDDL to the Ag nanoparticle surfaces to facilitate CO2 reduction. The electrochemical cathodic peak of the 0.1 Ag-BDDL electrode under CO2-saturated conditions suggests that CO2 reduction occurs on the Ag surface. Therefore, the intermediates of CO2 reduction must be adsorbed on the Ag surface37. It has been reported that Ag surfaces could stabilize anion radicals at certain negative potential depending on the experimental conditions, electrolyte and pH (standard redox potential for is −1.9 V)12,36,37,42. BDD can produce the anion radical in 0.1 M Na2SO4 and this radical can be stabilized as an intermediate product under suitable experimental conditions (high pressure)30. Therefore, the enhanced photoelectrochemical CO2 reduction of 0.1 Ag-BDDL is thought to be related to the formation of anion radicals by the transfer of photoexcited electrons to the Ag nanoparticles. The produced anion radicals are then immediately stabilized by a proton-coupled electron transfer mechanism at the Ag surface (Fig. 5) to produce the more stable products CO and H2O37.
In summary, the optimized Ag-BDDL electrode exhibits enhanced photoelectrochemical CO2 reduction activity for CO formation with minimal H2 production under 222 nm irradiation. This Ag-BDDL electrode exhibits superior activity to that of highly doped BDDH and GCE modified with Ag nanoparticles. The optimized Ag-BDDL electrode is stable after five cycles, retaining 44% of its initial activity for CO production. A high isotopic yield (94.17% after 5 h) revealed that the electrode is not corroded under 222 nm irradiation at −1.1 V. The content of surface oxygenated species decreases slightly during the photoelectrochemical reaction, which might increase the number of active sites for CO2 reduction. Therefore, the photocurrent increases during the first hour of photoelectrochemical reaction with the 0.1 Ag-BDDL electrode. Such surface photoreduction was not observed after 1 h of photoelectrochemical reaction. The enhanced activity of Ag-BDDL is ascribed to the synergistic interaction of the Ag nanoparticles with semiconducting BDDL. Ag is known for its selective CO2 reduction, while the very high energy of the CB of BDDL enhanced the formation of anion radicals on the Ag surface in aqueous electrolyte. The anion radicals were stabilized on the Ag surface at −1.1 V via a proton-coupled electron transfer mechanism to produce CO. The very high energy of the CB electrons of BDDL could be coupled with other co-catalysts to potentially improve the yield of several kinetically and thermodynamically limited photochemical conversions.
Deposition of Ag nanoparticles
BDD thin films were fabricated on silicon substrates using a reported microwave plasma chemical vapour deposition method35. Ag nanoparticles were electrodeposited on BDDL (effective area 1 × 1 cm) in a two-electrode system at −0.5 V with respect to a Pt counter electrode in aqueous AgNO3 (0.025–0.1 M) for 60 s. After Ag nanoparticle deposition, the electrodes were cleaned by sonication in deionized water (18.2 MΩ) for 15 min. After washing, electrodes were dried under a N2 flow at room temperature.
Photoelectrochemical CO2 reduction activity was investigated in an H-type electrochemical cell. Working and counter electrodes were separated by a proton-exchange Nafion membrane, and an aqueous solution of Na2SO4 was used as the supporting electrolyte. Ag/AgCl saturated with KCl was used as a reference electrode and Pt mesh served as a counter electrode. The scan rate was 100 mV/s. Excimer lamps (7 W) with wavelengths of 172, 222 and 308 nm were used in the photoelectrochemical studies. Light sources were positioned 2 cm away from the electrode and the reaction temperature was held at 25 °C throughout the experiments.
Gas chromatography analysis
Gaseous products were analysed by a gas chromatograph (GC-2014, Shimadzu, Japan) equipped with a thermal conductivity detector and flame ionization detector. Isotopic 13CO was measured by GC-MS (GC-2010 Plus, Shimadzu) and the relative amount was estimated by comparison with the amount of 12CO produced.
Ion chromatography analysis
An IC system (Dionex ICS-1000, Thermo Scientific) equipped with DionexIonPac AS12A and DionexIonPac CS12A columns was used to identify any detectable CO2 reduction liquid products obtained with the Ag-modified BDDL electrodes.
An FESEM (JEOL, JEM-3100F) was used to analyse the surface morphology of the Ag-BDDL electrodes. Energy-dispersive X-ray spectroscopy was conducted for elemental mapping of the Ag nanoparticle-modified BDDL substrates. Elemental states of Ag-BDDL electrodes before and after photoelectrochemical reaction at −1.1 V for 5 h were determined with an X-ray photoelectron spectrometer (ESCA-3400, Shimadzu) using a monochromatic Mg Kα source.
How to cite this article: Roy, N. et al. Boron-doped diamond semiconductor electrodes: Efficient photoelectrochemical CO2 reduction through surface modification. Sci. Rep. 6, 38010; doi: 10.1038/srep38010 (2016).
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
We thank Japan Science and Technology for financial support through the ACT-C program. C.T. acknowledges the partial financial support from Hosokawa Powder Technology Foundation and Tokyo Ohka Foundation for the Promotion of Science and Technology.
This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/