CO Selectivity for Au Nanoparticles during

Nov 3, 2017 - This degree of instability makes it impossible to correlate the structure of AuNPs determined prior to electrocatalysis to their catalyt...
2 downloads 17 Views 3MB Size
Article Cite This: J. Am. Chem. Soc. XXXX, XXX, XXX-XXX

pubs.acs.org/JACS

Size Stability and H2/CO Selectivity for Au Nanoparticles during Electrocatalytic CO2 Reduction Jamie A. Trindell, Jan Clausmeyer, and Richard M. Crooks* Department of Chemistry, Texas Materials Institute, The University of Texas at Austin, 105 East 24th Street, Stop A5300, Austin, Texas 78712-1224, United States S Supporting Information *

ABSTRACT: In this paper, we show that Au nanoparticles (AuNPs) stabilized with either citrate or by low-generation dendrimers rapidly grow during electrocatalytic reduction of CO2. For example, citrate-stabilized AuNPs and AuNPs encapsulated within sixth-generation, hydroxylterminated, poly(amidoamine) dendrimers (G6-OH DENs) having diameters of ∼2 nm grow substantially in size (to 6−7 nm) and polydispersity during just 15 min of electrolysis at −0.80 V (vs RHE). This degree of instability makes it impossible to correlate the structure of AuNPs determined prior to electrocatalysis to their catalytic function. In contrast to the G6-OH dendrimer, the higher generation G8-OH analogue stabilizes AuNPs under the same conditions that lead to instability of the other two materials. More specifically, G8-OH DENs having an initial size of 1.7 ± 0.3 nm increase to only 2.2 ± 0.5 nm during electrolysis in 0.10 M NaHCO3 at −0.80 V (vs RHE). Even when the electrolysis is carried out at −1.20 V, the higher-generation dendrimer stabilizes encapsulated AuNPs. This is presumably due to the compactness of the periphery of the G8-OH dendrimer. Although the G8-OH dendrimer nearly eliminates AuNP growth, the surface of the AuNP is still accessible for electrocatalytic reactions. The smaller, more stable G8-OH DENs strongly favor formation of H2 over CO. Some previous reports have suggested that AuNPs in the ∼2 nm size range yield primarily CO, but we believe these findings are a consequence of the growth of the AuNPs during catalysis and do not reflect the true function of ∼2 nm AuNPs.



INTRODUCTION Gold nanoparticles (AuNPs) in the ∼2 nm size range have been reported to be active for electrochemical CO2 reduction (to CO)1−5 or to preferentially result in proton reduction (to H2).6 We find, however, that AuNPs of the size used in these studies are unstable and increase substantially in diameter over a period of just minutes under typical electrochemical reaction conditions. Accordingly, it is unlikely that the ∼2 nm AuNPs characterized prior to CO2 reduction are actually the active electrocatalysts, and therefore, the relationship between structure and function remains an open question. To address this point, we synthesized dendrimer-encapsulated AuNPs (Au DENs) in this same size range and found that they remain structurally stable under conditions commonly used for electrochemical CO2 reduction. These stabilized AuNPs do not preferentially catalyze CO2 reduction but rather yield mostly H2 in accordance with previous findings by Strasser, Roldan Cuenya, and co-workers.6 Electrochemical CO2 reduction at bulk Au electrodes was reported by Hori in 1985,7 and it has been shown to proceed with faradaic efficiencies of up to 91% for formation of CO.8 More recent interest in electrocatalytic CO2 reduction has focused on NP catalysts1−3,5,6,9−16 and, in particular, how their morphology and size affect product selectivities. As discussed in the previous paragraph, conflicting results have been reported © XXXX American Chemical Society

in this regard; specifically, whether very small thiol-capped Au clusters (∼1 nm in diameter)1,3 or larger, aggregated particles are more active for CO formation.9,10,17,18 We hypothesized that these contradictions arise due to the analytical challenges associated with postcatalysis characterization of small AuNPs. Suggestive evidence for this comes from one report in which careful materials characterization was carried out after electrolysis. In that case, Manthiram et al. observed extensive aggregation following CO2 reduction by thiol-capped AuNPs having an initial size of 4.2 ± 0.5 nm.18 In the present study, we prepared Au DENs,19−22 containing an average of 147 atoms (Au147), using both sixth- and eighthgeneration, hydroxyl-terminated PAMAM dendrimers (G6-OH and G8-OH, respectively), and examined their electrocatalytic behavior toward CO2 reduction. We23−26 and others27−32 have previously shown that DENs are stabilized by the dendrimer framework but that reactants are still able to penetrate the dendrimer, encounter the NP surface, and undergo catalytic reactions. These characteristics make DENs good model systems for understanding a broad range of catalytic reactions.33−36 Received: June 29, 2017

A

DOI: 10.1021/jacs.7b06775 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Article

Journal of the American Chemical Society Our results show that ∼2 nm citrate-capped AuNPs (no dendrimer stabilization) and Au147 NPs encapsulated within lower generation G6-OH dendrimers increase in size (up to ∼6−7 nm) and polydispersity during CO2 reduction. This inherent instability of small AuNPs makes it difficult to quantitatively correlate structure (e.g., size effects) to function. We have found, however, that G8-OH dendrimers stabilize the size of AuNPs during electrocatalysis and significantly reduce the increase in polydispersity. Specifically, G8-OH(Au147) DENs grow only slightly: from 1.7 ± 0.3 to 2.2 ± 0.5 nm under the electrocatalytic conditions studied here. The results show that stabilized G8-OH(Au147) DENs, while exhibiting comparable current densities relative to the other catalysts tested, yield mostly H2, making them unattractive candidates as CO2 reduction catalysts.



chemical Analyzer (Austin, TX). Electrolysis studies were carried out in CO2- or N2-saturated 0.10 M NaHCO3 solutions using a twocompartment electrochemical cell (MFC set 100.25.3, Adams and Chittenden, Berkley, CA) incorporating a Nafion 117 cation-exchange membrane (∼25 cm2). Each compartment contained ∼120 mL of electrolyte solution. The anode compartment was outfitted with a 5 cm × 5 cm carbon-mesh counter electrode, and the cathode compartment held the catalyst-modified GCE (or a gold foil, ∼2 cm2) and a miniature, “leakless” Ag/AgCl/3.5 M KCl reference electrode (eDAQ, Colorado Springs, CO). Reported potentials were converted to the reversible hydrogen electrode (RHE) scale to take into account the different pH values of 6.7 and 9.0 measured for CO2and N2-saturated 0.10 M NaHCO3 solutions, respectively. A mass-flow controller (Tylan Model FC 280) was used to control the flow of gases into the cathode compartment (flow rate = 20 mL/min). The effluent gas stream was routed into a Shimadzu GC-2014 gas chromatograph (GC) equipped with both TCD and FID detectors. Gas chromatographic analysis of the product gas stream was conducted after running the electrolysis at either −0.80 V or −1.20 V (vs RHE) for 15 min. Linear sweep voltammograms (LSVs) of the electrolyte solutions were obtained after completion of the electrolysis experiments and after resaturation of the solution with CO2 or N2. The LSV scans were carried out between 0.50 V and the potential used for electrolysis studies, either −0.80 V or −1.20 V (vs RHE), respectively, at a scan rate of 5 mV/s. Following CO2 electrolysis and LSV, the electrochemically active surface areas (ECSA) of the Au catalysts used for the electrochemical studies were determined by cyclic voltammetry (CV) in 0.5 M H2SO4 (Figure S1). These experiments were carried out using the same electrochemical cell that was used for electrolysis. The charge corresponding to the Au reduction peak was used to calculate the ECSA of Au using the conversion factor of 390 μC/cm2.38 Current densities were calculated using these Au ECSA values. Scanning Transmission Electron Microscopy (STEM). Size analysis of the AuNPs before and after electrochemical experiments was carried out using a JEOL 2010F transmission electron microscope with a point-to-point resolution of 0.19 nm. Samples used in these studies were collected by wiping a lacey carbon-coated Cu grid (Electron Microscopy Sciences, Hatfield, PA) over the wetted surface of the electrode.39 Size distributions were determined by analyzing 200 particles from three independent electrolysis experiments, resulting in a total of 600 particles analyzed for each catalyst type.

EXPERIMENTAL SECTION

Chemicals. All chemicals were used as received unless otherwise noted. G6-OH and G8-OH PAMAM dendrimers were purchased from Dendritech, Inc. (Midland, MI). Dendrimers were received as 10−25% solutions in methanol and dried under vacuum prior to use. Citratecapped AuNPs were purchased from BBI Solutions (Madison, WI). HAuCl4·3H2O (≥99.9% trace metal grade), 5 wt % Nafion 117 solution, Au foil (0.025 mm thick, 99.99%), and NaBH4 (99.99%) were purchased from Sigma-Aldrich (St. Louis, MO). H2SO4 (95%) and 0.50 M NaOH (diluted to 0.30 M for DENs synthesis) were purchased from Fluka Analytical (Mexico City, Mexico). NaHCO3 powder (≥99.7%) was purchased from J.T. Baker (Avantor Performance Materials, Inc., Center Valley, PA). Nafion 117 membranes (183 μm thick) and Vulcan carbon (EC-72R) were purchased from Fuel Cell Store (College Station, TX). Unless otherwise indicated, all aqueous solutions were prepared using deionized (DI) water (18.2 MΩ Milli-Q water, Millipore, Bedford, MA). CO2 (99.998%, research grade) and N2 (99.999%, ultrahigh purity) were purchased from Praxair, Inc. (Austin, TX). All electrolyte solutions for electrochemical experiments were prepared using HPLCgrade, submicron-filtered water from Fisher Chemical (Pittsburgh, PA). Carbon mesh and glassy carbon rods (100 mm long, 6 mm in diameter) were purchased from Alfa Aesar (Tewksbury, MA). Electrodes were polished using 0.3 μm alumina (Baikowski International Corp., Charlotte, NC) and 0.05 μm alumina (Buehler, Lake Bluff, IL). Synthesis of G6-OH(Au147) and G8-OH(Au147) DENs. The syntheses of G6-OH(Au147) and G8-OH(Au147) DENs were carried out according to our previously published literature reports.37 First, 200 μL of 100 μM PAMAM dendrimer was added to 8.65 mL of 18.2 MΩ DI water with vigorous stirring. Next, 147 μL of a 20.0 mM HAuCl4 stock solution was added dropwise to the stirred PAMAM solution. The gold salt was allowed to sequester within the dendrimer for 2 min before reduction of the precursor complex with a ∼10-fold molar excess of NaBH4 mixed with 1.0 mL of 0.30 M NaOH solution. The reaction mixture was then stirred at 25 ± 2 °C for 12 h to deactivate excess NaBH4. The final reaction solution contained 10.0 mL of 2.0 μM Au147 DENs. Note that the notation “Au147” is not intended to suggest that these NPs contain exactly 147 atoms each but rather that this is the average Au/dendrimer ratio used in their preparation. Electrode Preparation. Composite inks were prepared by sonicating 400 μL of 2-propanol with ∼4 mg of Vulcan carbon and 48 μL of 5 wt % Nafion 117 solution. Next, 2.0 mL of 2.0 μM G6OH(Au147) or G8-OH(Au147) DENs solution was added to the mixture, and it was sonicated again for ∼10 min. Catalyst-coated electrodes were prepared by drop-casting 20 μL of this ink onto 6 mm diameter glassy carbon electrodes (GCEs). To maintain the same ratio of Vulcan carbon to AuNPs, the same method was used to modify electrodes with citrate-capped AuNPs. Electrochemical Measurements. Electrochemical measurements were obtained using a CH Instruments Model CHI700D Electro-



RESULTS AND DISCUSSION Preparation and Characterization of G6-OH(Au147) and G8-OH(Au147) Catalyst Inks. As discussed in the Experimental Section, G6-OH and G8-OH PAMAM dendrimers were used as templates to prepare Gn−OH(Au147) DENs, and Vulcan carbon and Nafion binder were used to prepare the Au DEN-containing composite inks.2−4 STEM analysis revealed that the size of the Au DENs (in the Vulcan/ Nafion ink) prior to electrocatalysis was 1.7 ± 0.3 nm for both the G6-OH(Au147) and G8-OH(Au147) DENs. For comparison, we also prepared a composite ink containing commercially available (nominally) 2 nm citrate-capped AuNPs that exhibited a measured particle diameter of 2.8 ± 0.5 nm. CO2 Reduction. The behavior of AuNP electrocatalysts during electrochemical CO2 reduction was examined under experimental conditions that have been widely used by others.1,3,5,6,13,18 Specifically, the electrolyses were carried out in aqueous, CO2-saturated 0.10 M NaHCO3 at potentials of either −0.80 or −1.20 V (vs RHE) for 15 min. We chose these potentials because they have been commonly used in previous studies of electrocatalytic CO2 reduction.2,3,5,13 We begin this section by discussing the results obtained at −0.80 V (vs RHE), and then compare these findings to those obtained at −1.20 V (vs RHE). B

DOI: 10.1021/jacs.7b06775 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Article

Journal of the American Chemical Society Current versus time (i−t) traces obtained at −0.80 V (vs RHE) for the three different AuNP catalysts are shown in Figure 1a (note that the shapes of these i−t trances are fully

Figure 1. Current density vs time traces for the Au catalysts shown in the legend. The electrode potential was set to (a) −0.80 V (vs RHE) or (b) −1.20 V (vs RHE), and the electrolyte was CO2-saturated 0.10 M NaHCO3. Note the use of different scales on the vertical axes of the two frames. The AuNPs and DENs were immobilized in a Vulcan/ Nafion ink. Current densities were calculated using the Au ECSA measured after all electrochemical experiments.

reproducible, Figure S2). There are two notable features in the i−t data. First, the G6-OH(Au147) DENs exhibit the highest average current density relative to the other catalysts. Second, the current for the G6-OH(Au147) catalyst decreases steadily throughout the duration of the 15 min electrolysis. This type of slow decrease in current density as a function of time has previously been attributed to loss of ECSA and, as we will show later, that is the likely cause here too.18,40 In contrast to the trace observed for the G6-OH(Au147) DENs, the i−t trace for the citrate-capped AuNPs reveals a sharp initial decrease in current density (Δj = 13 mA/cm2), followed by a near steady-state profile, as is typical of i−t traces reported in the literature by others for CO2 reduction at AuNPs.1,5,13,18 This rapid initial decrease is likely primarily due to rapid growth of the citrate-capped AuNPs (vide infra). There is also an initial decrease in the i−t trace for the G8-OH(Au147) DENs, but it is much smaller than the others (Δj = 2 mA/cm2) and likely arises primarily from capacitive charging. Note that capacitive charging is always observed, even in the absence of the catalysts (Figure S3). AuNP Stability. Samples for STEM analysis were collected from the working electrodes after 15 min of electrolysis at −0.80 V (vs RHE). The size distributions of the AuNPs were then analyzed and compared to the corresponding sizes of the same AuNPs prior to electrolysis (Figure 2). Both the average sizes and the polydispersity of citrate-capped AuNPs (Figure

Figure 2. STEM characterization of (a) citrate-capped AuNPs (b) G6OH(Au147) DENs, and (c) G8-OH(Au147) DENs immobilized in Vulcan/Nafion inks before (top row in each frame) and after (bottom row) electrolysis at −0.80 V (vs RHE) in CO2-saturated 0.10 M NaHCO3. Size distributions obtained prior to electrolysis were obtained by measuring a total of 200 NPs from each of the three inks. After electrolysis, the size distributions were determined using NPs from three independent electrolyses for each of the three inks shown (200 NPs/trial; 600 total NPs for each ink). C

DOI: 10.1021/jacs.7b06775 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Article

Journal of the American Chemical Society

hypothesize that the compact periphery of the G8-OH dendrimer is primarily responsible for limiting their growth.19,44−46 Evidence for this comes from previous studies aimed at the intentional extraction of AuNPs from different generation dendrimers, which showed that it was more difficult to remove AuNPs from higher generations.47 Other studies of the periphery of dendrimers have shown that small molecules are preferentially excluded from G8-OH compared to G6-OH dendrimers.25 One final point: when we changed the functional group on the Gn-OH periphery to −NH2 we found that the resulting G8-NH2(Au147) DENs were also stable during electrolysis, increasing in size from 1.7 ± 0.3 to just 2.2 ± 0.4 nm. Accordingly, the chemical nature of the periphery of the dendrimer does not seem to be important for stability. Catalytic Selectivity. The GC results of the product analysis for electrolyses carried out at −0.80 V (vs RHE) are shown in Figure 3a. Both Au foil and citrate-capped AuNPs produce mostly CO at −0.80 V (vs RHE) with comparable production rates (0.7 and 0.8 mL/(h·cm2), respectively). The G6-OH(Au147)-modified electrode generates the most gas overall and produces twice as much H2 (2.0 ± 0.5 mL/(h· cm2)) as CO (1.0 ± 0.1 mL/(h·cm2)). Previous studies of CO2

2a) and the (lower generation) G6-OH(Au147) DENs (Figure 2b) increase during the CO2 reduction experiments. Interestingly, regardless of the stabilizing ligand or the initial size distribution, the sizes of these two types of particles converge to the same average diameter of 6 ± 2 nm. As mentioned previously, this growth, and the corresponding loss of surface area, may be responsible for the current density decreases observed in Figure 1a. On the basis of the large size of the G6OH(Au147) particles after electrolysis, we conclude that the majority of the Au DENs are no longer confined inside the G6OH dendrimers, which have a diameter of only ∼6.7 nm.22,41 While neither the citrate capping agent nor the G6-OH dendrimers are sufficient to stabilize the AuNPs when they are exposed to a potential of −0.80 V (vs RHE), the shapes of the i−t traces shown in Figure 1a suggest different rates of particle growth for each catalyst.18,42 In contrast to the steep, immediate loss of current density observed for the citratecapped AuNPs, the i−t trace of the G6-OH(Au147) reveals a slow, steady decline in current density throughout the duration of the experiment. Therefore, while the G6-OH dendrimer does not completely stabilize the small size of the AuNPs during CO2 reduction, we hypothesize that it does reduce the rate of particle growth relative to the citrate-capped AuNPs. The growth mechanism of the citrate-capped and G6OH(Au147) DENs is not certain, but the formation of large, spherical particles is consistent with a dissolution and growth model.43 In a previous study, ∼4 nm alkanethiol-capped AuNPs were spin-coated onto a GCE and used to catalyze the reduction of CO2.18 This resulted in formation of dendritic aggregates, rather than the simple NP growth observed in our study. The authors interpreted their findings in terms of a twodimensional random-walk model. We thought it would be interesting to directly compare the fate of the alkanethiolcapped AuNPs to Au DENs, and therefore, we drop-cast G6OH(Au147) DENs directly onto a GCE (no Vulcan/Nafion ink) and carried out electrocatalytic CO2 reduction. The resulting aggregates are shown in Figure S4, and they mirror those reported by Manthiram et al.18 The interesting conclusion is that the AuNP growth mechanism depends on the support. Studies intended to better understand this correlation are presently underway in our laboratory. We carried out two control experiments to confirm that the increase in particle sizes observed for AuNPs in Vulcan/Nafion inks requires the negative potentials required for CO 2 reduction. First, exposure of G6-OH(Au147)-modified electrodes to the 0.10 M NaHCO3 electrolyte solution at open circuit potential (OCP) for 2 h did not result in significant changes in particle size (Figure S5). Second, we were concerned that the positive potentials used for the final measurement of the ECSA of the AuNPs could cause particle growth. Accordingly, the G6OH(Au147)-modified electrode was cycled to positive potentials of up to 1.70 V (vs RHE) in 0.5 M H2SO4. The results (Figure S6) revealed only minor changes in particle size (increase from 1.7 to 2.0 nm), indicating that this factor is not responsible for the ∼6 nm size observed after electrolysis. In contrast to the citrate-stabilized AuNPs and G6-OH(Au147) DENs, STEM analysis of G8-OH(Au147) DENs revealed that the particle size largely remained stable during CO2 electrolysis. Specifically, the G8-OH(Au147) DENs had an initial size of 1.7 ± 0.3 nm and grew to 2.2 ± 0.5 nm during electrolysis (Figure 2c). While further investigation is required to fully understand how the G8-OH dendrimer enhances the stability of encapsulated AuNPs during electrolysis, we

Figure 3. Bar graphs showing the amount of CO (red with black error bars) and H2 (black with blue error bars) formed during electrolysis using the indicated Au catalysts at (a) −0.80 V (vs RHE) and (b) −1.20 V (vs RHE). Note the use of different scales on the vertical axes. The AuNPs and DENs were immobilized in Vulcan/Nafion inks. Electrolysis was carried out for 15 min in 0.10 M NaHCO3, and gas chromatography analysis was conducted on the product stream. Results are reported as a function of the volume of gas produced per hour relative to the Au ECSA measured after all electrochemical experiments. No CO production was observed over the G8OH(Au147) DENs at −0.80 V (vs RHE). D

DOI: 10.1021/jacs.7b06775 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Article

Journal of the American Chemical Society reduction have revealed that AuNPs having an average diameter of ∼8 nm are more active for CO evolution than smaller AuNPs.6,13 This finding was rationalized by DFT calculations. Specifically, the calculations suggested that 8 nm AuNPs possess the optimum number of edge sites, thought to be the most active sites for CO2 reduction, for maximizing the CO:H2 ratio. By analogy, we propose that the growth of the citratecapped AuNP and G6-OH(Au147) DEN catalysts to sizes in this same range (∼6−7 nm) is responsible for the significant levels of CO evolution observed. The key finding from our study is that, in contrast to the AuNPs that undergo substantial growth, the more stable ∼2 nm G8-OH(Au147) DENs exhibit exclusive selectivity for H2 (2.1 ± 0.1 mL/(h·cm2)). Indeed, the amount of CO produced over the G8-OH(Au147) DENs was below the 100 ppm detection limit of the GC. This observation is consistent with DFT calculations, which showed that the increased relative number of under-coordinated corner sites to edge sites in AuNPs ∼2 nm in diameter make them more active for H2, thereby suppressing CO production.6 Potential Dependence of Catalytic Activity. To better understand the effect of potential on both the size stability and product selectivity of AuNPs, additional CO2 reduction experiments were performed at −1.20 V (vs RHE), which, in addition to −0.80 V (vs RHE), has been extensively used to study CO2 reduction.1−3,6,18 Comparison of the i−t traces at −0.80 and −1.20 V (vs RHE) (parts a and b, respectively, of Figure 1) reveals some notable differences. First, while the G8OH(Au147) DENs exhibit the highest average current density during electrolysis at −1.20 V (vs RHE), a significant decrease in current density occurs during the first 2 min of the electrolysis (the shapes of these i−t traces are reproducible, Figure S7). Second, while the current density of G6-OH(Au147) DENs was observed to decrease steadily over 15 min in the experiment carried out at −0.80 V (vs RHE) (Figure 1a), a steady-state current was achieved within ∼1 min at −1.20 V (vs RHE) (Figure 1b). Not surprisingly, after electrolysis at the higher potential of −1.20 V (vs RHE), STEM analysis of the different AuNP catalysts reveals that the citrate-capped AuNPs and the G6OH(Au147) DENs still both grow (6 ± 2 and 7 ± 2 nm, respectively, Figure S8). Therefore, the rapid decay of the current density (Δj = 11 mA/cm2), observed for the G6OH(Au147) catalyst in Figure 1b, suggests that the growth of the particles likely occurs more rapidly at −1.20 V (vs RHE) than for −0.80 V (vs RHE). Interestingly, STEM analysis also reveals that the stability of the G8-OH(Au147) DENs is slightly diminished under these more extreme electrocatalytic conditions. While size analysis (Figure S8) clearly shows that the vast majority of the particles retain diameters of ∼2 nm even after electrolysis at this more negative potential, a few larger particles are also observed. Another interesting difference between the data collected at −0.80 V (vs RHE) and −1.20 V (vs RHE) is observed in the postelectrolysis LSVs. Specifically, the LSV for the G8OH(Au147) catalyst exhibits a distinct peak centered around −0.9 V (vs RHE) (Figure 4). At the present time, we are unable to identify the origin of this peak, but it may be associated with the formation of a soluble product not detectable by GC. Experiments to identify soluble products, such as formate,48−50 are presently underway. A much smaller version of this peak is also apparent for LSVs obtained using the G6-OH(Au147) DENs, and it may originate from a low percentage of AuNPs

Figure 4. LSVs of the electrode materials indicated in the legend. The AuNPs and DENs were immobilized in Vulcan/Nafion inks. The LSVs were obtained in CO2-saturated 0.10 M NaHCO3 after electrolysis at −1.20 V (vs RHE) for 15 min. The working electrode was at the opencircuit potential (OCP) for 15 min during resaturation of the electrolyte with CO2 prior to collecting LSVs. As shown in Figure S5, there is essentially no particle growth at the OCP. The scans began at 0.50 V (vs RHE) and ended at −1.20 V (vs RHE), and the scan rate was 5 mV/s. Current densities were calculated using the Au ECSA measured after electrolysis experiments.

that remain encapsulated within the G6-OH dendrimers (i.e., those that do not grow). Figure 3b is a histogram that compares the gaseous product distributions for the different electrocatalysts after 15 min of electrolysis at −1.20 V (vs RHE). Interestingly, the Au foil, citrate-capped AuNPs, and G6-OH(Au147) DENs all produced primarily CO: 1.2 ± 0.4, 2.2 ± 0.5, and 2.0 ± 0.4 mL/(h·cm2), respectively. While the latter two AuNP catalysts appear to be highly active for CO evolution, their high degree of size polydispersity observed after electrolysis (Figure S8) makes it difficult to ascertain the exact contribution of the various catalyst sizes toward this observed activity. Finally, for the G8OH(Au147) DENs at −1.20 V (vs RHE), the formation of H2 gas (3.6 ± 0.7 mL/(h·cm2)) predominates over CO (1.5 ± 0.4 mL/(h·cm2)). While this observed product ratio (2.4:1) is in fair agreement with existing literature,6 the presence of a few large (5−10 nm) AuNPs may skew the product distribution.



SUMMARY AND CONCLUSIONS In summary, thorough particle size analysis using STEM demonstrates that both commercially available citrate-capped AuNPs and G6-OH(Au147) DENs increase substantially in size during CO2 reduction in 0.10 M NaHCO3 at −0.80 and −1.20 V (vs RHE). The results show that the size of these types of NPs converge to mean particle diameters of ∼6−7 nm. The resulting spherical particle shape and increased polydispersity are in agreement with a dissolution and growth mechanism. In contrast to the citrate-capped AuNPs and G6-OH(Au147) DENs, higher generation G8-OH and G8-NH2 dendrimers stabilize AuNPs during electrolysis. Specifically, the G8OH(Au147) DENs only increase in size from 1.7 ± 0.3 to 2.2 ± 0.5 nm after 15 min of electrolysis at −0.80 V (vs RHE). We hypothesize that the more sterically crowded periphery of the larger dendrimer minimizes particle growth while simultaneously allowing reactants and products to pass through the dendrimer superstructure and encounter the surface of the encapsulated AuNPs. E

DOI: 10.1021/jacs.7b06775 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Article

Journal of the American Chemical Society During electrolysis at −0.80 V (vs RHE), the stable, ∼2 nm G8-OH(Au147) DEN electrocatalysts yield exclusively H2 as a product (Figure 3a). This preferential reduction of H+ relative to CO2 is even higher than that observed in a previous report, demonstrating increased selectivity for H2 for small, ligand-free AuNPs.6 Taken together with previous reports of CO2 reduction using AuNPs, there are two main points that emerge from our study. First, the stability of electrocatalysts must be carefully considered before accurate structure−function correlations can be determined. Second, small (∼2 nm), stable AuNPs yield exclusively H2 as a product6 under conditions that have previously been reported to yield CO.1−5



(7) Hori, Y.; Kikuchi, K.; Suzuki, S. Chem. Lett. 1985, 14, 1695− 1698. (8) Hori, Y.; Murata, A.; Kikuchi, K.; Suzuki, S. J. Chem. Soc., Chem. Commun. 1987, 14, 728−729. (9) Chen, Y.; Li, C. W.; Kanan, M. W. J. Am. Chem. Soc. 2012, 134, 19969−19972. (10) Nursanto, E. B.; Jeon, H. S.; Kim, C.; Jee, M. S.; Koh, J. H.; Hwang, Y. J.; Min, B. K. Catal. Today 2016, 260, 107−111. (11) Lates, V.; Falch, A.; Jordaan, A.; Peach, R.; Kriek, R. J. Electrochim. Acta 2014, 128, 75−84. (12) Back, S.; Yeom, M. S.; Jung, Y. ACS Catal. 2015, 5, 5089−5096. (13) Zhu, W.; Michalsky, R.; Metin, Ö .; Lv, H.; Guo, S.; Wright, C. J.; Sun, X.; Peterson, A. A.; Sun, S. J. Am. Chem. Soc. 2013, 135, 16833− 16836. (14) Liu, M.; Pang, Y.; Zhang, B.; De Luna, P.; Voznyy, O.; Xu, J.; Zheng, X.; Dinh, C. T.; Fan, F.; Cao, C.; de Arquer, F. P. G.; Safaei, T. S.; Mepham, A.; Klinkova, A.; Kumacheva, E.; Filleter, T.; Sinton, D.; Kelley, S. O.; Sargent, E. H. Nature 2016, 537, 382−386. (15) Rogers, C.; Perkins, W. S.; Veber, G.; Williams, T. E.; Cloke, R. R.; Fischer, F. R. J. Am. Chem. Soc. 2017, 139, 4052−4061. (16) Burdyny, T.; Graham, P. J.; Pang, Y.; Dinh, C.-T.; Liu, M.; Sargent, E. H.; Sinton, D. ACS Sustainable Chem. Eng. 2017, 5, 4031− 4040. (17) Feng, X.; Jiang, K.; Fan, S.; Kanan, M. W. J. Am. Chem. Soc. 2015, 137, 4606−4609. (18) Manthiram, K.; Surendranath, Y.; Alivisatos, A. P. J. Am. Chem. Soc. 2014, 136, 7237−7240. (19) Crooks, R. M.; Zhao, M.; Sun, L.; Chechik, V.; Yeung, L. Acc. Chem. Res. 2001, 34, 181−190. (20) Myers, V. S.; Weir, M. G.; Carino, E. V.; Yancey, D. F.; Pande, S.; Crooks, R. M. Chem. Sci. 2011, 2, 1632−1646. (21) Niu, Y.; Crooks, R. M. C. R. Chim. 2003, 6, 1049−1059. (22) Scott, R. W. J.; Wilson, O. M.; Crooks, R. M. J. Phys. Chem. B 2005, 109, 692−704. (23) Pande, S.; Weir, M. G.; Zaccheo, B. A.; Crooks, R. M. New J. Chem. 2011, 35, 2054−2060. (24) Feng, Z. V.; Lyon, J. L.; Croley, J. S.; Crooks, R. M.; Vanden Bout, D. A.; Stevenson, K. J. J. Chem. Educ. 2009, 86, 368−372. (25) Niu, Y.; Yeung, L. K.; Crooks, R. M. J. Am. Chem. Soc. 2001, 123, 6840−6846. (26) Chechik, V.; Crooks, R. M. J. Am. Chem. Soc. 2000, 122, 1243− 1244. (27) Ye, R.; Yuan, B.; Zhao, J.; Ralston, W. T.; Wu, C.-Y.; Unel Barin, E.; Toste, F. D.; Somorjai, G. A. J. Am. Chem. Soc. 2016, 138, 8533− 8537. (28) Bingwa, N.; Meijboom, R. J. Phys. Chem. C 2014, 118, 19849− 19858. (29) Pozun, Z. D.; Rodenbusch, S. E.; Keller, E.; Tran, K.; Tang, W.; Stevenson, K. J.; Henkelman, G. J. Phys. Chem. C 2013, 117, 7598− 7604. (30) Lemo, J.; Heuzé, K.; Astruc, D. Inorg. Chim. Acta 2006, 359, 4909−4911. (31) Johnson, J. A.; Makis, J. J.; Marvin, K. A.; Rodenbusch, S. E.; Stevenson, K. J. J. Phys. Chem. C 2013, 117, 22644−22651. (32) Bernechea, M.; De Jesus, E.; Lopez-Mardomingo, C.; Terreros, P. Inorg. Chem. 2009, 48, 4491−4496. (33) Anderson, R. M.; Zhang, L.; Wu, D.; Brankovic, S. R.; Henkelman, G.; Crooks, R. M. J. Electrochem. Soc. 2016, 163, H3061− H3065. (34) Luo, L.; Zhang, L.; Duan, Z.; Lapp, A. S.; Henkelman, G.; Crooks, R. M. ACS Nano 2016, 10, 8760−8769. (35) Ostojic, N.; Thorpe, J. H.; Crooks, R. M. J. Am. Chem. Soc. 2016, 138, 6829−6837. (36) Luo, L.; Duan, Z.; Li, H.; Kim, J.; Henkelman, G.; Crooks, R. M. J. Am. Chem. Soc. 2017, 139, 5538−5546. (37) Kim, Y.-G.; Oh, S.-K.; Crooks, R. M. Chem. Mater. 2004, 16, 167−172. (38) Yancey, D. F.; Carino, E. V.; Crooks, R. M. J. Am. Chem. Soc. 2010, 132, 10988−10989.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/jacs.7b06775. Representative CV used for evaluation of Au ECSA; additional i−t traces for CO2 electrolysis experiments over (a) G6-OH(Au147) and (b) citrate-capped AuNPs at −0.80 V (vs RHE); current density versus time traces for Vulcan/Nafion and GCE control experiments; STEM image of G6-OH(Au147) after electrolysis with no Vulcan/Nafion ink; size distribution of G6-OH(Au147) DENs following exposure to CO2-saturated 0.10 M NaHCO3 for 2 h at open circuit potential; size distribution of G6-OH(Au147) DENs following ECSA measurements; additional i−t traces for CO2 electrolysis experiments over G8-OH(Au147) at −1.20 V (vs RHE); size distribution of all three AuNP catalysts after electrolysis studies −1.20 V (vs RHE) (PDF)



AUTHOR INFORMATION

Corresponding Author

*[email protected] ORCID

Richard M. Crooks: 0000-0001-5186-4878 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We gratefully acknowledge support from the Chemical Sciences, Geosciences, and Biosciences Division, Office of Basic Energy Sciences, Office of Science, U.S. Department of Energy (Contract: DE-FG02-13ER16428). We thank the Robert A. Welch Foundation (Grant No. F-0032) for sustained support of our research. J.C. is grateful for a RURS Gateway Fellowship.



REFERENCES

(1) Kauffman, D. R.; Thakkar, J.; Siva, R.; Matranga, C.; Ohodnicki, P. R.; Zeng, C.; Jin, R. ACS Appl. Mater. Interfaces 2015, 7, 15626− 15632. (2) Kauffman, D. R.; Alfonso, D.; Matranga, C.; Qian, H.; Jin, R. J. Am. Chem. Soc. 2012, 134, 10237−10243. (3) Andrews, E.; Katla, S.; Kumar, C.; Patterson, M.; Sprunger, P.; Flake, J. J. Electrochem. Soc. 2015, 162, F1373−F1378. (4) Andrews, E.; Flake, J.; Fang, Y. ECS Trans. 2015, 66, 67−70. (5) Zhu, W.; Zhang, Y.-J.; Zhang, H.; Lv, H.; Li, Q.; Michalsky, R.; Peterson, A. A.; Sun, S. J. Am. Chem. Soc. 2014, 136, 16132−16135. (6) Mistry, H.; Reske, R.; Zeng, Z.; Zhao, Z.-J.; Greeley, J.; Strasser, P.; Roldan Cuenya, B. J. Am. Chem. Soc. 2014, 136, 16473−16476. F

DOI: 10.1021/jacs.7b06775 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Article

Journal of the American Chemical Society (39) Yancey, D. F.; Zhang, L.; Crooks, R. M.; Henkelman, G. Chem. Sci. 2012, 3, 1033−1040. (40) Celorrio, V.; Montes de Oca, M. G.; Plana, D.; Moliner, R.; Lazaro, M. J.; Fermín, D. J. J. Phys. Chem. C 2012, 116, 6275−6282. (41) Jackson, C. L.; Chanzy, H. D.; Booy, F. P.; Drake, B. J.; Tomalia, D. A.; Bauer, B. J.; Amis, E. J. Macromolecules 1998, 31, 6259−6265. (42) Min, X.; Kanan, M. W. J. Am. Chem. Soc. 2015, 137, 4701−4708. (43) Ferreira, P. J.; la O’, G. J.; Shao-Horn, Y.; Morgan, D.; Makharia, R.; Kocha, S.; Gasteiger, H. A. J. Electrochem. Soc. 2005, 152, A2256− A2271. (44) Hu, J.; Xu, T.; Cheng, Y. Chem. Rev. 2012, 112, 3856−3891. (45) Fleming, C. J.; Liu, Y. X.; Deng, Z.; Liu, G. J. Phys. Chem. A 2009, 113, 4168−4174. (46) Gröhn, F.; Bauer, B. J.; Akpalu, Y. A.; Jackson, C. L.; Amis, E. J. Macromolecules 2000, 33, 6042−6050. (47) Garcia-Martinez, J. C.; Crooks, R. M. J. Am. Chem. Soc. 2004, 126, 16170−16178. (48) Plana, D.; Flórez-Montaño, J.; Celorrio, V.; Pastor, E.; Fermín, D. J. Chem. Commun. 2013, 49, 10962−10964. (49) Monzó, J.; Malewski, Y.; Kortlever, R.; Vidal-Iglesias, F. J.; SollaGullón, J.; Koper, M. T. M.; Rodriguez, P. J. Mater. Chem. A 2015, 3, 23690−23698. (50) Hahn, C.; Abram, D. N.; Hansen, H. A.; Hatsukade, T.; Jackson, A.; Johnson, N. C.; Hellstern, T. R.; Kuhl, K. P.; Cave, E. R.; Feaster, J. T.; Jaramillo, T. F. J. Mater. Chem. A 2015, 3, 20185−20194.

G

DOI: 10.1021/jacs.7b06775 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX