Revealing Catalyst Restructuring and Composition During Nitrate Electroreduction through Correlated Operando Microscopy and Spectroscopy

doi:10.21203/rs.3.rs-3310807/v1

Download PDF

Article

Revealing Catalyst Restructuring and Composition During Nitrate Electroreduction through Correlated Operando Microscopy and Spectroscopy

https://doi.org/10.21203/rs.3.rs-3310807/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Determining the active phase of an electrocatalyst at work is key to understanding its properties. However, the operating morphology of electrocatalysts is challenging to investigate because they can restructure into different motifs under applied potential due to changes in their oxidation state. These transformations will further alter their catalytic properties. Here, we employ a multi-modal approach centered on electrochemical liquid cell transmission electron microscopy (EC-TEM) to study the evolution of cubic Cu₂O pre-catalysts during the electrocatalytic nitrate reduction reaction and unveil how redox kinetics determine the working catalyst morphology. We found drastic differences in catalyst restructuring during operation and a strong dependency of its composition on the applied potential and the chemical environment. Moreover, by matching the timescales of morphological changes observed in EC-TEM with time-resolved chemical state information obtained from operando transmission soft X-ray microscopy, hard X-ray absorption spectroscopy and Raman spectroscopy, we reveal that Cu₂O can be kinetically stabilized for extended durations under moderately reductive conditions due to the formation of surface hydroxides. Finally, we rationalize how the interaction between the electrolyte and the catalyst influences the ammonia selectivity.

Physical sciences/Chemistry/Catalysis/Electrocatalysis

Physical sciences/Materials science/Materials for energy and catalysis/Electrocatalysis

Physical sciences/Materials science/Techniques and instrumentation/Imaging techniques

Physical sciences/Materials science/Techniques and instrumentation/Microscopy/Transmission electron microscopy

Electrocatalytic chemical conversion reactions such as carbon dioxide reduction reaction (CO₂RR)^1,2 and nitrate reduction reaction (NO₃RR)^3,4 are key to the advancement of various green energy solutions. However, it can be difficult to identify the active catalyst species in these reactions, even when the metallic state is supposed to be the stable phase, because the catalyst can change its oxidation state during reaction according to the external stimuli. Although the Pourbaix diagram⁵ can be used to rationalize the stable oxidation state/phase at different applied potentials and pHs, they are equilibrium diagrams which do not consider the kinetics of redox transitions and their effect on the catalyst morphology. For example, they do not include information about how oxide-to-metal transformations occur, how different facets can reconstruct differently under the same reaction conditions, how interactions between the catalysts and the electrolyte can alter catalyst surface or how reaction intermediates and products may lead to further changes.

The challenge here is two-fold. First, one must elucidate the working morphology of the electrocatalyst. Second, one needs to disentangle the role the observed morphological changes on the catalytic performance. There are, nonetheless, only a few methods^6,7 that can visualize the nanoscale restructuring dynamics of a catalyst and follow the catalyst as a function of the applied potential and electrolyte conditions. It is even more challenging to resolve the local chemical state of these features because most operando techniques for extracting chemical information, such as Raman spectroscopy and X-ray absorption spectroscopy (XAS), are “broad beam” methods, where the data is an ensemble signal derived from a large probed region. This gap between nanoscale imaging and ensemble-averaging spectroscopy limits our ability to rationalize how catalyst morphology impacts the overall performance of these complex but important reactions.

NO₃RR is foremost among various electrochemical conversion reactions in terms of its need for more insight into the electrocatalyst's phase during reaction. Notably, it is a promising strategy for mitigating freshwater pollution from agricultural fertilizer run-off and industrial waste⁸, and is also studied for its potential to produce NH₃^9–11, which is an important chemical in industry and a candidate carrier for green hydrogen^12,13. While Cu is one of the most studied electrocatalyst materials for NO₃RR due to its optimal nitrate adsorption energy^11,14, whether metallic Cu^15–18, Cu oxides or a Cu-Cu oxide interface¹⁹ are the key species for the selective NH₃ formation has remained largely unresolved. According to the Pourbaix diagram^20,21, the metallic phase of Cu should be the stable phase of Cu under typical NO₃RR conditions, but studies using in situ Raman spectroscopy had suggested that an oxide phase might exist during the reaction^14,19. Cu and its oxides are also known to be susceptible to etching²² and facet modification^23,24 by NH₃. Furthermore, it has been reported that NO₃RR can drive the dissolution and re-growth of single atom Cu catalysts³, and the clustering of small aggregates into larger nanoparticles (NPs)²⁵.

Here, we use electrochemical liquid cell transmission electron microscopy (EC-TEM) accompanied by correlated multi-modal operando investigations that include electrochemical liquid cell transmission X-ray microscopy (EC-TXM), operando XAS and operando Raman spectroscopy of the same pre-catalysts to visualize in real time how the structure and composition of Cu₂O cubes evolve as a function of the applied potential during NO₃RR. We found that the working electrocatalyst morphology was determined by three processes, (i) the dissolution of Cu₂O, (ii) the re-deposition of Cu from soluble Cu complexes^26,27 and (iii) the reduction of Cu₂O to metallic Cu. We also discovered a coexistence of Cu₂O with metallic Cu for extended reaction durations, thereby providing insight into the active Cu species during NO₃RR.

For the operando microscopy experiments, we prepared well-defined Cu₂O cubes on the carbon working electrode of the EC-TEM chips via electro-deposition^28,29 as shown schematically in Fig. 1a. The as-prepared Cu₂O cubes have an average size of 250 nm and consist of six {100} facets without the exposure of other minor facets such as {110} or {111} (Suppl. Figure 1). All voltages indicated in this paper are referenced against a Ag/AgCl electrode and then converted to the reversible hydrogen electrode scale (RHE) using the Nerst equation and the bulk pH of the electrolyte. Intriguingly, the image sequences show that the cubes do not undergo significant change in the commonly employed 0.1 M Na₂SO₄ + 8 mM NaNO₃ electrolyte for NO₃RR (Fig. 1b), during the initial sweep towards cathodic potentials. Cu₂O should reduce directly to metallic Cu at the higher overpotentials of the sweep, where according to the Pourbaix diagram, the redox potential for the transformation of Cu₂O to metallic Cu is 0 V_RHE in a solution with pH 7^20,21, and metallic Cu is the stable phase below − 0.2 V_RHE onwards. The stability of the Cu₂O cubes is remarkable because these redox transformations usually lead to morphological changes. For comparison, Fig. 1c depicts a Cu₂O cube under CO₂RR conditions in CO₂-saturated 0.1 M KHCO₃ at a potential similar to that applied in the NO₃RR experiment. As we reported previously²⁹, the latter cubes undergo fragmentation together with the re-deposition of small particles, a behavior that differs from the morphologically much more stable NO₃RR samples at the same applied potentials. The linear sweep voltammograms acquired during these two experiments are available as Suppl. Figure 2.

Next, we studied these Cu₂O cubes systematically at different sustained potentials from − 0.2 V_RHE to -0.6 V_RHE (Fig. 2a-e) to probe further their morphological stability during NO₃RR. For these extended experiments, we adopted an intermittent imaging protocol (images captured at 15-minute intervals with the electron beam blanked the rest of the time) to minimize beam-induced dissolution of the Cu₂O cubes (see discussion in Suppl. Note 1) and ensure that the catalyst re-structuring kinetics we extract from the collected data are as accurate as possible. The electrochemical current profiles over time at each potential measured in these EC-TEM experiments are provided as Suppl. Figure 3. At -0.2 V_RHE (Fig. 2a), the cubes were stable during our entire observation, with no significant restructuring observed. From − 0.2 to -0.5V_RHE, dissolution/re-deposition is the main restructuring pathway. At -0.3 V_RHE, the cubic form persisted for almost 135 min (Fig. 2b), while the cube completely dissolved after 140 min at -0.4 V_RHE (Fig. 2c) and after 90 min at -0.5 V_RHE (Fig. 2d). The lighter contrast of the cube exterior in Fig. 2c compared to the middle of the cube at 60 and 80 min is explained by the cube corners and edges being etched first. The weaker contrast of the dissolving Cu₂O cubes compared to that of the growing Cu NPs also suggests that the dissolving cubes were still in oxide form.

Two Cu₂O cubes were captured in the images acquired at -0.6 V_RHE (Fig. 2e). One cube shrank and restructured into a smaller cube with a void in the center and then became rougher due to small NPs attaching to its surface, while another completely dissolved within the same time frame. We also highlight that the intensity of the cubes in the TEM images obtained from − 0.3 to -0.5 V_RHE gradually decrease, whereas the intensity of the cube at -0.6 V_RHE is brighter, implying that cube-like frame at -0.6 V_RHE is metallic. Moreover, the interplay of dissolution/re-deposition and direct reduction at the more cathodic potentials means that the terminal catalysts morphologies of oxide pre-catalysts vary depending on the applied potential.

Next, we repeated the NO₃RR experiments in a H-type cell with Cu₂O cubes electrodeposited on carbon paper to check for the consistency of the EC-TEM results with standard reaction geometries. Figures 2f and 2g show lower magnification images of samples from the EC-TEM experiments with scanning electron microscopy (SEM) images of samples extracted from H-type cell experiments after 2 hours of reaction at three different applied potentials, -0.2, -0.4 and − 0.6 V_RHE. Electron diffraction patterns taken from samples extracted after reaction show that the cubes did not undergo significant restructuring at -0.2 V_RHE and remain Cu₂O whereas samples reacted at -0.6 V_RHE were largely metallic (Suppl. Figure 4). Conversely, samples reacted at -0.4 V_RHE show a mixture of residual Cu₂O and metallic Cu structures (Suppl. Figure 4–6). The morphological differences between the sample after reaction in the H-type cell at -0.4 V_RHE and − 0.6 V_RHE further support that the catalysts indeed restructure through different pathways as described by our EC-TEM experiments. Inductively coupled plasma mass spectrometry measurements of the electrode and the electrolyte in the H-type cell after reaction further show that Cu dissolution happens at all the applied potentials (Suppl. Figure 7). Therefore, these experiments indicate that the Cu₂O cubes undergo a gradual dissolution under NO₃RR conditions, which in turn leads to the re-deposition of metallic particles elsewhere on the working electrode with shapes and sizes that are modulated by the applied potential.

To obtain unambiguously the oxidation state of the catalyst species present during reaction and rule out the possibility that the ex situ identified Cu₂O phase are the result of re-oxidation during the return to open circuit potential^30,31 (OCP), we performed operando EC-TXM measurements on the Cu₂O cubes by transferring our EC-TEM holder into a TXM at the BESSY II synchrotron facility as illustrated in Suppl. Figure 8. This unique arrangement maintains the same reaction environment between the two experiments, while enabling time-resolved operando measurements of Cu absorption edges under applied potential without compromising the sustained electrolyte flow as X-rays are attenuated less by the electrolyte and enclosing membranes. Thus, the evolution of the electrode’s composition can be tracked during NO₃RR. Figures 3a-d show the time-resolved evolution of the Cu₂O catalysts during NO₃RR at -0.4 V_RHE as observed by EC-TXM in the form of the colored maps that were reconstructed from a XAS image stack using linear combination fitting (LCF)³². Cu₂O is depicted in red and metallic Cu in yellow. The maps show that Cu₂O and Cu are the dominant phases present over the duration of the NO₃RR and that the oxide and metallic phases coexist under specific reaction conditions, but they are spatially separated. CuO is also detected (blue color) but it is present only in small quantities and is not clearly visible from the maps. The corresponding decomposed spectra are shown in Fig. 3e-p, where the yellow, red and blue colored lines represent the respective Cu species. The total contribution of the individual spectra in Fig. 3e-p represents the amount (thickness) of each species, which indicates that the content of metallic Cu species increases (Fig. 3e-h) during reaction, whereas Cu₂O decreases (Fig. 3i-p). The details of the data acquisition and processing are discussed in Suppl. Note 3. Most importantly, these results confirm the sluggish reduction kinetics of the large Cu₂O cubes in the Na₂SO₄ + NaNO₃ electrolyte, and that the metallic phase forms when the dissolved Cu species re-deposit on the working electrode due to the reductive potential employed.

We further verified that the slow reduction of the Cu oxide cubes extends to larger reaction volumes with operando hard XAS measurements of samples electro-deposited on carbon paper in our home-built electrochemical XAS cell³³. In Fig. 3q, we plot the changes in the Cu K-edge valence states (from 8950 to 9105 eV) that were obtained from operando XAS. The weight of the Cu valence state is extracted by LCF of the X-ray absorption near-edge structure (XANES) of the oxide-derived Cu catalyst collected at a constant potential of -0.4 V_RHE in 0.1 M Na₂SO₄ + 8 mM NaNO₃ electrolyte. As seen in Fig. 3r, the fraction of Cu₂O decreased but not completely after more than 2 hours of electrolysis, while the fraction of metallic Cu increased correspondingly, eventually to almost a 1:1 ratio of Cu₂O:Cu. The XANES results agree with the persistence of Cu₂O and the continual evolution of the Cu species seen in the EC-TEM (Fig. 2c) and EC-TXM (Cu-L₃ edges in Fig. 3r) results at -0.4 V_RHE. Minute amounts of the CuO species were also detected during the experiment. The change in the weights of the three species over time exhibit similar trends in both TXM and XAS, confirming that the results we obtain in the EC-TEM cells indeed extrapolate to a larger ensemble of catalyst particles.

This overall agreement between different methods and experimental geometries means that we can use the in situ TEM image sequences to quantify the potential-dependent dissolution and re-deposition rates. Our method for fraction extraction from the EC-TEM images and additional analysis of the re-deposited particles is described in the Suppl. Note 4. As shown in Fig. 4a, the cube fraction decreases over time at an increasing rate as the potential decreases from − 0.2 V to -0.5 V_RHE. The sample at -0.6 V_RHE deviates from this trend (dark purple line) because of the direct reduction of Cu₂O to metallic Cu. In Fig. 4b, we use the cube dissolution rate to estimate the Cu₂O to Cu ratio at a certain potential and use it to visualize the majority phase (greater than 50%) at different times. We further compare the Cu₂O/Cu ratio with NH₃ conversion activity (current density) and selectivity (Faradaic Efficiency, FE) obtained from our bench top electrochemistry measurements. In Fig. 4c, we plot the linear sweep voltammogram of the Cu₂O cubes prepared on carbon paper and in Fig. 4d-e, their product distribution as a function of the applied potentials. The measured yield rate and the FE towards NH₃ was much higher at -0.6 V_RHE as compared to -0.2 V_RHE and − 0.4 V_RHE, implying that the change in catalytic selectivity is related to the faster rate of oxide to metal conversion at -0.6 V_RHE (Fig. 4b).

Next, we performed EC-TEM studies in various electrolyte compositions to elucidate the mechanism behind Cu₂O stabilization. Suppl. Figure 9a and 9b describe experiments using pure 0.1 M Na₂SO₄ and 0.1 M Na₂SO₄ + 8 mM NaNO₂ (i.e. nitrite reduction) respectively. In both cases, the cubes behaved similarly to their behavior with NO₃RR where they gradually reduced in size until they fragment/reduce at longer reaction duration, which means the Cu₂O stability is related to the Na₂SO₄ supporting electrolyte and not the reactant. Re-deposition was, however, much less in Na₂SO₄ compared to its NaNO₃/NaNO₂ containing counterparts. We attribute this difference to how the local pH during electro-reduction differs in the presence and absence of NO_x species. Under applied cathodic potentials, the pH at the electrocatalyst surface increases as hydroxyl ions form^34,35 due to the reduction of e.g. H₂O, O₂, NO₂⁻ and NO₃⁻. In particular, NO_xRR results in higher currents and consequently a steeper rise in the local pH as compared to when only hydrogen reduction takes place. This process can bring the pH of a neutral electrolyte above 12³⁴, triggering the formation of soluble Cu hydroxides. To look at the effect of electrolyte pH, we further performed experiments in 0.1 M Na₂SO₄ where the pH was increased to 10 by adding NaOH. As shown in Suppl. Figure 9c, it altered the amount of re-deposition observed. Lastly, to probe the influence of ammonium ions, we deliberately added NH₄OH into the 0.1 M Na₂SO₄ carrier electrolyte, which led to rapid restructuring of the cubes as shown in Suppl. Figure 9d.

To explain these results, we consider the phase stability of Cu as a function of pH and in the presence of NH₃. A complex series of reactions encompassing different acid–base chemistries, Cu(OH)₂ precipitation, and complex ion formation are known for the Cu-NH₃ system^36,37 (see Suppl. Note 5). Specifically, the equilibrium between solid Cu(OH)₂ and the Cu(NH₃)₄²⁺ complex depends on the NH₃ concentration, with Cu(OH)₂ precipitation being favored at low concentrations due to the poor solubility of Cu(OH)₂³⁶. We hypothesize that the sluggish reduction observed may be the result of transient surface Cu(OH)₂ formation induced by the interfacial pH rise in the course of the NO₃RR. Cu(OH)₂ formation may also be more favorable in Na₂SO₄, due to the electrolyte’s inability to buffer the local pH increase from electro-reduction³⁴, as compared to the KHCO₃ electrolyte used in CO₂RR, thereby leading to the differences in re-structuring behaviors. To validate this hypothesis, we studied the chemical changes taking place on the surface of the Cu₂O cubes with operando Raman spectroscopy measurements.

Figure 5a shows the results of experiments at constant applied potentials performed with cubes electrodeposited on glassy carbon plates. At OCP, the Raman spectrum shows three features of bands centered at 415, 520, and 630 cm^− 1 respectively, which are in good agreement with the reported values of Cu₂O^38,39. When − 0.2 V_RHE was applied, the band intensity at 520 and 630 cm^− 1 decreased over time but continued to persist, which is consistent with the gradual dissolution of Cu₂O. At -0.4 V_RHE, a new peak at 475 cm^− 1 started to emerge, while the peak at 630 cm^− 1 flattened, indicating oxide to metal transition. At -0.6 V_RHE, the characteristic bands of Cu₂O were less pronounced and additional weak bands emerged at 450, 475 and 590 cm^− 1. The band around 475 cm^− 1 can be assigned to the Cu-O-H vibration^39–41 or Cu(OH)₂⁴⁰, while the peak at 590 cm^− 1 is often assigned to CuO³⁸ or adsorbed O-species on Cu⁴¹. These results suggest the formation of a transient intermediate oxide or hydroxide phase during electrolysis, which is also supported by the extended presence of Cu(I) signatures and the weak but persistent Cu(II) signatures in the EC-TXM and operando XAS measurements in Fig. 3. It was, however, difficult to obtain the Raman signatures of absorbates with these samples due to the relatively low loading of the cubes that we were able to electrodeposit and the overlap of the D/G bands of the glassy carbon support with surface absorbate bands, which limits the signal-to-noise ratios at those bands. Thus, we repeated the Raman measurements with cubes electrodeposited at higher loading on carbon paper to improve the signal-to-noise ratios to identify surface adsorbed species or intermediate species in the vicinity of electrode. These results are discussed in Supplementary Note 6.

Hence, we arrive at a restructuring mechanism illustrated in Fig. 5b where surface hydroxides first form on the Cu₂O cubes due to the pH increase induced by electrolysis, which delays the oxide reduction. The subsequent NH₃ production and added pH rise from continued NO₃RR then destabilizes this hydroxide layer and forms soluble Cu complexes, thereby initiating catalyst evolution via re-deposition from Cu complex reduction or the aggregation of migrating NPs. We also performed more EC-TEM experiments to check that the delayed restructuring kinetics extend to other pre-catalyst geometries. Suppl. Figure 10 describes the evolution during NO₃RR of Cu₂O truncated octahedrons and metallic frames created by pre-reducing the Cu₂O cubes. Both samples are stable during early-stage reaction. With more time, re-deposition similar to that seen in the cubes was observed in the Cu₂O octahedrons whereas little re-deposition was noticeable with the metallic frames, which was likely due to less dissolution occurring when we start from metallic pre-catalysts.

The correlated microscopy and spectroscopy experiments presented here therefore indicate that the morphology of the Cu catalysts during NO₃RR at a given pH is governed by a complex, time- and potential-dependent interplay of three processes: (i) dissolution of oxide and hydroxide species, (ii) metal re-deposition, and (iii) oxide catalyst reduction. According to the Pourbaix diagram^20,21, metallic Cu is the only stable species under the specific applied potentials and pH of our experiments, but we have shown that oxidic and metallic phases can co-exist over extended durations and over a broad range of applied potentials, which has significant implications in terms of determining the active species for producing NH₃. It has been suggested previously¹⁹, based on operando spectroscopy measurements of CuO pre-catalysts, that Cu/Cu₂O interfaces are responsible for NH₃ production¹⁹, but these methods cannot differentiate the distribution of these species on the nanoscale. As we have shown in this work, the presence of both spectroscopic signatures in ensemble-averaging measurements does not necessarily mean that the two phases are spatially connected. Furthermore, we have demonstrated that the de-coupled metallic Cu and Cu oxide phases can persist for the extended durations at mild cathodic potentials (less than − 0.5 V_RHE), and that a high residual abundance of Cu₂O in the operando XAS measurements corresponded to a low NH₃ production efficiency in our electrolysis data of equivalent samples. The improvement of NH₃ selectivity with increasing overall metallic character of the samples, therefore, advocates that metallic Cu is the active phase for producing NH₃ compared to Cu₂O, in agreement with recent work on the topic^15–18. In this case, the strong stability of the Cu₂O cubes and their sluggish reduction kinetics in the often-used Na₂SO₄ carrier electrolyte, are detrimental for the NH₃ production by delaying the onset of selective NH₃ formation.

By showing the diverse behaviors that can be derived in different electrolytes and reaction conditions, our work also illustrates the critical need to pay attention to how the electrolyte can influence the restructuring of catalysts and the stability of oxide, hydroxide, and metallic phases before we attempt to generalize results across different studies and reactions. So far, the description of electrolyte effects in electrolysis has been largely confined to cation adsorption effects^42–45 and re-structuring induced by aggressive halide anions^44–46, whereas studies of pH had focused on its impact on reaction mechanisms and NH₃ selectivity^8,10, and not catalyst phase stability. Significant additional microscopy work, such as the one presented here will be required to separate the impact of electrolyte-driven morphological transformation from the much better understood associated electronic and chemical changes. Furthermore, current computational models still cannot rationalize the impact of an explicit complex electrolyte on the catalyst restructuring and its associated influence on the creation of active sites. Efforts to improve these models and advance the theory describing electrocatalytic processes, will undoubtedly require more accurate representations of dynamic catalyst surfaces. The challenge here is serious, since theoretical mechanistic insight must consider two simultaneously occurring dynamic processes, namely the one that the catalyst material experiences, in parallel to that undergone by the reactants, both of which being coupled and driven by the local chemical potential.⁴⁷ Our results revealing phase co-existence also opens the possibility that different species may be responsible for activating specific steps of the conversion reaction. Hence, we expect operando approaches that incorporate chemically resolved microscopy within multi-modal spectroscopic investigations, as we demonstrated here, to play a vital role moving forward in the understanding of electrocatalytic processes by providing a path towards mapping such complexity.

In summary, operando EC-TEM and EC-TXM measurements have revealed that the morphologies of Cu₂O pre-catalysts during NO₃RR and their evolutionary pathways are sensitive to the reaction time, applied potential and the nature of the electrolyte. As expected, the rate of oxide reduction accelerates with increasing negative applied potentials, but spatially-separated oxide and metallic phases can co-exist over extended reaction times under moderately reductive potentials. More importantly, the kinetics of the different re-structuring processes, which have been unveiled here, determine that final morphology of the catalysts. Our results also indicate that the nature of the electrolyte can introduce time-dependent selectivity changes in the early stages of the catalyst restructuring, which will help resolve on-going controversies regarding the active state of Cu for selective NH₃ production. Finally, this work impacts our understanding of how electrocatalysts evolve under reaction condition through the discovery of Cu oxide and hydroxide stability. In addition, we unveiled local structural and chemical heterogeneities that develop under electrochemical working conditions, even on a pre-catalyst sample initially characterized by a narrow size, shape and compositional distribution. Thus, our findings emphasize the need of operando characterization methods to establish connections between materials’ structural and compositional characteristics under specific reaction environments and external stimuli and their electrocatalytic performance.

Specimen preparation

The Cu₂O cubes (250 nm) were synthesized on the polished glassy carbon (vitreous, SPI) plates, carbon paper and the carbon electrode of the Hummingbird Scientific EC-TEM chips using an electro-deposition protocol we previously developed.^28,29 The deposition solution consists of a mixture of 5 mM copper sulfate-pentahydrate (CuSO₄·5H₂O, Sigma Aldrich) and 12.5 mM of potassium chloride (KCl, Sigma Aldrich). After the synthesis, the samples were rinsed with ultrapure water and then used for the subsequent NO₃RR experiments.

Electrolyte preparation for nitrate reduction

The electrolyte used for nitrate reduction experiments is an aqueous solution of 0.1 M Na₂SO₄ (anhydrous, 99.99%, Suprapur)) + 8 mM NaNO₃ (Sigma-Aldrich, ReagentPlus®, ≥ 99.0%).

Operando EC-TEM

The EC-TEM experiments were performed in a Thermo Fisher Scientific 300 kV Titan TEM operated in STEM mode and a Hummingbird Scientific Bulk Liquid Electrochemistry TEM holder with Pt counter and Ag/AgCl (3 M KCl) reference electrodes. The EC-TEM top chips have a 50 nm thick silicon nitride membrane windows and were also produced by Hummingbird Scientific. The bottom chips have a 50 nm silicon window and 250 nm spacers. The electrochemistry experiments were performed using a Biologic SP-200 potentiostat. The potentials were measured against the built-in Ag/AgCl reference.

The TEM holder was pre-filled with 0.1 M Na₃SO₄ + 8mM NaNO₃ solution during cell assembly to fill the entire fluid path with electrolyte. After loading into the TEM, the syringe was filled with 0.1 M Na₃SO₄ + 8mM NaNO₃ and introduced at a flow rate of 1.25 mL min^− 1. Linear sweep voltammetry from − 0.5 V to − 1.1 V_AgAgCl (repeated twice) was first used to determine the onset potential for the NO₃RR using a scan rate of 15 mV s^− 1 and to ensure that the applied potential was consistent between experiments, followed by chronoamperometry for up to 2 hours at − 0.2, -0.3, -0.4, -0.5.1 and − 0.6 V_RHE (converted from Ag/AgCl scale using the Nerst equation) with the new catalyst specimen and fresh electrolyte at each condition.

In situ imaging was always performed under conditions with electrolyte in the cell, as determined from the image contrast. We stayed under an electron flux of 1.75 e⁻ Å⁻² s^− 1 at all times to minimize electron beam-induced artifacts with an electron probe current of approximately 220 pA. The acquired images have a resolution of 1024 × 1024 pixels. For continuous imaging, the images were acquired at 1 frame per second. For intermittent imaging, the images were acquired every 15 minutes with the electron beam blanked-in between. The image segmentation for the EC-TEM movies was performed using built-in functions and scripting in MATLAB (See Suppl. Note 4 for details).

NO₃RR product analysis and detection

The electrochemistry experiments for product analysis and ex situ imaging were conducted by using an Autolab potentiostat (PGSTAT 302N) and a custom-made H-type electrochemical cell, whereby the cathodic and anodic compartments were separated by a piece of anion exchange membrane (Selemion AMV, AGC Inc.). A platinum gauze (MaTecK, 3600 mesh cm^− 2) and a leak-free Ag/AgCl electrode (LF-1, Alvatek, potential 0.198 V vs. standard hydrogen electrode) were used as the counter and the reference electrodes, respectively. The anodic compartment (with counter electrode) was filled with 18 mL 0.1 M Na₂SO₄ electrolyte while the cathodic compartment (with the working electrode) was filled with 18 mL 0.1 M Na₂SO₄ electrolyte + 8 mM NaNO₃. Before the electrochemical tests, the anodic and cathodic solutions were de-aerated by continuously bubbling Ar (grade 6.0, 99.9999%) with a 20 mL min^− 1 flow rate (Bronkhorst). During the chronoamperometric measurements, a constant Ar flow (10 mL min^− 1) was used to maintain the inert atmosphere. Linear sweep voltammetry was performed with a scan rate of 5 mV s^− 1, and chronoamperometry (2 hours reaction time) was used for product distribution measurements.

UV-Vis spectroscopy (Agilent Cary 60) was used to detect and quantify ammonia and nitrite in the electrolyte following the procedures established in previous literature ^48,49. The liquid samples were diluted to match the suitable detection range for spectrophotometric analysis of each analyte and the sample absorbance was measured in the range 400–800 nm.

The indophenol-blue method was used for the determination of ammonia^48,49. For nitrite (NO₂^–) quantification, 3 mL of the diluted sample were added to a glass vial containing 35 mg of white powder from a commercial nitrite test kit (photometric 0.002–1.00 mg L^− 1 NO₂-N, 0.007–3.28 mg L^− 1 NO₂⁻, Spectroquant, Merck). Details on quantification can be found in the Suppl. Note 2.

Ex situ TEM and SEM measurement

The ex situ TEM imaging was also performed with the Thermo Fisher Scientific 300 kV Titan TEM. The EC-TEM chips were loaded using a Hummingbird Scientific Tomography holder for before and after reaction comparisons. The EC-TEM chips were rinsed in ultrapure water after they were disassembled from the EC-TEM holder and immediately transferred into the TEM. The ex situ SEM imaging of the bulk samples was performed using a Thermo Fisher Scientific Apreo SEM.

Operando EC-TXM measurement

Operando EC-TXM experiments were conducted at the U41-TXM beamline in BESSY II (Berlin, Germany). The beam-size was 26 µm × 26 µm with a nominal resolution of 20 nm. The image stacks were collected using a charge coupled detector (CCD) at 1340 pixel × 1300 pixel and the exposure time of 1 second per energy. 10 nm monochromator slit was used. The intensity of the incident radiation was monitored and adjusted to have a photon count constant (approximately 15000 count per pixel) at the background area (no specimen) when liquid is fully filled. Image stacks were acquired as the beam energies were scanned from 926 to 965 eV, which encompassed both Cu-L₃ and Cu-L₂ edges.

A Hummingbird Scientific electrochemistry holder was used for the operando measurements. The applied potential was controlled with a Biologic potentiostat. The reference electrode was a Pt-pseudo reference on the chip and the counter was Pt. The reference potential was calibrated against an external Ag/AgCl electrode to ensure a potential comparable to the EC-TEM experiments were applied.

Details regarding the data processing, including accurate alignment, atomistic background subtraction, data normalization, spectra averaging and linear combination fitting (LCF) of the spectra images can be found in Suppl. Note 2.

Operando XAS

Operando time-resolved X-ray absorption fine-structure spectroscopy (XAFS) experiments were performed at P64 beamline of the PETRA III synchrotron (Hamburg,Germany) in quick XAFS (QXAFS) mode. Measurements were performed at the Cu K-edge (8979 eV). The intensity of the incident radiation was monitored by a gas ionization chamber filled with pure N₂. Additional ionization chambers were used to acquire spectra of a Cu foil in transmission mode for calibration purposes at the beginning of each QXAFS scan. The beam-size was less than 2 × 2 mm. The XAS data were collected in fluorescence mode using a PIPS detector at one spectrum per second and one spectrum per 5 seconds.

For these operando XAS experiments, we used a home-made single-compartment electrochemical cell³³. The applied potential was controlled with a Biologic potentiostat. Argon was flowed into gas compartment at 10 ml/min. The sample was Cu₂O cubes prepared on the carbon paper electrode and the 0.1 M Na₂SO₄ + 8 mM NaNO₃ electrolyte was continuously circulated through the cell using a double-channel peristaltic pump.

Data extraction and calibration were performed using the JAQ software of the P64 beamline. Further data processing and analysis of the XANES spectra were performed according to the procedures described previously in reference ⁵⁰.

Operando Raman measurement

The operando Raman spectra were obtained using a Renishaw (InVia Reflex) confocal Raman microscope and a water immersion objective with a long working distance (Leica microsystems, 63x, numerical aperture of 0.9) was chosen. The glassy carbon experiments used a 785 nm laser with 0.1% laser power (0.36 mW). The carbon paper experiments used a 633 nm laser with laser power of about 5%. The objective was protected from the electrolyte by a Teflon film (DuPont, film thickness of 0.013 mm). Then, a drop of water was used to drive away the air between the film and the objective to match the refractive index to ensure efficient excitation and collection of the Raman signal.

The electrochemical measurements were performed in a home-built spectro-electrochemical cell made of Teflon and controlled by a Biologic SP-240 potentiostat. The cell was equipped with a reference electrode (leak-free Ag/AgCl, Alvatek), a counter electrode (Pt ring), and a working electrode with the catalyst electrodeposited on glassy carbon. A 15 ml Ar-purged 0.1 M Na₂SO₄ + 8mM NaNO₃ solution was used as an electrolyte. During the experiment, the Raman spectra were acquired every 5 minutes over 1 hour of reaction.

Contributions

AY, SWC, and BRC conceived the project, planned the experiments, wrote the manuscript. AY prepared specimens, conducted the operando EC-TEM studies and operando EC-TXM studies, ex situ TEM and ex situ SEM analysis, and analyzed the data. LB measured NO₃RR selectivity and analyzed the electrochemical data. FY assisted with the sample preparation and performed some of the operando EC-TEM experiments. FF provided the initial inspiration for the project and developed the experimental protocol for H-type cell measurements and product analysis. CZ and LB conducted operando Raman measurements and analyzed the results. MR and JT planned and collected the operando XAS data and performed the analysis and interpretation. CP and SW planned the operando EC-TXM measurements with AY and SWC, operated the TXM and helped with analysis of the TXM data. MM and HS were involved in the XAS measurements and helped with analysis and interpretation of the results.

Acknowledgements

AY and CZ thank the Alexander von Humboldt Foundation (AvH) for supporting them with an AvH postdoctoral research grant. L. Bai acknowledges the support from the Early Postdoc Mobility Fellowship (P2ELP2_199800) of Swiss National Science Foundation. FY acknowledges the Chinese Scholarship Council for sponsoring her PhD. This work was partially funded by the German Federal Ministry for Education and Research (BMBF) under the grant Catlab (03EW0015B), the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under project no. 406944504 – SPP 2080 and Germany´s Excellence Strategy – EXC 2008 – 390540038 – UniSysCat. AY also thanks the SPP 2080 “DynaKat” Early Career Research Scholarships for Female Scientists for partial funding. We are very grateful to Dr. Antonia Herzog, Dr. Uta Hejral, Dr. Arno Bergmann, and Shih-Yu Fu for their time and help with the XAS and TXM beamtime measurements. We further acknowledge Dr Andrea Martini and Dr. Christoph Scheurer for their helpful discussions with regards to analysis of the TXM data, Dr Eduardo Ortega for his help with some of the TEM data analysis and Mr. Walter Wachsmann for the ICP-MS measurements. Finally, we thank the Helmholtz-Zentrum Berlin für Materialien und Energie (Berlin, Germany), and DESY (Hamburg, Germany), members of the Helmholtz Association HGF, for the allocation of dedicated synchrotron radiation beamtime.

Gao D, Arán-Ais RM, Jeon HS (2019) Roldan Cuenya, B. Rational catalyst and electrolyte design for CO2 electroreduction towards multicarbon products. Nat Catal 2:198–210
Popović S et al (2020) Stability and Degradation Mechanisms of Copper-Based Catalysts for Electrochemical CO2 Reduction. Angew Chemie - Int Ed 59:14736–14746
Yang J et al (2022) Potential-Driven Restructuring of Cu Single Atoms to Nanoparticles for Boosting the Electrochemical Reduction of Nitrate to Ammonia. J Am Chem Soc
Hu Q et al (2021) Reaction intermediate-mediated electrocatalyst synthesis favors specified facet and defect exposure for efficient nitrate–ammonia conversion. Energy Environ Sci 14:4989–4997
Pourbaix M (1974) Atlas of Electrochemical Equilibria in Aqueous Solutions
Bergmann A (2019) Roldan Cuenya, B. Operando Insights into Nanoparticle Transformations during Catalysis. ACS Catal 9:10020–10043
Zhu Y, Wang J, Chu H, Chu Y-C, Chen HM (2020) Situ/Operando Studies for Designing Next-Generation Electrocatalysts. ACS Energy Lett 5:1281–1291
Wang Z, Richards D, Singh N (2021) Recent discoveries in the reaction mechanism of heterogeneous electrocatalytic nitrate reduction. Catal Sci Technol 11:705–725
Jiao F, Xu B (2019) Electrochemical Ammonia Synthesis and Ammonia Fuel Cells. Adv Mater 31:1805173
van Langevelde PH, Katsounaros I, Koper MTM (2021) Electrocatalytic Nitrate Reduction for Sustainable Ammonia Production. Joule 5:290–294
Jung W, Hwang YJ (2021) Material strategies in the electrochemical nitrate reduction reaction to ammonia production. Mater Chem Front 5:6803–6823
Valera-Medina A, Xiao H, Owen-Jones M, David WIF, Bowen PJ (2018) Ammonia for power. Prog Energy Combust Sci 69:63–102
Chen JG et al (2018) Beyond fossil fuel–driven nitrogen transformations. Science vol. 360 at https://doi.org/10.1126/science.aar6611
Bae SE, Stewart KL, Gewirth AA (2007) Nitrate adsorption and reduction on Cu(100) in acidic solution. J Am Chem Soc 129:10171–10180
Zhou N et al (2023) Potential-Induced Synthesis and Structural Identification of Oxide-Derived Cu Electrocatalysts for Selective Nitrate Reduction to Ammonia. ACS Catal 13:7529–7537
Anastasiadou D et al (2023) Morphology Changes of Cu2O Catalysts During Nitrate Electroreduction to Ammonia**. ChemCatChem 15, e202201503
Costa GF et al (2024) Identifying the active site of Cu/Cu2O for electrocatalytic nitrate reduction reaction to ammonia. Chem Catal. 4
Bai L et al (2024) Electrocatalytic Nitrate and Nitrite Reduction toward Ammonia Using Cu2O Nanocubes: Active Species and Reaction Mechanisms. J Am Chem Soc 146:9665–9678
Wang Y, Zhou W, Jia R, Yu Y, Zhang B (2020) Unveiling the Activity Origin of a Copper-based Electrocatalyst for Selective Nitrate Reduction to Ammonia. Angew Chemie - Int Ed 59:5350–5354
Celante VG, Freitas MBJG (2010) Electrodeposition of copper from spent Li-ion batteries by electrochemical quartz crystal microbalance and impedance spectroscopy techniques. J Appl Electrochem 40:233–239
Puigdomenech I, Taxén C (2000) Thermodynamic data for copper. Implications for the corrosion of copper under repository conditions
Hoar TP, Rothwell GP (1970) The potential/pH diagram for a copper-water-ammonia system: its significance in the stress-corrosion cracking of brass in ammoniacal solutions. Electrochim Acta 15:1037–1045
Siegfried MJ, Choi K-S (2006) Elucidating the effect of additives on the growth and stability of Cu2O surfaces via shape transformation of pre-grown crystals. J Am Chem Soc 128:10356–10357
Luo Q, Mackay RA, Babu SV (1997) Copper Dissolution in Aqueous Ammonia-Containing Media during Chemical Mechanical Polishing. Chem Mater 9:2101–2106
Fu X et al (2020) Alternative route for electrochemical ammonia synthesis by reduction of nitrate on copper nanosheets. Appl Mater Today 19:100620
Schroder D, Schwarz H, Wu J, Wesdemiotis C (2001) Long-lived dications of Cu(H2O)2 + and Cu(NH3)2 + do exist! Chem Phys Lett 3343:258–264
Ducéré JM, Goursot A, Berthomieu D (2005) Comparative density functional theory study of the binding of ligands to Cu + and Cu2+: Influence of the coordination and oxidation state. J Phys Chem A 109:400–408
Grosse P, Yoon A, Rettenmaier C, Chee SW (2020) Roldan Cuenya, B. Growth Dynamics and Processes Governing the Stability of Electrodeposited Size-Controlled Cubic Cu Catalysts. J Phys Chem C 124:26908–26915
Grosse P et al (2021) Dynamic Transformation of Cubic Copper Catalysts during CO2 Electroreduction and its Impact on Catalytic Selectivity. Nat Commun 12:6736
Li Y et al (2020) Electrochemically scrambled nanocrystals are catalytically active for CO2-to-multicarbons. Proc. Natl. Acad. Sci. U. S. A. 117, 9194–9201
Yoon A, Poon J, Grosse P, Chee SW (2022) Roldan Cuenya, B. Iodide-mediated Cu catalyst restructuring during CO2 electroreduction. J Mater Chem A 10:14041–14050
Velasco-Vélez JJ et al (2020) On the Activity/Selectivity and Phase Stability of Thermally Grown Copper Oxides during the Electrocatalytic Reduction of CO2. ACS Catal 10:11510–11518
Timoshenko J, Roldan Cuenya B (2021) In Situ/ Operando Electrocatalyst Characterization by X-ray Absorption Spectroscopy. Chem Rev 121:882–961
Nobial M, Devos O, Mattos OR, Tribollet B (2007) The nitrate reduction process: A way for increasing interfacial pH. J Electroanal Chem 600:87–94
Monteiro MCO, Koper MT (2021) M. Measuring local pH in electrochemistry. Curr Opin Electrochem 25
Johnson AR, McQueen TM, Rodolfa KT (2005) Species distribution diagrams in the copper-ammonia system: An updated and expanded demonstration illustrating complex equilibria. J Chem Educ 82:408–414
Giannopoulou I, Panias D, Paspaliaris I (2009) Electrochemical modeling and study of copper deposition from concentrated ammoniacal sulfate solutions. Hydrometallurgy 99:58–66
Deng Y, Handoko AD, Du Y, Xi S, Yeo BS (2016) In Situ Raman Spectroscopy of Copper and Copper Oxide Surfaces during Electrochemical Oxygen Evolution Reaction: Identification of CuIII Oxides as Catalytically Active Species. ACS Catal 6:2473–2481
Niaura G (2000) Surface-enhanced Raman spectroscopic observation of two kinds of adsorbed OH – ions at copper electrode. Electrochim Acta 45:3507–3519
Chan HYH, Takoudis CG, Weaver MJ (1999) Oxide film formation and oxygen adsorption on copper in aqueous media as probed by surface-enhanced Raman spectroscopy. J Phys Chem B 103:357–365
Bodappa N et al (2019) Early Stages of Electrochemical Oxidation of Cu(111) and Polycrystalline Cu Surfaces Revealed by in Situ Raman Spectroscopy. J Am Chem Soc 141:12192–12196
de Moura M, Kortlever R (2019) Electrolyte Effects on the Electrochemical Reduction of CO2. ChemPhysChem 20, 2926–2935
Marcandalli G, Monteiro MCO, Goyal A, Koper MTM (2022) Electrolyte Effects on CO2 Electrochemical Reduction to CO. Acc Chem Res 55:1900–1911
Nitopi S et al (2019) Progress and Perspectives of Electrochemical CO2 Reduction on Copper in Aqueous Electrolyte. Chem Rev 119:7610–7672
Katsounaros I, Kyriacou G (2007) Influence of the concentration and the nature of the supporting electrolyte on the electrochemical reduction of nitrate on tin cathode. Electrochim Acta 52:6412–6420
Arán-Ais RM et al (2018) Structure- and Electrolyte-Sensitivity in CO2 Electroreduction. Acc Chem Res 51:2906–2917
Chee SW, Lunkenbein T, Schlögl R (2023) Roldán Cuenya, B. Operando Electron Microscopy of Catalysts: The Missing Cornerstone in Heterogeneous Catalysis Research? Chem Rev 123:13374–13418
Andersen SZ et al (2019) A rigorous electrochemical ammonia synthesis protocol with quantitative isotope measurements. Nature 570:504–508
Wang Y, Wang C, Li M, Yu Y, Zhang B (2021) Nitrate electroreduction: mechanism insight, in situ characterization, performance evaluation, and challenges. Chem Soc Rev 50:6720–6733
Jeon HS et al (2021) Selectivity Control of Cu Nanocrystals in a Gas-Fed Flow Cell through CO2 Pulsed Electroreduction. J Am Chem Soc 143:7578–7587

There is NO Competing Interest.

SupplementaryMovie1.avi
Supplementary Movie 1
SupplementaryMovie2.avi
Supplementary Movie 2
SupplementaryMovie3.avi
Supplementary Movie 3
SupplementaryMovie4.avi
Supplementary Movie 4
NO3RRSIResubmissionClear.pdf

Download PDF

Version 1

posted

You are reading this latest preprint version

Revealing Catalyst Restructuring and Composition During Nitrate Electroreduction through Correlated Operando Microscopy and Spectroscopy

Status:

Version 1

Abstract

Figures

Introduction

Results & Discussion

Methods

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1