Continuous ammonia production with plasma assisted electrochemical reactor

Abstract

A continuous-flow plasma-electrochemical system for redox reactions includes a flow-type plasma reactor configured to receive feedstock and to generate an output comprising intermediate species. An electrochemical reactor is configured to receive the intermediate species from the plasma reactor and to generate an output having the desired product (e.g., ammonia).

Description

Attorney Docket No.: 011520.01941 CONTINUOUS AMMONIA PRODUCTION WITH PLASMA ASSISTED ELECTROCHEMICAL REACTOR Cross-Reference to Related Applications

[0001] This application claims priority to U.S. Provisional Application No.63 / 655,056, filed on June 2, 2024, now pending, the disclosure of which is incorporated herein by reference. Field of the Disclosure

[0002] The present invention relates generally to the field of chemical synthesis, and more particularly to systems and methods for ammonia production. Background of the Disclosure

[0003] Ammonia is an essential chemical for the agricultural sector, primarily produced through the energy-intensive Haber−Bosch process. The electrochemical nitrogen reduction reaction (NRR) presents a promising alternative route, enabling the conversion of N2 into ammonia under ambient conditions using water as the proton source, significantly reducing the carbon footprint of ammonia production. However, current electrochemical NRR research is hindered by low Faradaic efficiency to NH3and, more crucially, a low ammonia production rate. To address these challenges, research on nitrogen reduction chemistry has recently diversified into Li-mediated nitrogen reduction and electrochemical reduction of nitrate. Although initial results are promising, it is imperative to broaden the technological spectrum to tackle the ammonia production problem effectively.

[0004] One emerging strategy for nitrogen activation is the incorporation of non-thermal plasma. Non-thermal plasma, produced via dielectric barrier discharge at room temperature, offers a swift and direct approach to activating nitrogen triple bonds. Applications of non- thermal plasma in ammonia production, CO2conversion, methane activation, and chemical synthesis have been demonstrated. Nonetheless, plasma chemistry is complex, often producing molecular ion fragments or excited-state compounds with limited control over reaction outcomes. This challenge can be tackled by integrating a catalyst into the dielectric barrier discharge reactor, enabling some control over reaction pathways. For example, during CO2conversion, the incorporation of a catalyst in the plasma reactor can increase the reactionAttorney Docket No.: 011520.01941 selectivity from 2.5% (without catalyst) to 75% (with catalyst). The field of plasma catalysis is still very much in its infancy.

[0005] Additionally, a two-step approach combining plasma and electrochemical reactors has also been reported. In this method, nitrogen molecules are first broken down into more reactive components, such as NO₂⁻ and NO₃⁻, and collected in a solution. This NOx⁻ solution is then used as an electrolyte in an electrochemical reaction for ammonia production. While this two-step method is viable, it involves a more complex setup that could increase capital costs during scaling. Furthermore, cycling the plasma and electrochemical reactors on and off to switch electrolyte solutions can be challenging. Nonetheless, further advancements in this approach may lead to rewarding outcomes. Brief Summary of the Disclosure

[0006] The present disclosure provides a cascading plasma-electrochemical process to generate ammonia at room temperature and pressure with water as the proton source instead of hydrogen, which significantly lowers the overall carbon footprint. Compared to other electrochemical reaction, the present plasma electrochemical process generated the highest ammonia production rate among reported literature. Furthermore, the present process is able to take air directly as feedstock instead of purified nitrogen. Overall, this technology is a much greener way to produce ammonia with high rate that can lead to commercial value.

[0007] In the presently described continuous flow plasma-electrochemical process for converting nitrogen in air into ammonia, nitrogen molecules are first converted to a mixture of NOx species in a plasma reactor. We employed a graph theory approach with first-principles calculations to exhaustively enumerate all pathways from N2 to NH3, pinpointing key intermediates (NH2* and NO*). A density functional theory (DFT) modeling further screened different electrocatalysts capable of stabilizing the key intermediates post-plasma reactor. An optimized CuPd catalyst was used in the electrochemical reaction, demonstrating an ammonia production rate of 81.2 mg h−1cm−2, two orders of magnitude higher than normal electrochemical nitrogen reduction alone. The stability of the plasma-electrochemical system was also demonstrated for over 1000 hours at a 2 A applied current for the continuous conversion of atmospheric nitrogen to ammonia.Attorney Docket No.: 011520.01941 Description of the Drawings

[0008] For a fuller understanding of the nature and objects of the disclosure, reference should be made to the following detailed description taken in conjunction with the accompanying drawings.

[0009] Figure 1. Investigation of plasma reaction conditions. (a) illustration of an experimental embodiment of a continuous flow plasma-electrochemical system according to the present disclosure: only air was purged in the plasma reactor and aqueous NH4+can be collected at the outlets at room temperature. (b) The effect of feed gas to the plasma reactor on the NH4+production rate. (c) The effect of N2 / O2 feed ratio to the plasma reactor on the NH4+yield rate. All N2 / O2feed ratios are humidified with water. (d) Mass spectra of the products generated by the plasma reactor with humidified air. Liquid products post-plasma reactor were collected by bubbling them into methanol for GCMS analysis. Orange and red lines represent the gaseous products, and the purple line corresponds to the liquid products.

[0010] Figure 2. Schematic diagram of our proposed plasma electrocatalyst-integrated system for high NH4+production rate.

[0011] Figure 3. FTIR spectroscopy of different chemicals dissolved in 1 M KOH solutions. NO2−(~1354 cm−1) and NO3−(~1233 cm−1) peaks can be directivity observed from the products generated by the plasma reactor. The peaks for N2H4 and NH2OH have overlapped by the peak for KOH (~1635 cm−1).

[0012] Figure 4. (a) UV-visible absorption spectroscopy of our plasma generated NO3−. (b) UV-visible absorption spectroscopy of standard solutions with different concentrations of NaNO3. (c) Calibration curves for NO3−concentration.

[0013] Figure 5. (a) UV-visible absorption spectroscopy of our plasma generated NO2−- Griess products. (b) UV-visible absorption spectroscopy of standard solutions with different concentrations of NaNO2. (c) Calibration curves for NO2−concentration.

[0014] Figure 6. Optimization for experimental configurations. (a) The NO2−and NO3−yield rate under different applied power in 1 M KOH absorption solution. (b) The concentrations of NO2−and NO3−as a function of plasma running time in 1 M KOH absorption solution. TheAttorney Docket No.: 011520.01941 error bars represent the standard deviations of measurements obtained from three distinct samples, each prepared under the same conditions.

[0015] Figure 7. The output voltage waveforms for AC power supply.

[0016] Figure 8. The NO2−and NO3−yield rates under different flow rates in 1 M KOH absorption solution. The error bars represent the standard deviations of measurements obtained from three distinct samples, each prepared under the same conditions.

[0017] Figure 9. Comparison of1H NMR spectra under different experimental configurations. The blue line represents plasma itself cannot generate NH4+. The purple line verifies that N species produced by plasma transport through the catalysts (without current density) cannot be transformed into NH4+. The pink line demonstrates that N species produced by plasma transport through the catalysts (with current density) can be successfully transformed into NH4+.

[0018] Figure 10.1H NMR spectra of obtained NH4+using gaseous14N2and15N2as the source to generate N species for the subsequent electrocatalytic process.

[0019] Figure 11. Reaction graph from N2to NH3. Each node represents one reactant, intermediate or products; and each edge represents one reaction. purple and red nodes stand for the species with two and one nitrogen atoms respectively. Red, purple, and the yellow edges indicate the hydrogenation / dehydrogenation, oxidation, and the dissociation reactions.

[0020] Figure 12. The hydrogenation steps of NH2*to NH3*on different alloys (a) Cu, (b) CuAg, (c) CuAu, (d) CuRu.

[0021] Figure 13. The hydrogenation steps of NO*to NOH*on different alloys (a) Cu, (b) CuAg, (c) CuAu, and (d) CuRu. This step is always regarded as the rate determined by previous research for the NRR with NO as the reactant.

[0022] Figure 14. The initial and transition states of each hydrogenation step with N2 as reactant on CuPd. (a) ^^∗ ^ ^^ ^^ ^ ^^ି → ^^ ^^∗ ^ ^^^^ି. (b) ^^ ^ ∗ ି ∗ ିଶ ଶ ଶ ଶ ^ ^ ^^ଶ^^ ^ ^^ → ^^ଶ^^ଶ ^ ^^^^ .(c) ^^ଶ^^∗ଶ ^ ^^ଶ^^ ^ ^^ି → ^^ଶ^^∗ଷ ^ ^^^^ି. (d) ^^ ∗ ି ∗ ିଶ^^ଷ ^ ^^ଶ^^ ^ ^^ → ^^ଶ^^ସ ^ ^^^^ . (e) ^^ଶ^^ସ∗^^^ ି ∗ ି ∗ ି ିଶ^^ ^ ^^ → ^^^^ଶ ^ ^^^^ଷ ^ ^^^^ . (f) ^^^^ଶ ^ ^^ଶ^^ ^ ^^ → ^^^^ଷ^^^^ ^ ^^^^ .Attorney Docket No.: 011520.01941

[0023] Figure 15. The initial state and transition state of each hydrogenation steps with NO2 as reactant on CuPd. Due to the hydrogenation steps of NO are incorporated in these steps, we don’t repeat it. (a) ^^^^∗ ^ ^^ ^^ ^ ^^ି → ^^^^ ^^∗ ^ ^^^^ି ∗ ି ∗ଶ ଶ ଶ . (b) ^^^^ଶ^^ ^ ^^ଶ^^ ^ ^^ → ^^^^ ^^^^^ି ^ ^^ ^^. (c) ^^^^∗ ^ ^^ ି ∗ ି ∗ ି ∗ ିଶ ଶ^^ ^ ^^ → ^^^^^^ ^ ^^^^ . (d) ^^^^^^ ^ ^^ଶ^^ ^ ^^ → ^^ ^ ^^^^ ^^^ଶ^^. (e) ^^∗ ^ ^^ଶ^^ ^ ^^ି → ^^^^∗ ^ ^^^^ି. (f) ^^^^∗ ^ ^^ଶ^^ ^ ^^ି → ^^^^∗ ିଶ ^ ^^^^ . (g) ^^^^ଶ∗^^^ ^^ ^ ^^ି → ^^^^ ^^^^ ^ ^^^^ିଶ ଷ .

[0024] Figure 16. The initial state and transition state of each hydrogenation steps withNH2OH as reactant on CuPd. (a) ^^^^ଶ^^^^∗ ^ ^^ଶ^^ ^ ^^ି → ^^^^ ∗ଶ ^ ^^^^ି ^ ^^ଶ^^. (b) ^^^^ଶ∗^^^ ^^ ^ ^^ି → ^^^^ ^^^^ ^ ^^^ ିଶ ଷ ^ .

[0025] Figure 17. The initial state and transition state of hydrogen evolution reaction(HER) on CuPd. (a) Volmer step ^^ଶ^^ ^ ^^ି → ^^∗ ^ ^^^^ି. (b) Heyrovsky step ^^∗ ^ ^^ଶ^^ ^^^ି → ^^ ^^^^ ^ ^^^^ି. (c) Taf ∗ ∗ଶ el step ^^ ^ ^^ → ^^ଶ^^^^.Figure 18. The optimized adsorption structures of intermediates for eNORR. (a) N2*, (b) N2H*, (c) N2H2*, (d) N2H3*, (e) N2H4*, (f) NO2*, (g) NO2H*, (h) NO*, (i) NOH*, (j) N*, (k) NH2OH*, (l) NH*, (m) NH2*, (n) NH3*, (o) H*on CuPd.

[0027] Figure 19. DFT and microkinetic simulation. (a) the mechanism of major products pf plasma to ammonia synthesis is depicted. The arrows indicate the rate-determining steps. (b) the activation barriers of the two rate-determining steps along different pathways to ammonia. (c) the microkinetic kinetic simulation about selectivity of ammonia relative to hydrogen for various species at pH = 14, with a potential of −0.18 V versus RHE. (d) the microkinetic kinetic simulation for rate of ammonia production when using different species as reactants, in comparison to when nitrogen gas is used as the initial reactant.

[0028] Figure 20. Pictures of (a) Cu foam and (b) Ni foam, SEM images of (c) Cu foam and (d) Ni foam.

[0029] Figure 21. (a) SEM images of crystalline substitutional CuPd bimetallic alloys under substitutional time of 20 min. (b) EDS mappings of Cu and Pd distributions. (c) SEM images of crystalline substitutional CuAu bimetallic alloys under substitutional time of 20 min. (d) EDS mappings of Cu and Au distributions.Attorney Docket No.: 011520.01941

[0030] Figure 22. (a) SEM images of crystalline substitutional CuAg bimetallic alloys under substitutional time of 20 min. (b) EDS mappings of Cu and Ag distributions. (c) SEM images of crystalline substitutional CuRu bimetallic alloys under substitutional time of 20 min. (d) EDS mappings of Cu and Ru distributions.

[0031] Figure 23. (a) SEM images of crystalline substitutional CuPd bimetallic alloys under substitutional time of 10 min. (b) EDS mappings of Cu and Pd distributions. (c) SEM images of crystalline substitutional CuPd bimetallic alloys under substitutional time of 30 min. (d) EDS mappings of Cu and Pd distributions.

[0032] Figure 24. The high-resolution Cu 2p XPS spectra of Cu foam.

[0033] Figure 25. High−resolution XPS characterization of different crystalline substitutional surface alloys catalysts under substitutional time of 20 min. (a) Au 4f and (b) Cu 2p XPS spectra of CuAu alloy. (c) Ag 3d and (d) Cu 2p XPS spectra of CuAg alloy. (e) Ru 3d and (f) Cu 2p XPS spectra of CuRu alloy. The atomic percentages of Au, Ag, and Ru in alloys are 19.25%, 17.34%, and 5.68%, respectively.

[0034] Figure 26. High-resolution XPS characterization of different crystalline substitutional CuPd surface alloys as a function of reaction time. (a) Pd 3d and (b) Cu 2p XPS spectra of CuPd-10 min. (c) Pd 3d and (d) Cu 2p XPS spectra of CuPd-20 min. (e) Pd 3d and (f) Cu 2p XPS spectra of CuPd-30 min. The atomic percentages of Pd in CuPd-10min, CuPd-20min, and CuPd-30min alloys are 11.92%, 15.62%, and 22.27%, respectively.

[0035] Figure 27. The JV curves for different alloy catalysts under the voltage range of −0.5~−3.0 V in MEA. (a) Ni foam, Cu foam, and CuRu bimetallic alloys. (b) CuAu, CuAg, and CuPd bimetallic alloys. The chronoamperometry experiment was applied for 3 mins and the current densities were averaged for the last minute. The error bars represent the standard deviations of measurements obtained from three distinct samples, each prepared under the same conditions.

[0036] Figure 28. Evaluation of N-species reduction catalysts and conversion mechanisms. (a) comparison of NH4+production rate of different catalysts after 30 min at −200 mA cm-2in MEA, with an air flow rate of 400 sccm. (b) the long-term stability of our single pass plasma-electrochemical process with time-dependent current curve and NH4+yieldAttorney Docket No.: 011520.01941 rate in a MEA at −3.0 V over 1000 h. CuPd bimetallic alloy was applied as the catalyst. (c) comparison of NH4+yield rate at different voltages in the aqueous solution of 1 M KOH with 0.5 M NaNO3, NaNO2, and NH2OH, respectively. (d) the in-situ Raman spectra at different voltages for CuPd bimetallic alloys in the electrolyte of 1 M KOH with 0.5 M NaNO3. (e) the in- situ Raman spectra at different voltages for Cu foam in the electrolyte of 1 M KOH with 0.5 M NaNO3. The error bars represent the standard deviations of measurements obtained from three distinct samples, each prepared under the same conditions.

[0037] Figure 29. Comparison of NH4+production rate of different bimetallic alloy catalysts (a) CuPd catalysts with different Pd contents, and (b) CuAu catalysts with different Au contents, after 30 min at −400 mA cm−2. The error bars represent the standard deviations of measurements from three distinct catalyst samples, each prepared under the same conditions.

[0038] Figure 30. Schematic diagram of the long stability test by single-pass plasma- electrocatalyst integrated system.

[0039] Figure 31. (a) the in-situ Raman spectra at different voltages for CuPd bimetallic alloys in the electrolyte of 1 M KOH with 0.5 M NaNO2. (b) the in-situ Raman spectra at different voltages for Cu foam in the electrolyte of 1 M KOH with 0.5 M NaNO2.

[0040] Figure 32. (a) the in-situ Raman spectra at different voltages for CuPd bimetallic alloys in the electrolyte of 1 M KOH with 0.5 M NH2OH. (b) the in-situ Raman spectra at different voltages for Cu foam in the electrolyte of 1 M KOH with 0.5 M NH2OH.

[0041] Figure 33. (a) the in-situ Raman spectra at different voltages for bimetallic alloys in the electrolyte of 1 M KOH with 0.5 M N2H4. (b) the in-situ Raman spectra at different voltages for Cu foam in the electrolyte of 1 M KOH with 0.5 M N2H4.

[0042] Figure 34. Table 1: All intermediates used for graph theory from N2 to NH3.

[0043] Figure 35. Illustrations of Haber-Bosch (top) and plasma & electrochemical (bottom) production of NH3.

[0044] Figure 36. Illustrations of various components of an embodiment of the present disclosure.

[0045] Figure 37. Table 3: Comparison of reported ammonia synthesis.Attorney Docket No.: 011520.01941

[0046] Figure 38. CV curves of different electrocatalysts in a MEA electrolyzer. a, Ni. b, Cu. c, CuRu. d, CuAg. e, CuPd. f, CuAu, after 10 cycles. The CV curves for different catalysts in the MEA electrolyzer align with the j-V curves, showing peaks around -1.0 V corresponding to the redox reactions of Cu or Ni. Insets show zoomed-in views of the CVs, with the oxidation / reduction peaks of the metal catalysts labelled.

[0047] Figure 39. Comparison of the HER Faradaic efficiency of different electrocatalysts at −200 mA cm−2, with an air flow rate of 400 sccm.

[0048] Figure 40. a, EIS spectra for the different electrocatalysts tested in experimental studies. b, zoomed-in image of a. Both Rs and Rct values for all catalysts shows similar values suggesting that the reaction kinetics rate are similar.

[0049] Figure 41. a. Typical CO2 reduction with no plasma. b. CO2 reduction reaction with plasma pre-activation (overall reaction yield improved ~ 20 %). c. Current density graph of plasma vs. no plasma CO2 reduction reaction. With plasma pre-activation step, current density of the reaction was higher.

[0050] Figure 42. A chart depicting a method according to another embodiment of the present disclosure. Detailed Description of the Disclosure

[0051] Embodiments of the present continuous-flow plasma-electrochemical system integrate a plasma reactor (e.g., for nitrogen fixation) and an electrochemical reactor for ammonia synthesis. This integrated approach leverages the advantages of both plasma processing and electrochemical reduction while mitigating their individual limitations.

[0052] The system operates by first converting atmospheric nitrogen into reactive nitrogen-containing species (NOx and NOxHy) in the plasma reactor, then electrochemically reducing these species to ammonia in the electrochemical reactor. This two-stage approach avoids the need for high-temperature, high-pressure conditions while achieving selectivity for ammonia production.

[0053] With respect to Figure 1a, in an aspect, the present disclosure may be embodied as a continuous-flow plasma-electrochemical system 10. Such a system may be used to produceAttorney Docket No.: 011520.01941 ammonia from atmospheric nitrogen. The system 10 includes a flow-type plasma reactor 20. The plasma reactor 20 is configured to receive nitrogen, for example, air (i.e., atmospheric nitrogen), and to generate reactive nitrogen species—an output comprising NOxand NOxHyspecies. It should be noted that the plasma reactor may be operated in liquid or gas phase (for convenience only, the present disclosure is generally described using examples of gas-phase operation, and such examples are not intended to be limiting).

[0054] The plasma reactor may be a dielectric barrier discharge (DBD) plasma reactor. With reference to Figure 2, the plasma reactor 220 may have an inner electrode 222 and an outer electrode 224 separated by a dielectric barrier 226. In this way, in various embodiments, when a current is applied between the electrodes (e.g., a high voltage alternating current), a non-thermal plasma is generated in the gap. The inner electrode may serve a dual function as both an electrical conductor and a plasma catalyst to enhance the nitrogen reactions. The inner electrode may be made from a metal such as, for example, aluminum, copper, titanium, tungsten, or nickel, or an alloy. The inner electrode may be coated with platinum, ruthenium, iridium, or palladium, or the like. The plasma reactor may have an outer electrode 224.

[0055] The plasma reactor may be configured with various flow patterns, for example, to optimize gas-plasma contact and enhance nitrogen fixation efficiency. For example, the plasma reactor may be configured with a flow pattern involving one or more spirals, porous packed particles, cascading plates, gyroid, or other flow pattern or combinations. In non-limiting examples, these flow patterns include spiral configurations that create helical gas flow paths, porous packed particle beds that increase surface area and residence time, cascading plate arrangements that may provide multiple plasma exposure zones, gyroid structures that create complex three-dimensional flow patterns, and other engineered geometries that may promote turbulent mixing and uniform plasma exposure. The residence time of gas in the plasma reactor is selected to maximize the conversion of nitrogen to reactive NOxand NOxHyspecies while minimizing energy consumption. Typical residence times range from milliseconds to seconds, depending on the specific reactor configuration and operating conditions. The gas pressure within the plasma reactor can be controlled between, for example, 1 atm to 10 atm (inclusive) via pressure regulators or flow restrictors. This may allow enhancement of plasma characteristics and nitrogen fixation efficiency for different operating conditions.Attorney Docket No.: 011520.01941

[0056] The system 10 includes an electrochemical reactor 30 (illustrated in an exploded view in Figure 1). The electrochemical reactor 30 is configured to receive the NOx and NOxHy species from the plasma reactor 20. For example, the electrochemical reactor may be fluidically coupled to the plasma reactor. The electrochemical reactor 30 is configured to generate an output comprising ammonia. The electrochemical reactor may be a membrane electrode assembly having an anode 32, a cathode 34, and a membrane 36 separating the anode 32 and the cathode 34. The membrane may be an anion exchange membrane, a cation exchange membrane, or a bipolar membrane.

[0057] The cathode may be a catalyst made up of a metal or an alloy. For example, the catalyst (cathode) may be made from copper-gold (CuAu), copper-palladium (CuPd), copper- silver (CuAg), or copper-ruthenium (CuRu). The catalyst may be a bimetallic alloy, a trimetallic alloy, or a high entropy alloy. For example, the catalyst may be a bimetallic alloy of CuPd, CuAu, CuAg, or CuRu supported on a conductive support. The conductive support is made from a conductive material, such as, for example, a carbon support or a copper support (e.g., Cu foam). The catalyst may be a metal or alloy compositions that exhibit high activity and selectivity for NOx reduction.

[0058] The anode may facilitate oxidation reactions, such as, for example, an oxygen evolution reaction, a hydrogen oxidation, a chloride oxidation, an organic oxidation, etc.

[0059] The membrane separates the anode and cathode compartments while allowing selective ion transport to maintain charge balance. The choice of membrane depends on the specific electrolyte system and operating conditions. For example, anion exchange membranes (AEMs) may be advantageous for alkaline electrolyte systems and allow hydroxide ion transport from cathode to anode. Cation exchange membranes (CEMs) may be advantageous for acidic or neutral electrolyte systems and allow proton transport from anode to cathode. Bipolar membranes, which comprise both anion and cation exchange layers, can be used to, for example, maintain different pH conditions at the anode and cathode, potentially optimizing the reactions at each electrode. These are example considerations, and other considerations may be used to select a suitable membrane in addition to or instead of these.

[0060] The NOx and NOxHy species generated in the plasma reactor may be introduced into the electrochemical reactor as part of an electrolyte. In some embodiments, the output of theAttorney Docket No.: 011520.01941 plasma reactor 220 is purged into a KOH solution 242 to create a catholyte for the subsequent electrochemical reaction (Figure 2).

[0061] In some embodiments, the system may include a humidifier 240 configured to humidify the air received by the plasma reactor 220 (Figure 2). The presence of water was shown to increase NH4+yield as further described below. Additionally, the system may include a manifold 244 for introducing additional gas or liquid components to tune the plasma reaction outcome. These additional components may include water vapor, argon, carbon dioxide (CO₂), methane (CH₄), hydrogen gas, oxygen gas, or other reactive or inert species that can modify plasma characteristics, alter reaction pathways, or enhance the formation of specific nitrogen- containing products. The humidification level and concentration of additional components may be selected to balance enhanced nitrogen fixation with potential negative effects such as increased energy consumption, reduced plasma stability, etc.

[0062] In another aspect, the present disclosure may be embodied as a method 100 for generating ammonia. The method 100 includes introducing 103 nitrogen (e.g., atmospheric nitrogen—air) into a flow-type plasma reactor. For example, the plasma reactor may be a dielectric barrier discharge plasma reactor. The plasma reactor is used to generate 106 an output stream containing NOx and NOxHy species. The output stream (containing NOx and NOxHy species) is provided 109 to an electrochemical reactor, which is configured to generate ammonia as an output. The plasma reactor and the electrochemical reactor may be of the types described above, and may use components (e.g., catalysts, anodes, membranes, etc.) as described above.

[0063] In some embodiments, the method 100 further includes humidifying 112 the air before introduction 103 into the plasma reactor. In some embodiments, the method 100 further includes mixing 115 additional gas or liquid components into the input air stream. For example, a manifold may be provided to facilitate such mixing.

[0064] In some embodiments, the method 100 further includes purging 118 the output stream of the plasma reactor into a KOH solution to create a KOH electrolyte.

[0065] Although described with reference to ammonia generation, the present plasma- electrochemical reactor system and method may apply to other reactions as well (e.g., redox reactions, such as, for example, carbon dioxide reduction, methane oxidation, etc.) For example, Figure 41 includes plots showing a CO2 reduction reaction and the advantage of using plasmaAttorney Docket No.: 011520.01941 pre-activation (~20% improvement in yield). Suitable catalysts may be determined by methods including, for example, the graph theory approach described herein.

[0066] As such, the system may be described as a continuous-flow plasma- electrochemical system for redox reactions includes a flow-type plasma reactor configured to receive feedstock and to generate an output comprising intermediate species. An electrochemical reactor is configured to receive the intermediate species from the plasma reactor and to generate an output having the desired product (e.g., ammonia). The method may be similarly generalized.

[0067] Additional Discussion

[0068] The following discussion provides additional details of the present disclosure and of particular experimental embodiments. Configurations, materials, ranges, values, and other particulars are provided to illustrate certain embodiments and are not intended to be limiting.

[0069] Embodiments of the present disclosure provide a continuous-flow plasma- electrochemical reactor that converts air directly into ammonia at room temperature. In some embodiments, only water is used as an additional reactant. In some embodiments, a dielectric barrier discharge reactor is used to process nitrogen continuously, generating a mixture of NOxspecies that flows directly into a membrane-electrode-assembly (MEA) style electrochemical reactor. Both setups are easily scalable and tunable. To navigate the complex chemistry of plasma reactions, we use a graph theory approach to enumerate all possible nitrogen species and their pathways towards NH₃, identifying key intermediates. Density functional theory (DFT) calculations and microkinetic modeling were then employed to identify the optimal catalyst surface for binding the NOx species mixture to maximize the yield of ammonia. Experimental demonstration of our plasma-electrochemical reactor achieves an ammonia production rate of 81.2 mg h⁻¹ cm⁻², two orders of magnitude higher than normal electrochemical nitrogen reduction alone. More significantly, our starting material is atmospheric nitrogen, eliminating the need for specialized or purified nitrogen gas. To the best of our knowledge, this is the first report of such a continuous air-to-ammonia conversion reactor in room condition. This system design opens a new path for high-rate ammonia production at room temperature and has the potential to revolutionize other challenging chemical reactions, paving the way for new chemical synthesis routes.

[0070] Plasma-electrochemical Reactor Design 12 000160.01237 Business 9990964v1Attorney Docket No.: 011520.01941

[0071] An embodiment of the present plasma-electrochemical system is depicted in Figure 1a and Figure 2. Moist air was firstly activated by a dielectric barrier discharge plasma reactor and converted into diverse NOxand NOxHyspecies, then purged into 1 M KOH solution, which was directly used as catholyte for the subsequent electrochemical reaction in a MEA reactor with a Cu catalyst. The NH4+yield rate was quantified by1H nuclear magnetic resonance (NMR). A proof-of-concept system was built and used different gas compositions to determine the effectiveness of the presently disclosed technique (Figure 1b). With pure dried N2, no detectable amount of NH4+was found. In comparison, NH4+could be produced with either O2 or H2O added to the mixture, and the NH4+yield rate reached the highest during the experiments, 20.4 mg h−1cm−2, with humidified air. This result suggests that plasma-activated N*alone are too short-lived, and the addition of activated O*and H*via O2and H2O can generate various NOx and NOxHy radicals that can be converted by the electrochemical reactor. We further examined the NH4+yield rates with different N2 / O2ratios (Figure 1c), by controlling the N2and O2 flow rate via a mass flow controller. The highest ammonia production rate during the experiment was achieved under N2:O2 = 4:1, closely resembling the composition of air. Therefore, moist air was chosen as the feed gas for plasma reactors in the subsequent experiments. The use of air is also advantageous for cost savings of the system.

[0072] To qualify the molecular fragments after the plasma reactor, gas chromatography- mass spectrometry (GC-MS) was used. Figure 1d shows the molecular compounds detected by injecting gas samples directly into the GC-MS, including NO, N2O, NO2, N2O2, and N2H4. Since moisture was present in the plasma reactor, the outlet gas was also collected in liquid phases and N2H4, NH2OH and H2O2 were detected. Fourier Transform Infrared Spectroscopy (FTIR) (Figure 3) and UV-vis (Figures 4-5) further verified the existence of NO2−and NO3−. Considering the gaseous NxOy can be immediately dissolved in KOH solutions to form NO2−and NO3−, we quantified the concentration of NO2−and NO3−via UV-vis to explore advantageous plasma configurations. The total concentration of NO2−and NO3−was found to increase almost linearly with the applied plasma power (Figure 6a) and operation time (Figure 6b), suggesting the stability of the plasma reactor. The output voltage waveform for the plasma reactor is shown in Figure 7. In comparison, the NOx−yield rate reached a maximum at a flow rate of 400 sccm (Figure 8) and tapered off to a steady state upon further increasing the air flow rate. Consequently, moist air with a flow rate of 400 sccm was chosen as the feed gas for the plasma system for subsequent study.Attorney Docket No.: 011520.01941

[0073] We also performed a series of control and isotope experiments to confirm the authenticity of the NH3 produced. With the plasma reactor alone or with the plasma reactor plus electrochemical reactor with no current, no ammonia was detected in NMR (Figure 9), illustrating that NH4+was not coming from the plasma alone or background contamination. Furthermore, when15N2isotope was used as the reactant, only15NH4+was detected in the1H NMR spectra (Figure 10), further confirming the authenticity of the products.

[0074] As further described below and throughout this disclosure, in a first aspect, the present disclosure may be embodied as a system for generating ammonia. The system includes a plasma reactor. For example, the plasma reactor may be a dielectric barrier discharge plasma reactor. The plasma reactor is configured to receive air. In some embodiments, the air is humidified. For example, the air may have a moisture content of greater than 30%, 40%, 50%, 60%, 70%, or 80%, although the moisture content may be higher, lower, or between these values. The system may include a humidifier configured to humidify the air received by the plasma reactor (e.g., to humidify the air before it reaches the plasma reactor). The plasma reactor is configured to generate an output comprising NOxand NOxHyspecies.

[0075] The system includes an electrochemical reactor. The electrochemical reactor is configured to receive the NOx and NOxHy species from the plasma reactor and to generate an output comprising NH3. The electrochemical reactor may be a membrane electrode assembly. The membrane electrode assembly may have a catalyst comprising or consisting of Cu metal or a Cu alloy. For example, the catalyst may include or may be made up of CuAu, CuPd, CuAg, or CuRu. In some embodiments, the catalyst is a bimetallic alloy, such as, for example, a bimetallic alloy of CuPd, CuAu, CuAg, or CuRu on Cu foam.

[0076] The electrochemical reactor (e.g., membrane electrode assembly) may receive the NOxand NOxHyspecies as an electrolyte (e.g., a catholyte). For example, the output of the plasma reactor may be purged into a KOH solution to create a KOH electrolyte. The membrane electrode assembly may then receive the NOxand NOxHyspecies as the KOH electrolyte.

[0077] In another aspect, the present disclosure may be embodied as a method for generating ammonia. The method includes passing air into a plasma reactor configured to generate an output comprising NOx and NOxHy species. The plasma reactor may be a dielectric barrier discharge plasma reactor. The method may include humidifying the air before passing into the plasma reactor. For example, the air may have a moisture content of greater than 30%,Attorney Docket No.: 011520.01941 40%, 50%, 60%, 70%, or 80%, although the moisture content may be higher, lower, or between these values.

[0078] The NOx and NOxHy species from the plasma reactor are provided as an input to an electrochemical reactor (e.g., a membrane electrode assembly). The electrochemical reactor is configured to generate an output comprising NH3. The membrane electrode assembly may have a catalyst comprising or consisting of Cu metal or a Cu alloy. For example, the catalyst may include or may be made up of CuAu, CuPd, CuAg, or CuRu. In some embodiments, the catalyst is a bimetallic alloy, such as, for example, a bimetallic alloy of CuPd, CuAu, CuAg, or CuRu on Cu foam.

[0079] The electrochemical reactor (e.g., membrane electrode assembly) may receive the NOx and NOxHy species as an electrolyte (e.g., a catholyte). For example, the output of the plasma reactor may be purged into a KOH solution to create a KOH electrolyte. The membrane electrode assembly may then receive the NOx and NOxHy species as the KOH electrolyte. Graph theory-based conversion pathways

[0080] While the activation of nitrogen gas via plasma is effective in breaking the nitrogen triple bond, it also generates a wide variety of products, which are challenging to utilize as a reactant mixture. A solid understanding of the competing reaction pathways is necessary for rational catalyst design. Given the numerous combinations of N-O-H post plasma reactor, we employed a graph theory approach to systematically explore all potential intermediates and reaction pathways from nitrogen to ammonia. Similar work has been performed for the CO2 reduction pathway. This comprehensive approach will map out the complex network of reactions, providing a clearer understanding of the process to guide catalyst development.

[0081] We began from N2species and analyzed all possible chemical intermediates up to a maximum of N ≤ 2, H ≤ 6, and O ≤ 4 atoms using CATKIT. With this approach, we obtained the structures, formula, and the number of electrons transferred to N2of a total of 102 intermediates through the N2-to-ammonia reaction. Table 1 (Figure 34) provides a list of the resulting intermediates. To determine the reaction pathway, we chose the most unsaturated atom on a particular intermediate as the binding atom and calculated the adsorption energy of the intermediate on the atop, bridge, hcp hollow, and fcc hollow binding site on the Cu (100) surface, using an adsorption vector algorithm. Following the optimization using densityAttorney Docket No.: 011520.01941 functional theory (DFT), we checked the intermediates to ensure that there was no dissociation or reconstruction, and any dissociated or reconstructed species were removed from the species list. This step was performed to remove any nonphysical intermediates under the N ≤ 2, H ≤ 6, and O ≤ 4 combinations. Through this computational routine, we obtained the full reaction network from N2to NH3, shown in Figure 11. Each node in the graph represents a reaction intermediate and each edge represents a possible reaction pathway. From the network diagram, it can be seen that the intermediates from nitrogen to ammonia undergo the step of NH2hydrogenation to NH3. Thus, for the nitrogen reduction to ammonia, NH2* is an advantageous intermediate and an efficient catalyst should have favorable binding to NH2*. At the same time, the nitrogen-hydrogen-oxygen intermediates undergo the transformation to NO* before further reaction to ammonia. Thus, NO* may also be a beneficial intermediate product among all the nitrogen-hydrogen-oxygen species that requires consideration during catalyst design.

[0082] Despite the complexity of the products mixture from a plasma reactor, we have identified that the NOx or NOxHy species analyzed go through NO* or NH2* as an intermediate step before reaching NH3. Thus, a catalyst with favorable binding to NO* and NH2* should give advantageous catalytic activity for NH3. Density Function Theory (DFT) and microkinetic simulations

[0083] DFT and microkinetic simulations were used to identify suitable electrocatalysts for the experimental investigation. Based on the bottleneck intermediates identified from graph theory, we focused on two essential hydrogenation steps: NOH*-to-N* and NH2*-to-NH3, both are common key steps for most intermediates. A Cu catalyst was used as a base model as it was used during our initial testing, and it is considered to be an advantageous electrocatalyst for nitrate reduction. To address the weak bonding interactions between NOx − and Cu atoms, we also screened various Cu-based alloy catalysts (CuAg, CuAu, CuPd, CuRu) to modify the surface electronic structure and create bifunctional surface adsorption sites. The optimized structure for each hydrogenation step for each alloy catalyst is shown in Figures 12−18.

[0084] The first step was to elucidate the underlying mechanism in the nitrogen reduction reaction (NRR) process. Recognizing that NO2converts to NO during the NRR, the reaction rate for NO2 must be less than or equal to that of NO. Therefore, we used NO as a proxy for both NO and NO2. Figure 19a presents a comprehensive diagram of reduction pathways starting from N2, NO, NH2OH, and N2H4. Notably, converting nitrogen to ammonia requires three hydrogenationAttorney Docket No.: 011520.01941 steps, with the initial step necessitating significant energy to overcome the nitrogen-nitrogen triple bond barrier. The transition from NO to ammonia involves five hydrogenation steps. For N2H4to ammonia conversion, two hydrogenation steps were needed, each producing one ammonia molecule. Similarly, transitioning from NH2OH to NH3 used two hydrogenation steps for each ammonia molecule. The energy barriers of each Cu-derived catalyst are illustrated in Figure 19b Among the selected alloys, CuAu outperformed others in both reactions, followed by CuPd and then CuAg, while pure Cu and the CuRu alloy exhibited the least efficacy.

[0085] In an in-depth analysis of each product, we calculated the transition state barriers and ΔG for the entire pathway from NO, N2, NH2OH, and N2H4to NH3, aiding microkinetic simulations. All energy barriers and corresponding energies are detailed in Figures 12-18 and Table 1 (Figure 34), where we also compare them against the hydrogen evolution reaction (HER) energy (Figure 19c). We observed that in the N2 to NH3 conversion, the largest barrier to break is the nitrogen-nitrogen triple bond, leading to low NRR against HER. For NO, NH2OH, and N2H4hydrogenation to NH3, the energy barriers of their respective hydrogenation steps are low, effectively suppressing HER and boosting NH3selectivity. Further exploration of reaction kinetics revealed that NH2OH, due to its less required hydrogenation step and low barrier, exhibited the highest reactivity, NO closely follows, and N2H4lags behind NO due to the extra N-N bond-breaking step (Figure 19d). The turnover frequency for N2 is several orders of magnitude lower than that of the main plasma products. Thus, our theoretical insights suggest that future ammonia synthesis from air should focus on enhancing the selectivity of NH2OH from plasma, owing to its superior selectivity and activity among all intermediates from N2to NH3. Electrocatalytic performance

[0086] To further improve the plasma-electrochemical system, a range of bimetallic alloy catalysts, including CuPd, CuAu, CuAg, and CuRu on Cu foam, were synthesized via a galvanic substitution method according to the computational results. Representative scanning electron microscopy (SEM) and energy-dispersive X-ray spectroscopic (EDS) images (Figures 20-23) of the as-synthesized catalysts confirmed the existence of homogeneously distributed bimetallic alloy on the Cu foam substrate. X-ray photoelectron spectra (XPS) were further employed to analyze the chemical composition of the bimetallic alloy catalysts (Figures 24 - 27), determining the atomic contents (Table 2 below) of Pd, Au, Ag, and Ru as 15.6, 19.3, 17.3, and 5.7 at. %,Attorney Docket No.: 011520.01941 respectively. The Cu foam showed Cu 2p peaks at 932.3 and 952.0 eV in a 2:1 area ratio, which could readily be assigned the Cu 2p3 / 2 and Cu 2p1 / 2 signals, respectively, of metallic Cu. For the CuAu and CuPd alloy catalysts, the Cu 2p peaks were shifted to slightly higher binding energies (a 0.2 eV positive shift) compared to that of Cu foam, revealing a change in surface electronic structure through alloying. The Au and Pd atoms in the alloy catalysts were determined by XPS to also be in metallic state. Table 2. The atomic percentages of different metal elements for bimetallic alloy catalysts are determined by SEM–EDS and XPS. SEM XPS

[0087] The j–V curves of the alloy catalysts are shown in Figure 27, which were tested similarly to Figure 1. The highest current densities of −574.1 mA cm−2at −3.0 V were obtained from the CuAu bimetallic alloy catalyst. Although Ni metal foam, as a control, had higher current densities than CuPd, CuAg, and CuRu catalysts at −3.0 V, Ni metal only generated a small amount of NH3, suggesting HER is the dominant reaction. The onset potential for CuPd, CuAu, CuAg, and CuRu catalysts positively shifted to above −1.2 V from −1.8 V compared to Ni foam, confirming a favorable energy barrier for NOxreduction compared to HER. This was further supported by the j-V curves tested with only a KOH electrolyte (plasma off) vs. KOH electrolyte with plasma products (Figure 38), where the CuRu, CuAg, CuPd, and CuAu catalysts all exhibited higher current density with a KOH electrolyte containing NOxHy products from the plasma reactor. Figure 39 confirms that over 95% of electrons are transferred toward NOx orAttorney Docket No.: 011520.01941 NOxHy reduction with minimal transfer to the hydrogen evolution reaction (HER) for the CuAg, CuAu, and CuPd electrocatalysts. As seen from Figure 28a, the CuAu bimetallic alloy catalyst showed the highest NH3production rate, 40.9 mg h−1cm−2, while CuPd was a close second, within the error bar range. The experimental ammonia production rate and HER FE closely match the trend predicted by Figure 19b, further validating the presently disclosed mechanism. By changing the galvanic reaction exchange time, the Pd surface content could be tuned between 10 and 20 at. % (Figure 29a). The maximum NH4+yield was achieved for Cu0.78Pd0.22at 81.2 mg h−1cm−2. Similar optimization was also done for CuAu, achieving the highest NH4+ production rate of 78.5 mg h−1cm−2(Figure 29b).

[0088] The N2 utilization in our plasma reactor, determined by the ratio of the input vs output air flow rate, is ca.1%. The N species conversion rate for the electrochemical reactor, determined by the ammonia production rate over the NOx and NOxHy input rates, is ca.0.1%.

[0089] To demonstrate the stability of the present plasma-electrochemical system, we tested constant voltage electrolysis via a single-pass process (illustrated in Figure 30). A KOH electrolyte with post-plasma N species was circulated to the MEA reactor in a single pass, and a 4 cm2CuPd catalyst was used. The plasma-electrochemical system was stable for over 1000 h (Figure 28b) with constant current response and NH4+yield rate. Other than a small burn-in period at the beginning, the raw cell current was stable around ca. −2 A, or 500 mA / cm2. The NH4+production rate also stabilized at ca.30 mg h−1cm−2. Table 3 (Figure 37) compares the recently reported state-of-the-art ammonia production systems, including NO3RR, lithium- mediated approaches, eNRR and plasma-assisted systems. The present NH4+yield rate was ~3 times higher than NOxRR with similar applied current density. Mechanism investigation

[0090] To explore the ammonia production mechanism, we first investigated the potential dependent NH4+product rates with NO3−, NO2−, N2H4, and NH2OH each as reactants individually (Figure 28c). As the potential increased from 0 to −0.6 V vs. RHE (reversible hydrogen electrode), the NH4+yield rates of NO3−, NO2−, and NH2OH were all gradually enhanced. At all applied potentials, NH2OH exhibited the highest NH4+yield rate, followed by NO2−and NO3−, demonstrating that preference-ordering of consumption for generated N-species is NH2OH, NO2−, and NO3−. The favorable selectivity of NH2OH is also consistent with computational results in Figure 19d. The potential dependent NH4+yield rates of N2H4 were alsoAttorney Docket No.: 011520.01941 tested but were generally very low. Given that N2H4can auto-decompose N2, and H2, the result may not be reliable and we did not include it in Figure 28c.

[0091] To further verify the presently disclosed mechanism from our computational model, in situ Raman spectroscopy was conducted in a mixed solution of 1 M KOH with 0.5 M NO3−, NO2−, NH2OH, or N2H4 respectively over Cu foam and CuPd alloy catalysts. Figures 28d and 28e show the Raman spectra in 1 M KOH and 0.5 M NO3−for CuPd and Cu respectively. The peak at 1040 cm−1can be assigned to NO3−. The Raman band at 798 cm−1, appeared at 0.36 V vs. RHE, can be assigned as NO2−bending. An additional metal-N-O bonds also appear at 0.36 V vs. RHE, 620 cm−1for CuPd catalyst which was not observed for Cu. In both Figures 28d and 28e, the NO3−peak was strong for both CuPd and Cu, suggesting that NO3−was well adsorbed on both surfaces. As the overpotential increases, the NO2−peak becomes visible on the CuPd alloy catalyst due to conversion of NO3−into NO2−. However, the NO2−peak was not observed on the Cu surface. These results would suggest that a higher NO2−surface concentration is present on the CuPd, making it a better catalyst. Similarly in Figure 31a, when 1 M KOH and 0.5 M NO2−are used as electrolyte, NO2−peak is similarly weaker on Cu surface compared to CuPd. For NH2OH reduction (Figure 32a), a peak for NH3 at 1200 cm−1is observed beginning at 0.21 V. vs RHE for CuPd but barely visible for Cu, which suggest again CuPd has higher catalytic activities compared to Cu. For N2H4 conversion (Figure 33a), three bands at 580, 1019, 1363 cm−1were assigned to N2H4. The NH3and HNH, peaks at 715 and 1317 cm−1, started to form at 0.15 V and increased with applied voltages. On the contrary, no obvious peak for Cu-N-H can be seen and the band for NH3was much weaker (Figure 33b) on the Cu surface. Overall, CuPd catalyst shows better catalytic activities and has higher surface concentration of intermediates over Cu in all the NO3−, NO2−, NH2OH, or N2H4reduction experiment. These results confirm our computational investigation that CuPd catalyst can efficiently bind to a mixture of NOxspecies coming from our plasma reactor. Experimental Example - Additional Details

[0092] The following are non-limiting details related to certain experiments performed to create and characterize example embodiments.Attorney Docket No.: 011520.01941 Materials

[0093] All chemicals were directly used without further purification. Anion exchange membrane (AEM) was purchased from Fuel Cell Store. Potassium hydroxide (KOH), Ruthenium chloride hydrate (RuCl3), gold chloride trihydrate (HAuCl4), potassium hexachloropalladate (K2PdCl6), and Griess reagent were purchased from Sigma Aldrich. H2SO4 was obtained from Baker Analyzed. Silver nitrate (AgNO3), hydrazine hydrate (N2H4), hydroxylamine (NH2OH), sodium nitrite (NaNO2), and sodium nitrate (NaNO3) were bought from Thermo Scientific. Cu foam and Ni foam were purchased from MTI Corp. Catalyst Preparations

[0094] The alloy catalysts were synthesized by a metal ion substitution method. Typically, a piece of Cu foam was immersed in 5 ml of 10 mM solution with K2PdCl6, RuCl3, HAuCl4, or AgNO3solution for 20 mins at room temperature. Then, the Cu foams with different surface alloys were dried overnight and then directly used as the working electrodes for the subsequent electrochemical process. Materials Characterization

[0095] The scanning electron microscopy (SEM) and energy-dispersive X-ray spectroscopic (EDS) images were carried out by a Carl Zeiss AURIGA CrossBeam. The ultraviolet-visible (UV) absorbance spectra were conducted on an Agilent Cary spectrometer. The 1H-NMR spectra were measured by a Bruker 500-MHz system. Fourier transform infrared spectroscopy (FTIR) was acquired using PerkinElmer Spectrum Two with Attenuated Total Reflectance Accessory. The elemental surface compositions of different alloys were investigated by X-ray photoelectron spectra (XPS) (PHI 5000 Versaprobe). In situ Raman Spectroscopy

[0096] In situ Raman spectroscopy was carried out with a StellarNet Raman spectrometer coupled with an Olympus microscope and a 785 nm laser light. The electrochemical process was conducted by a three-electrode system using CuPd or Cu foam as the working electrode, Ag / AgCl electrode as the reference electrode, and Ni wire as the counter electrode.1 M KOH with 0.5 M NO3−, NO2−, N2H4or NH2OH were employed as electrolytes.Attorney Docket No.: 011520.01941 Plasma Generated N-Fragments Experiments

[0097] The plasma experimental configurations were illustrated in Figure 2, which contains a power supply, a plasma reactor, and an oscilloscope. Our plasma reactor is made up of a glass tube serving as the dielectric barrier and two stainless steel tubes as the high-voltage electrodes. Humid air was utilized as the feeding gas for the plasma reactor. Post-plasma reactor gas was bubbled into a 1 M KOH solution, which was used directly as the electrolyte for the electrochemical process. Electrochemical Experiments

[0098] Unless otherwise noted, all electrochemical evaluations were performed at room temperature in a membrane-electrode-assembly configuration (MEA) with a Squidstat Plus Potentiostat (Admiral Instrument). All electrochemical experiments were carried out in triplicate to provide an error bar in the manuscript. For the MEA electrolyzer, alloy catalyst and Ni foam were applied as the cathode and anode, respectively, separated by an AEM. The reaction area of the electrode was set as 1 cm2. During operation, the plasma-generated species in 1 M KOH aqueous solution was used as catholyte. The JV curves of the different alloy catalysts were conducted by chronoamperometry experiment at different voltages for 15 minutes. The current densities reported were the averaged stable current densities in the last minute. The NH4+production rates in Figure 28a were performed at a current density of −200 mA cm−2, while the current density for different loading of the CuPd catalysts was at −400 mA cm−2.

[0099] The comparison of NH4+yield rate at different voltages in the aqueous solution of 1 M KOH with 0.5 M NaNO3, NaNO2, and NH2OH (Figure 28c) was conducted in a flow cell setting with Ag / AgCl reference electrode to provide accurate voltage comparison. A 1 M KOH solution was employed as the anolyte, while a 1 M KOH solution with 0.5 M NaNO3, NaNO2, or NH2OH was employed as the catholyte. The reaction area of cathode / anode was set as 1 cm2. The voltages shown in Figure 28c are iR-corrected voltages converted to RHE.

[0100] For the single-pass stability test (Figure 28b), the experimental setup is illustrated in Figure 30. A CuPd alloy catalyst and Ni foam were used as the cathode and anode, respectively, separated by an AEM. The catholyte was 1 M KOH solution with post plasma N species and it is not recycled back to the original container. The reaction area was expanded toAttorney Docket No.: 011520.01941 4 cm2and the NH4+production rate and corresponding current densities were collected under a constant voltage of −3 V. Quantitative determination of ion concentrations

[0101] NO3−quantification. To quantitatively determine the amount of NO3−, a series of NaNO3 solutions with known concentrations were prepared. NO3−shows absorbance to ultraviolet light at the wavelength of 220 and 275 nm. The absorbance values were calculated by the absorbance at 220 nm minus the absorbance at 275 nm. The corresponding calibration curves are displayed in Figure 4. The pH of plasma-generated solutions (in 1 M KOH) was first adjusted to pH = 7 by adding 1 M HCl before UV vis measurement.

[0102] NO2−quantification. The concentrations of generated NO2−are determined by Griess reagent. Typically, 5 ml standard solutions of NaNO2 with known concentrations or diluted plasma−generated solutions were added into the test tube, and then 0.1 ml Reagent A was added to the solution. After standing for 10 min in a dark place, 0.1 ml Reagent B was added to the mixed solution. Following shaking and left standing for 30 min in darkness, the concentration of NO2−was determined by UV-vis absorption spectra. Standard concentration- absorbance curves were acquired from a series of NaNO2solutions with known concentrations at the wavelength of 540 nm, and the corresponding calibration curves are shown in Figure 5.

[0103] NH4+quantification. The concentrations of NH4+were determined by1H NMR spectrometer. NMR samples were prepared by the addition of 600 μl electrolyte after electroreduction and 60 μl 9 M H2SO4 to the NMR tube. Then a coaxial insert containing 3.1 mM DMSO (dissolved in D2O) was placed into the NMR samples tube as reference. The1H spectra were collected with a water suppression method and 64 scans. The production rates of NH4+(mg h−1cm−2) were calculated according to the following formula: ^^ேுశ ∗ ^^ ா ∗ 1000 ^^^ெௌை ∗ ^^^ெௌை ∗ ^^ேெோಿಹశ^^ேெோ^^^^^^ ^^^ ^^శర∗ర∗^ାுమௌைర∗^ெௌைmgாrepresent the reaction time and reaction area, ^^^ெௌைand ^^ேுరశ correspond to the numbers of H in DMSO (6) and NH4+(4), ^^^ெௌைis the concentration of DMSO (3.1 mM), ^^ேெோಿಹశరand ^^ேெோವಾೄೀare the area integral value of1H NMR peaks detected for NH4+and DMSO,Attorney Docket No.: 011520.01941 ^^ேெோ^^^^^^^^and ^^ேெோ^^^^^^^ାுమௌைర^are the volume of NMR sample (600 μl) and the volume after pH adjustment (660 μl), ^^^ெௌைand ^^^^^^^^is ascribed to the volume of the stem coaxial insert (60 μl) and the volume of NMR sample tube minus the volume of the coaxial insert tube (530 μl). Isotope labeling experiments

[0105] For the isotope experiment, 99%15N2 was used as the N-source while the remaining setup is the same to previous tests. The15N2gas was connected to the plasma reactor at a flow rate of 10 sccm mixing with O2 gas at a flow rate of 2.5 sccm to create a roughly 4:1 mixture similar to our14N2setting. After electroreduction, the obtained electrolyte was characterized by1H NMR spectrometer also in the same setting. Graph theory methods

[0106] In our study, we utilized directed graphs to identify all possible simple paths from the reactant node to the product nodes. This analysis was conducted using a modified depth-first search algorithm, as implemented in the networkx package. This approach allowed us to enumerate the number of feasible simple pathways.

[0107] For the graph-based enumeration of intermediate structures, we employed a method integrated into the Catkit software. This method considers atoms in an intermediate as nodes and bonds as edges. The unsaturation level in a molecule was defined based on the coordination numbers of nitrogen, oxygen, and hydrogen. The degree of unsaturation was calculated by comparing these preset coordination numbers with the actual ones in the molecule. The site with the highest degree of unsaturation was classified as the most unsaturated.

[0108] Post-optimization, we compared the connectivity matrix of each intermediate with that of the initial structure. Differences in these matrices indicated structural reconstruction or dissociation in the intermediates. Intermediates that reconstructed or dissociated in all geometries were deemed unstable and subsequently removed from our node list. Correspondingly, reactions involving these unstable intermediates were also excluded.Attorney Docket No.: 011520.01941 DFT Simulation

[0109] In this study, we employed Density Functional Theory (DFT) calculations which were executed using the ab initio simulation program VASP, incorporating the Perdew-Burke- Ernzerhof (PBE) exchange-correlation functional. Electron-ion interactions were modeled using the Projector-Augmented Wave (PAW) method, and the plane-wave basis set was defined with a cut-off energy of 450 eV. To accurately represent long-range dispersion interactions between the adsorbates and catalysts, the D3 correction method by Grimme et al. was utilized. Brillouin zone integration was performed using a 3×3×1 Monkhorst-Pack k-point mesh. We optimized all adsorption geometries employing a force-based conjugate gradient algorithm. The crystal structure of copper was specifically optimized to a lattice parameter of 3.63 Å. For the surfaces Cu (100), we used a four-layer p(3×3) slab model, where the top two layers were relaxed and the bottom two layers were kept fixed. The initial structures of all intermediate states were generated using the Catkit package, a tool developed by Stanford University (available at CatKit GitHub Repository). All the adsorption geometries were optimized using a force-based conjugate gradient algorithm, while transition states (TSs) were located with a constrained minimization technique. A dipole correction was applied to decouple the electrostatic interaction between the periodically repeated slabs. Gibbs free energy correction was applied using the harmonic oscillator approximation. Ab initio molecular dynamics simulations

[0110] As an electrochemical interface is very complex, it is difficult to confirm that the explicit water structure is unique and the most stable one. Therefore, we performed ab initio molecular dynamics calculations at 298.15 K and 10 ps (10,000 steps) to generate a series of electrode–water interfaces using a slab of all alloys covered with a thick layer of water of ~10 Å (denoted as a multilayer). The ion effects are also taken into consideration in the same methods with previous research. Electrochemical barrier calculations

[0111] We utilized the computational hydrogen electrode model to calculate reaction energies as a function of potential. At a potential of U = 0 V versus RHE, protons and electrons are in equilibrium with hydrogen gas (H2) under standard conditions of 101325 Pa and 298 K, applicable across all pH levels:Attorney Docket No.: 011520.01941 ^^ା ^ ^^ି → ^^ଶ^^^

[0112] At a given U≠ 0V versus RHE, ^^^^^ା^ ^ ^^^^^ି^1 ൌ 2^^^^^ଶ^ െ ^^^^

[0113] Therefore, the free energy change (take the generation of NOH* as example) of the proton-electron transfer can be calculated through: ∆^^ ൌ ^^^^^^^^^^ െ ^^^^^^^^ െ ^1 2^^^^^ଶ^ െ ^^^^^

[0114] As andifficult to confirm that the explicit water structure is unique and the most stable one, while a hydrogen-bonding network is needed along the active site for proton transfer (H2O is considered as the proton donor in alkaline conditions). For each initial reactants (N2, NO2, and NH2OH), we performed the long-time minima-hopping simulation to obtain the stable structures. Steady-state equation

[0115] The rate equation for step i, NxHyOz* +H+↔ NxHy+1Oz*, can be written as: ^^^ ൌ ^^^^^ே^ு^ை^^^^ା^ െ ^^ି^^^ே^ு^శభை^

[0116] Here, the rate of each step, denoted as ^^^, is defined as the difference between the forward and reverse reaction rates. The surface coverages of NxHyOz* and NxHy+1Oz* are represented as ^^ே^ு^ை^and ^^ே^ு^శభை^respectively. The forward reaction rate constant, ki,and the reverse reaction rate constant, k-i, are both derived from transition state theory (TST), which is given by: kൌ^^^^^ ^^ି∆ீಯ,° / ^ಳ்e ∆^^ஷ,wher°is the change of energy barrier at 0 V vs. SHE. We describe the favorability of a reaction step using the reaction energies. For a reaction with energy∆^^°, the forward and reverse reaction barrier are determined as max ^∆^^°, 0^ andmax ^െ∆^^°, 0^, respectively.Attorney Docket No.: 011520.01941

[0117] We employed the steady-state approximation in our model, which posits that the rate of change in the surface coverage of all species X is zero at steady-state: ^^^^௫^^^^ൌ ^ ^^^ െ ^ ^^ ^ ൌ 0

[0118] In this context, ^^^reactions generating andconsuming the species X, respectively.

[0119] With all the coverages above, the coverage of free sites is obtained using: ^^∗ ൌ 1 െ ^^^^

[0120] For our kineticand activation energies were calculated at a coverage of 1 / 9 ML, reflecting a low coverage regime. For the steady-state solver, we adopted the CATKINAS. For the concentration / partial pressure of reactants and products, the partial pressures of reactant and products are 0.1 bar and 10−20bar, respectively, to describe the reaction condition at the beginning. The partial pressure of liquid water was determined using the ideal gas approximation at its vapour pressure at room temperature, 3534 Pa (3.49×10−2bar). Conclusion

[0121] In conclusion, we have demonstrated a continuous-flow plasma-electrochemical system capable of reducing atmospheric nitrogen to ammonia. Atmospheric nitrogen is directly activated into a mixture of NOx species. Graph theory and microkinetic modeling guided the design of a CuPd catalyst that is able to stabilize key intermediates in the NOxmixture for electrochemical reaction. Experimental results show a high ammonia production rate sustained for over 1000 hours, directly from air. In some embodiments, the present plasma-electrochemical system can be configured into modular units for distributed chemical production in remote locations. Additionally, this system could be applicable for other challenging chemical reactions, such as methane oxidation or CO2reduction. The findings of this study provide a new approach for the green synthesis of chemicals, addressing global environmental challenges.

[0122] Although the present disclosure has been described with respect to one or more particular embodiments, it will be understood that other embodiments of the present disclosure may be made without departing from the spirit and scope of the present disclosure. The termAttorney Docket No.: 011520.01941 “high entropy alloy” is a term of art, which includes alloys having multiple principal elements in relatively high concentrations (e.g., rather than alloys made up of relatively small amounts of secondary elements added to a primary element). For example, the principal elements (typically five or more elements, but may be fewer) are often present in equal amounts or relatively large proportions.

Claims

Attorney Docket No.: 011520.01941 What is claimed is:

1. A continuous-flow plasma-electrochemical system for producing ammonia from atmospheric nitrogen, comprising: a flow-type plasma reactor configured to receive air and generate an output comprising NOxand NOxHy species; and an electrochemical reactor configured to receive the NOxand NOxHyspecies from the plasma reactor and generate an output comprising ammonia.

2. The system of claim 1, wherein the flow-type plasma reactor is a dielectric barrier discharge plasma reactor.

3. The system of claim 2, wherein the flow-type plasma reactor comprises an inner electrode to act as a plasma catalyst and an outer electrode.

4. The system of claim 3, wherein the inner electrode comprises aluminum, copper, titanium, tungsten, or nickel and / or is coated with platinum, ruthenium, iridium, or palladium.

5. The system of claim 1, wherein the electrochemical reactor is a membrane electrode assembly having an anode, a cathode, and a membrane.

6. The system of claim 5, wherein the cathode is a catalyst comprising metal or an alloy.

7. The system of claim 6, wherein the catalyst comprises copper-gold (CuAu), copper-palladium (CuPd), copper-silver (CuAg), or copper-ruthenium (CuRu).

8. The system of claim 6, wherein the catalyst is a bimetallic alloy, a trimetallic alloy, or a high entropy alloy.

9. The system of claim 6, wherein the catalyst is a bimetallic alloy of CuPd, CuAu, CuAg, or CuRu supported on a conductive support.

10. The system of claim 9, wherein the conductive support is Cu foam or a carbon support.

11. The system of claim 5, wherein the membrane is an anion exchange membrane, a cation exchange membrane, or a bipolar membrane.Attorney Docket No.: 011520.01941 12. The system of claim 1, wherein the electrochemical reactor receives the NOxand NOxHyspecies as an electrolyte.

13. The system of claim 1, wherein the output of the plasma reactor is purged into a KOH solution to create a KOH electrolyte; and wherein the electrochemical reactor receives the NOxand NOxHy species as the KOH electrolyte.

14. The system of claim 13, wherein the system suppresses the hydrogen evolution reaction (HER) with over 95% of electrons transferred toward NOx or NOxHy reduction.

15. The system of claim 1, further comprising a humidifier configured to humidify the air received by the plasma reactor.

16. The system of claim 1, wherein the electrochemical reactor is configured to receive the mixture of NOxand NOxHyspecies and electrochemically reduce the species to ammonia using water as a proton source 17. A method for generating ammonia, the method comprising: introducing air into a flow-type plasma reactor; generating, using the plasma reactor, an output stream comprising NOxand NOxHyspecies; and providing the output stream having NOxand NOxHyspecies to an electrochemical reactor configured to generate an output comprising ammonia.

18. The method of claim 17, wherein the plasma reactor is a dielectric barrier discharge plasma reactor.

19. The method of claim 17, wherein the electrochemical reactor is a membrane electrode assembly.

20. The method of claim 19, wherein the membrane electrode assembly has a catalyst comprising Cu metal or a Cu alloy.

21. The method of claim 20, wherein the catalyst comprises CuAu, CuPd, CuAg, or CuRu.

22. The method of claim 20, wherein the catalyst is a bimetallic alloy.Attorney Docket No.: 011520.01941 23. The method of claim 22, wherein the catalyst is a bimetallic alloy of CuPd, CuAu, CuAg, or CuRu on Cu foam.

24. The method of claim 17, wherein the electrochemical reactor receives the NOx and NOxHy species as an electrolyte.

25. The method of claim 17, further comprising purging the output of the plasma reactor into a KOH solution to create a KOH electrolyte; and wherein the electrochemical reactor receives the NOx and NOxHy species as the KOH electrolyte.

26. The method of claim 17, further comprising humidifying the air before introducing into the plasma reactor.

Images