Methods and systems for hydrogen production

Abstract

A portable device for generating hydrogen from ammonia includes a first reactor layer having an ammonia inlet a retentate port, and a chamber containing an ammonia decomposition catalyst. A first gas-collecting layer has a manifold with a hydrogen outlet. A first hydrogen-selective membrane is disposed between the first reactor layer and the first gas-collecting layer. In this way, hydrogen gas generated in the chamber of the first reactor layer will permeate through the hydrogen-selective membrane into the manifold of the first gas-collecting layer. A burner layer is adjacent to the first reactor layer and separated from the chamber by a first conduction plate. The burner layer includes an intake port and an exhaust port. The intake port is in fluid connection with the retentate port of the first reactor layer.

Description

METHODS AND SYSTEMS FOR HYDROGEN PRODUCTION Statement Regarding Federally Sponsored Research

[0001] This invention was made with government support under grant no. DE- AR0001479 awarded by the Department of Energy. The government has certain rights in the invention. Field of the Disclosure

[0002] The present disclosure relates to techniques for generation hydrogen from Ammonia. Background of the Disclosure

[0003] The global H2generation market size is predicted to rise from a value of $242.7 billion in 2023 to $410.6 billion by 2030, expanding at a CAGR of 7.8% [Hydrogen Market, Global Industry Size Forecast, [Latest] (marketsandmarkets.com)]. The rising demand for renewable sources of energy and low-carbon fuels, for decarbonized energy end-use, government regulation for desulphurization of refinery activities, and increased demand for H2in the transportation sector are projected to boost market growth.

[0004] Currently, greater than 95% of the H2used commercially is produced by the steam reforming (SR) of a fossil source such as naphtha or CH4 (SMR) followed by various cleanup steps. IEA [Executive summary – Global Hydrogen Review 2023 – Analysis - IEA] reported a rapid increase in the number of announced projects for low-emission H2 production. The annual production of low-emission H2could reach 38 million metric tonne (MT) in 2030. Of the total projects included in the 2022 Global Hydrogen Review, 27 MT are based on electrolysis and low-emission electricity and 10 MT on fossil fuels with carbon capture, utilization, and storage.

[0005] Hydrogen is a commonly found element, one of the most abundantly available in the universe but in minuscule concentrations in the air. Unlike CH4 or other fossil fuels, there is essentially no production of H2from naturally occurring sources of pure H2on Earth, but the possibilities are being explored with greater enthusiasm since some discoveries in Mali andelsewhere have confirmed its existence (Hand, 2023, Chiaramonte, 2021). For many petrochemical / chemical products, it is consumed as a process reactant or fuel.

[0006] Currently, the global annual H2 demand is approximately 95 million tons (1 trillion m3(NTP)). Bulk petrochemicals—NH3and CH3OH—and processing / refining of gasoline consume nearly two-thirds of the annual H2 demand. These facilities typically have H2 requirements of up to 500,000 Nm3 / h. The remaining H2demand is directed to many different process industries including fats and oils processing, chemicals, pharmaceuticals, metallurgy semiconductor production, and aerospace industries.

[0007] In agriculture, H2 is used to produce NH3 by the Haber process, along with methanol and cyclohexane, which are used to produce pharmaceuticals and plastic. Hydrodesulfurization of fuels in oil refining processes is a catalytic chemical reaction used to reduce sulfur from petroleum products including gasoline or petrol, kerosene, jet fuel, fuel oils, diesel, and natural gas. Desulphurization is used to create fuels including ultra-low-sulfur diesel, to curb sulfur dioxide emission from aircraft, automotive vehicles, ships, railroad locomotives, and oil or gas power plants, among various other forms of petroleum combustion.

[0008] Captive systems, including on-site generation, may negate several problems linked to the conveyance and distribution of large volumes of H2 required daily. Recently, on- site H2generation in small-scale industries has gained more popularity owing to new technologies being offered at affordable costs in comparison to delivered distributed channels. In terms of volume, this segment is expected to expand at a CAGR of 9.2% over the forecast period.

[0009] The merchant systems segment was valued at $88 billion in 2022 with a 7.4% CAGR. Merchants supply H2 to small- to medium-sized industries at competitive rates, in cylinders or other carriers. Small- to medium-sized industries may not necessarily afford setup costs and upkeep of generational facilities. Hence, they opt to buy H2 from dealers.

[0010] There is a need for improved methods of efficiently generating hydrogen, while minimizing undesired byproducts.Brief Summary of the Disclosure

[0011] Our overall technical approach is to effectively combine an active catalyst for NH3 decomposition with a selective H2 membrane into a system that can operate in an industrial application. This novel integrated membrane reactor design, together with a low-cost, highly active Ru-based catalyst, is advantageous over previous designs. This applied research and development work, therefore, represents an innovative and transformational solution to H2generation from thermal catalytic NH3 decomposition. The presently disclosed membrane reactor can provide high-purity H2 for a fuel cell-powered vehicles (e.g., automobiles, forklifts, etc.) The cracking technology can open new markets for H2 utilization. Description of the Drawings

[0012] 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.

[0013] Figure 1A. An exploded-view diagram of a hydrogen generator according to an embodiment of the present disclosure.

[0014] Figure 1B. An exploded-view diagram of a second reactor layer, a second hydrogen-selective membrane, a second gas-collecting layer, and additional components according to another embodiment. The second reactor layer may be attached to the device shown in Figure 1A (i.e., connected to the bottom side of the burner layer of Figure 1A and where the depicted bottom of the burner layer will be replaced by a second conduction plate (or is a second conduction plate)).

[0015] Figure 3. Schematic of membrane reactor for H2generation from thermal catalytic NH3 decomposition.

[0016] Figure 3. Process flow diagram of the catalyticNH3decomposition system for H2generation.

[0017] Figure 4. Ammonia cracking applications.

[0018] Figure 5. System process flow diagram of another embodiment according to the present disclosure

[0019] Figure 6. A system according to another embodiment of the present disclosure.

[0020] Figure 7.H2 chamber unit

[0021] Figure 8. A system according to another embodiment of the present disclosure.

[0022] Figure 9. Embodiment of a burner design for 0.5 kgH2 / day

[0023] Figure 10. Prototype main components

[0024] Figure 11. Initiation, ammonia preheating tube and external insulation

[0025] Figure 12. Prototype preliminary results

[0026] Figure 13. Maximum recoverable hydrogen determination - small scale reactor

[0027] Figure 14. > 99.9% NH3decomposition using H2burner and external H2fuel

[0028] Figure 15. Stable NH3decomposition for > 6h using H2burner and generated H2from ammonia

[0029] Figure 16. Schematic of self-sustained NH3decomposition reactor systems according to some embodiments of the present disclosure: (A) version 1 and (B) version 2.

[0030] Figure 17. Simulation results for the self-sustained NH3 decomposition system.

[0031] Figure 18. Temperature profiles along the external surface of a ¼ inch burner tube as a function of time and H2 flowrate. (A, B) holes are made along 7 cm of the tube while (C, D) holes are made along 14 cm of the tube.

[0032] Figure 19. Temperature profiles along the external surface of a ¼ inch burner tube as a function of H2 flowrate. (A) spiral staggard holes (B) straight holes.

[0033] Figure 20. V1 Reactor NH3decomposition results as a function of NH3flowrate and H2 fuel flowrate. Burner fuel is 30% H2 in N2 balance.

[0034] Figure 21. V2 NH3 reactor decomposition results where the H2 burner is fed with 300 sccm H2 (30% in N2 balance).

[0035] Figure 22. Mass spectroscopy signals for the decomposition product of V2 reactor using 600 and 300 sccm NH3 and H2 flowrates, respectively.

[0036] Figure 23. Simulation results for the self-sustained NH3 decomposition system.

[0037] Figure 24. A schematic of the proposed burner design and a picture of the micro mill MF 70 from Proxxon used in hole perforation

[0038] Figure 25. Photographs of a developed self-sustained NH3 decomposition reactor system. (A) version 1

[0039] Figure 26. SEM images of SS tube surface before (A) and after (B) coating with Pd layer. The scale bar is 10 microns.

[0040] Figure 27. H2burner initiation time as a function of its starting temperature: The burner is fed with 100 SCCM mixed with air and the thermocouple reads the temperature outside the tube at the front of the catalyst layer while the tube surface is covered with silica wool insulation.

[0041] Figure 28. Temperature profiles along the external surface of a ⅜ inch burner tube as a function of H2 flowrate for 24 cm long burner

[0042] Figure 29. Temperature profiles along the external surface of ¼ inch SS burner tube: temperature at the front of a short catalyst bed (~3 cm) and along the bed respectively as a function of time and H2flowrate (A & B). Temperature profiles as a function of time along a 24 cm long catalyst bed loosely packed and densely packed respectively where the burner is fed with 100 sccm H2(C&D). All measurements were taken on the surface of the SS tube without external insulation using 1Wt.% platinum on gamma alumina powder from Sigma Aldrich.

[0043] Figure 30. Ammonia conversion and hydrogen productivity of the used Ru-based catalyst as a function of test temperature at 1 atm and 6000 GHSV.

[0044] Figure 31. Gas analysis of V1reactor exhaust from the burner during burner start up

[0045] Figure 32. Simulation results for the self-sustained NH3decomposition system

[0046] Figure 33. Schematic of the self-sustained NH3 decomposition membrane reactor.

[0047] Figure 34. Schematic of the setup for testing the developed CMR.

[0048] Figure 35. Process flow diagram for the self-sustained NH3 decomposition system.

[0049] Figure 36. NH3reactor performance as a function of NH3flowrate and the recovered H2.

[0050] Figure 37. CMR surface temperature and temperature inside the burner as a function of the recovered hydrogen at different NH3 flowrates.

[0051] Figure 38. Product purity and content of NOx in the burner exhaust as a function of NH3flowrate and recovered H2.

[0052] Figure 39. Effect of the H2 to O2 ratio on the CMR performance tested using 2.6 SLPM NH3flowrate.

[0053] Figure 40. Cyclic performance of the developed CMR using 2.6SLPM NH3feed.

[0054] Figure 41. Photographs for the basic components of the initial CMR.

[0055] Figure 42. The perforated tube developed for the membrane reactor burner.

[0056] Figure 43. Photographs for the reactor insulation steps.

[0057] Figure 44. A photograph for the membrane reactor setup components.

[0058] Figure 45. Reactor surface temperature and temperature inside the burner tube during heating tape initiation.

[0059] Figure 46. A hydrogen fuel cell including a hydrogen generation device. Detailed Description of the Disclosure

[0060] With reference to Figure 1A, in a first aspect, the present disclosure may be embodied as a device 10 for generating hydrogen from ammonia. For example, the device may be a portable device. The device 10 has a first reactor layer 12. The first reactor layer has achamber 18. The first reactor layer 12 has an ammonia inlet 14, which is in fluid communication with the chamber 18. A retentate port 16 is in fluid communication with the chamber 18. The chamber 18 is configured to contain an ammonia decomposition catalyst. In some embodiments, the chamber 18 contains an ammonia decomposition catalyst (not shown, but see Figure 33 for example). The ammonia decomposition catalyst may include, for example, ruthenium (e.g., a ruthenium-based catalyst). The first reactor layer (i.e., the chamber of the first reactor layer) may be configured to operate at a temperature of 400 °C to 450 °C. The first reactor layer (i.e., the chamber of the first reactor layer) may be configured to operate at a pressure of 1-4 bar, inclusive.

[0061] The device 10 includes a first gas-collecting layer 20 having a manifold 22 (i.e., an interior volume). The first gas-collecting layer 20 has a hydrogen outlet 24 in fluid communication with the manifold 22. A first hydrogen-selective membrane 26 is disposed between the first reactor layer 12 and the first gas-collecting layer 20. In other words, the first hydrogen-selective membrane 26 separates the chamber 18 of the first reactor layer 12 and the manifold 22 of the first gas-collecting layer 20. The first hydrogen-selective membrane 26 is configured such that hydrogen gas generated in the chamber 18 of the first reactor layer 12 will permeate through the hydrogen-selective membrane 26 into the manifold 22 of the first gas- collecting layer 20. The first hydrogen-selective membrane may provide near-absolute selectivity for hydrogen with virtually complete exclusion of nitrogen and ammonia. For example, various embodiments of the first hydrogen-selective layer may have an H2 selectivity of > 250, 1,000, or 10,000 (for example, compared to N2). In another example, in various embodiments, the first hydrogen-selective membrane may have a permeance for N2 or NH3 of less than 10-11mol·m-2·s-1·Pa-1nitrogen or ammonia--may be trace or undetectable. These are only sample values and the membrane may have different properties which are greater or less than the example selectivity and / or permeance values provided. The first hydrogen-selective membrane may be made from or include palladium or a palladium silver alloy.

[0062] The device 10 includes a burner layer 30 adjacent to the first reactor layer 12. The burner layer 30 is separated from the chamber 18 by a first conduction plate 32. The first conduction plate may have a palladium coating on a burner side (the side of the conduction plate which faces the burner layer. The burner layer 30 has an intake port 34 and an exhaust port 36. The intake port 34 is in fluid connection with the retentate port 16 of the first reactor layer 12. For example, a tube may fluidically couple the intake port 34 with the retenate port 16.

[0063] A sealing member may be located at an interface of the first reactor layer and the first hydrogen-selective membrane. Another sealing member may be located at an interface of the first gas-collecting layer and the first hydrogen-selective membrane. The sealing member(s) may be a gasket (i.e., an O-ring) such as, for example, a graphite gasket.

[0064] In some embodiments, the chamber 18 contains an ammonia decomposition catalyst (not shown, but see Figure 33 for example). The ammonia decomposition catalyst may include, for example, ruthenium (e.g., a ruthenium-based catalyst). The chamber may further contain an inert material configured to confine movement of the ammonia decomposition catalyst. The inert material may be, for example, silica wool. The inert material may further protect the first hydrogen-selective membrane from damage caused by contact with the ammonia decomposition catalyst.

[0065] The device may include a heat exchanger for pre-heating an ammonia gas supplied to the first reactor layer (see, for example, cross exchanger of Figure 3). For example, the heat exchanger may have a first channel in fluid communication with the exhaust port of the burner layer and a second channel in fluid communication with the inlet of the first reactor layer. In this way, the heat exchanger is configured to use exhaust gas from the burner layer to pre-heat ammonia gas supplied to the first reactor layer.

[0066] With reference to Figure 1B, in some embodiments, the device 10 may have a second reactor layer 40 adjacent to the burner layer 30. For example, the second reactor layer 40 may be on a side of the burner layer 30 opposite the side of the first reactor layer 12. The second reactor layer has a chamber 48. A second conduction plate separates the burner layer 30 from the chamber 48 of the second reactor layer 40. The second reactor layer 40 has an ammonia inlet 44, which is in fluid communication with the chamber 48 of the second reactor layer. A retentate port 46 is in fluid communication with the chamber 48 of the second reactor layer. The chamber 48 of the second reactor layer is configured to contain an ammonia decomposition catalyst. In some embodiments, the chamber 48 of the second reactor layer contains an ammonia decomposition catalyst (not shown, but see Figure 33 for example). The ammonia decomposition catalyst may include, for example, ruthenium (e.g., a ruthenium-based catalyst). The second reactor layer (i.e., the chamber of the second reactor layer) may be configured to operate at a temperature of 400 °C to 450 °C. The second reactor layer (i.e., the chamber of the second reactor layer) may be configured to operate at a pressure of 1-4 bar, inclusive.

[0067] In embodiments having a second reactor layer 40, the device 10 may include a second gas-collecting layer 50 having a manifold 52 (i.e., an interior volume). The second gas- collecting layer 50 has a hydrogen outlet 54 in fluid communication with the manifold 52 of the second gas-collecting layer 50. A second hydrogen-selective membrane 56 is disposed between the second reactor layer 20 and the second gas-collecting layer 50. In other words, the second hydrogen-selective membrane 56 separates the chamber 48 of the second reactor layer 20 and the manifold 52 of the second gas-collecting layer 50. The second hydrogen-selective membrane 56 is configured such that hydrogen gas generated in the chamber 48 of the second reactor layer 20 will permeate through the second hydrogen-selective membrane 56 into the manifold 52 of the second gas-collecting layer 50. The second hydrogen-selective membrane may provide near- absolute selectivity for hydrogen with virtually complete exclusion of nitrogen and ammonia. For example, various embodiments of the second hydrogen-selective layer may have an H2 selectivity of > 250, 1,000, or 10,000 (for example, compared to N2). In another example, in various embodiments, the second hydrogen-selective membrane may have a permeance for N2 or NH3 of less than 10-11mol·m-2·s-1·Pa-1nitrogen or ammonia--may be trace or undetectable. These are only sample values and the membrane may have different properties which are greater or less than the example selectivity and / or permeance values provided. The second hydrogen- selective membrane may be made from or include palladium or a palladium silver alloy.

[0068] An additional sealing member 62 may be located at an interface of the second reactor layer and the second hydrogen-selective membrane. Another additional sealing member 62 may be located at an interface of the second gas-collecting layer and the second hydrogen-selective membrane. The sealing member(s) may be a gasket (i.e., an O-ring) such as, for example, a graphite gasket.

[0069] The device may include a heat exchanger for pre-heating an ammonia gas supplied to the first reactor layer and the second reactor layer (see, for example, cross exchanger of Figure 3). For example, the heat exchanger may have a first channel in fluid communication with the exhaust port of the burner layer and a second channel in fluid communication with the inlet of the first reactor layer and the second reactor layer. In this way, the heat exchanger is configured to use exhaust gas from the burner layer to pre-heat ammonia gas supplied to the first reactor layer and the second reactor layer.

[0070] In another aspect, the present disclosure may be embodied as a hydrogen fuel cell which includes a device similar to that described above (see Figure 46). In this way, the fuel cell may receive a source of hydrogen from the device.

[0071] It will be apparent in light of the present disclosure that additional burner layers, reactor layers, and gas-collecting layers may be provided. Additional description and embodiments (including embodiments having tubular components) are provided in the discussion below.

[0072] Additional Discussion

[0073] The following discussion and examples are provided for illustrative purposes only and are not intended to limit the scope of the present disclosure. The examples, theories, and embodiments described herein are meant to aid in understanding certain aspects of the disclosure but should not be construed as restricting the claims or limiting alternative implementations.

[0074] Ammonia decomposition or cracking is shown in Equation 1. 2NH3↔ 3 H2+ N2(∆H = +46 kJ / mole NH3or 2.7 MJ / kg NH3) Equation 1

[0075] In an aspect, the present disclosure may be embodied as hollow fiber membranereactor, as shown in Figure 2, comprising: i) a low-cost, highly active ruthenium (Ru)-basedcatalyst for NH3decomposition, ii) a H2selective membrane on ceramic hollow fibers with high surface to volume ratio for extracting H2from N2(simultaneously shifting NH3decomposition reaction), and iii) a catalytic H2burner to provide thermal energy for NH3decomposition (to be developed in the current program).

[0076] The novel use of a H2selective membrane in an NH3cracking reactor provides high-purity H2 as well as shifting the NH3 decomposition equilibrium to achieve high conversion. Moreover, the retentate, containing lower concentrations of H2, can be utilized efficiently in an optimized catalytic H2 burner to uniformly heat the catalyst bed thereby providing thermal energy for NH3decomposition and improving the system energy efficiency. The innovative integration of these three components (Ru-based catalyst, H2-selective membrane, and catalytic H2burner) results in a compact and modular membrane reactor that achieved all technical goals of the ARPA-E REFUEL Program. Table 1 shows the extent a prototype system exceeded the program goals. This was achieved at a relatively low temperature(< 450 °C) to deliver H2at a high rate per volume (> 0.15 g H2 / h / cm3) and a low delivery cost (< $3.5 / kg) for commercial-scale modules. The projected source-to-use energy cost (SUE) was $0.263 / kWh, exceeding ARPA-E’s target of < $0.3 / kWh. Table 1. Performance of the earlier prototype exceeded program goals. Items Goal Achieved Energy efficiency >80% 87% H2high purity>99% >99.9%0.15 g H / h / cm30.16-0.47 g H / h3Deliver volumetric rate 2 2 / cm NH3conversion>99% >99.5%Product NH3concentration<100 ppb <10 ppb

[0077] The presently disclosed novel hollow fiber membrane reactor (for example, as depicted in Figure 2) may be used in a process design capable of achieving high-energy efficiency. An example process flow diagram is shown in Figure 3.

[0078] The process may feature the following advantages / characteristics: ^ Pressurized NH3 vapor at 10~15 bar (Stream 1) used as feedstock; ^ Low-cost, highly active ruthenium (Ru)-based catalysts used in a compact fixed bed reactor for high-rate NH3 decomposition (2^^^^3 ↔ ^^2 + 3^^2) at a reaction temperature below 450 °C; ^ H2selective membrane on ceramic hollow fibers with a high surface-to-volume ratio (2,700 m2 / m3) used as a reactor boundary to extract high-purity H2 (> 99%) from the reaction product; the removal of the H2will also shift the reaction towards higher conversion; ^ Optionally, the active catalyst may also be loaded in the ceramic hollow fiber to decompose residual NH3 as desired to improve H2 quality specifications; ^ Lower concentration residual H2in the retentate (Stream 4) may be burned with air in a catalytic burner to provide thermal energy for NH3 decomposition;^ Permeated high-purity H2(Stream 3) and exhaust from catalytic H2combustion (Stream 6) at elevated temperatures will be used to preheat NH3 feed (Stream 1); and ^ High purity H2 (> 99%) (Stream 3) after a cross exchanger can be compressed to 30 bar for delivery to a PEM fuel cell or other uses.

[0079] The present disclosure may be embodied as a device for generating hydrogen from ammonia. The device includes a reactor chamber having an ammonia inlet. An ammonia decomposition catalyst is disposed within the reactor chamber. The ammonia decomposition catalyst may be a ruthenium-based catalyst (e.g., containing ruthenium). The catalyst may have any form suitable for use in a reactor chamber. For example, the catalyst may be in pellet form (e.g., pellets having a diameter of 0.4 mm to 1.5 mm, inclusive—though in some applications, the pellets may be smaller or larger). In another example, the catalyst may be in fiber form. The catalyst may have any catalyst support structure configuration (e.g., advantageously providing a large surface area for a gas-phase process).

[0080] One or more cylinders (sometimes referred to herein as fibers) of porous substrate are contained within the reactor chamber. The cylinders may be made from a ceramic material, such as, for example, Al2O3. In some embodiments, the cylinders are hollow. Each of the one or more cylinders may be coated with an H2 selective membrane. The H2 selective membrane may include palladium or a palladium alloy. The H2selective membrane may have a hydrogen / nitrogen selectivity of 250 or more.

[0081] At least one end of each of the one or more cylinders is in fluidic communication with an H2outlet. In this way, H2gas can pass through the H2selective membrane into the interior of a porous substrate cylinder and be collected via the H2 outlet.

[0082] In some embodiments, the device includes an H2burner configured to heat ammonia gas prior to introduction of the ammonia gas into the reactor chamber via the ammonia inlet. The H2burner may be in fluid communication with a flue gas outlet of the reactor chamber to receive flue gas (having, for example, N2 and H2) for combustion.

[0083] In some embodiments, the reactor chamber is configured to operate at a temperature of 400 °C to 450 °C. In some embodiments, the reactor chamber is configured to operate at a pressure of 1–4 bar, inclusive.

[0084] In another embodiment a device for generating hydrogen from ammonia includes a reactor chamber having an ammonia inlet. An ammonia decomposition catalyst is disposed within the reactor chamber. The ammonia decomposition catalyst may be a ruthenium-based catalyst (e.g., containing ruthenium). The catalyst may have any form suitable for use in a reactor chamber. For example, the catalyst may be in pellet form (e.g., pellets having a diameter of 0.4 mm to 1.5 mm, inclusive—though in some applications, the pellets may be smaller or larger). In another example, the catalyst may be in fiber form. The catalyst may have any catalyst support structure configuration (e.g., advantageously providing a large surface area for a gas-phase process).

[0085] One or more cylinders (sometimes referred to herein as fibers) of porous substrate are contained within the reactor chamber. The cylinders may be made from a ceramic material, such as, for example, Al2O3. In some embodiments, the cylinders are hollow. In some embodiments, the hollow interior of the one or more cylinders may contain ammonia decomposition catalyst the same or different ammonia decomposition catalyst. Each of the one or more cylinders may be coated with an H2 selective membrane. The H2 selective membrane may include palladium or a palladium alloy. The H2 selective membrane may have a hydrogen / nitrogen selectivity of 250 or more.

[0086] At least one end of each of the one or more cylinders is in fluidic communication with an interior of the reactor chamber. In this way, the products of ammonia cracking can enter the one or more cylinders and the hydrogen may pass through the hydrogen-selective membrane.

[0087] The device includes one or more impermeable tubes. Each of the one or more impermeable tubes surrounds an exterior of theH2 selective membrane of a corresponding cylinder of the one or more cylinders of porous substrate. The one or more impermeable tubes are configured to containH2 gas that passes through theH2 selective membrane. Each of the one or more impermeable tubes includes anH2 outlet.

[0088] A second end of each of the one or more cylinders of porous substrate is an exhaust. In some embodiments, the second end of each of the one or more cylinders is in fluid communication with an air inlet of the reactor chamber.

[0089] In some embodiments, the device includes an H2burner configured to heat ammonia gas prior to introduction of the ammonia gas into the reactor chamber via the ammonia inlet. The H2burner may be in fluid communication with a flue gas outlet of the reactor chamber to receive flue gas (having, for example, N2 and H2) for combustion.

[0090] In some embodiments, the reactor chamber is configured to operate at a temperature of 400 °C to 450 °C. In some embodiments, the reactor chamber is configured to operate at a pressure of 1–2 bar, inclusive.

[0091] The required H2 purity may vary by application. While an attractive market for the present technology is H2for fuel cells and fuel cell vehicles, there may be other market opportunities that have lower purity requirements. There are additional ways to classify H2, including total H2present, quality verification levels (QVLs), and fuel cell grade. Gaseous H2is denoted as Type I and liquid H2 as Type II. Common purity nomenclature for H2 can be stated with respect to the number of nines as seen in Table 2. For example, 3 nines refer to the minimum H2 concentration of 99.9% and an impurity limit of 1000 ppm, while 4 nines would be 99.99% and 100 ppm. Table 2. Common Purity Nomenclature. Common Minimum H2Impurity limit excluding name concentration helium m y 2

[0092] The Compressed Gas Association (CGA) has developed standards for H2 purity depending on the use. Their standard, G-5.3 – 2017 Commodity Specification for Hydrogen –7th Edition, provides specifications for gaseous and liquid H2by grade. It also includes general sampling methods and analytical procedures for impurities.

[0093] Table 3 shows the maximum level for each component by Type and QVL. A blank cell does not imply that the limiting characteristic is or is not present but merely that the test is not required for compliance with the specification. By standard practice, the supplier ensures the QVL of the H2product. The typical use categories are shown in Table 4.

[0094] One such application is retail H2 sales for fuel cell vehicles. SAE J2719, Hydrogen Fuel Quality for Fuel Cell Vehicles specifies H2impurities that are detrimental to fuel cell vehicle systems. These are shown in Table 5.

[0095] Traditionally, QVL are measured at bulk custody transfer points where piping leads to the customer equipment. In the case of fuel cell (FC) grade, H2 can be collected by sample container at the dispenser nozzle. Hydrogen suppliers may choose to provide higher purity H2 with lower QVL values than required for FC grade. Table 3. Directory of Limiting Characteristics Quality Verification Levels Limiting Characteristic Maxima for Type I (gaseous) hydrogen Maxima for Type II (liquid) hydrogen T i l U B1D E F1L FC A B C E .9 0 0 0Table 4. Typical Uses Type I Gaseous Type II Liquid QVL T ical uses QVL T ical usesStandard industrial fuel, fuel cell, and propellant B General industrial applications A applications per MIL-PRF-27201E Fuel, hydrogenation, and water chemistry D B Hi h rit ind tri l f l f l ll nd li tions

[0096] NOTE-Hydrogen supply to the motor fuel dispensing station shall meet or exceed the minimum standards shown in Table 3 as some contamination of the H2at the dispenser nozzle to vehicle interface with atmospheric gases, moisture, and particulates is unavoidable.

[0097] The QVL limits shown in Table 2 are the minimum requirements that shall be met by the H2 motor fuel supplier. In most cases, H2 in compliance with the FC QVL values in Table 3 also complies with the impurities listed in Table 5, based on the current H2production methods. Table 5. Trace Impurities Detrimental to Fuel Cells, but not Typically Analyzed Constitute Chemical formula SAE J2719 limits, ppm Total sulfur Includes: H2S COS CS20.004Ammonia Cracking

[0098] Hydrogen is expensive to transport over long distances. There is a growing interest in NH3use as a H2carrier. Ammonia is gaseous under standard conditions but easily liquefies at 8 bar pressure or −33 °C. Ammonia is the second-largest chemical produced globally with 18 million tons of NH3currently stored and transported each year. More than 200 gas carriers capable of shipping NH3 and liquid petroleum gas are currently in operation around the world and 150-200 ports are equipped with terminals for unloading and storing traded NH3.

[0099] On a molecular basis, NH3is comprised of 75% H2molecules and 25% N2but on a mass basis, NH3 is 82% N2 and only 18% H2. Stoichiometrically as shown in Equation 1, the maximum yield for NH3cracking is 0.1776 kg H2 / kg NH3. Upon cracking, a blend of H2and N2is formed. This mixture is referred to as “forming gas”. Reaction 1 is endothermic, requiring the addition of energy to proceed. This energy can be supplied by combusting part of the NH3feedstock via Reaction 2, combusting part of the purified H2 product via Reaction 2, electrical heating, or otherwise. 4NH3 + 2O2 → 2 N2 + 6H2O (∆H = −293 kJ / mole NH3) Equation 2 2NH3 + O2 → 2H2O (∆H = −242 kJ / mole H2) Equation 3

[0100] The extent of the cracking reaction is controlled by equilibrium. The equilibrium conversion of NH3 at 600 °C and atmospheric pressure is 99.9% resulting in 0.1% of the residual, unconverted NH3. Higher pressures decrease the equilibrium conversion. A catalyst is required to accelerate the reaction at lower temperatures.

[0101] The global NH3 cracking market is projected to increase from $200 million in 2023 to 1.6 billion in 2032 at a CAGR of 26% (See “Ammonia Crackers: Powering a Greener Future with Hydrogen? Market Set to Reach $1,574.44 Million by 2032” (prnewswire.com)). NH3cracking applications can require “partially cracked NH3” (a mixture of NH3, H2, and N2), forming gas, fully cracked gas with residual unconverted NH3, or purified H2. Many of these applications benefit from the properties of a H2 / NH3blend. The lower burning temperature and lower burning velocity of NH3 versus hydrocarbons mean 100% NH3 burning in large-volume engines and turbines is currently not feasible. Research is underway to first partially crack the NH3 fuel to H2 and N2 before entering the combustor, especially for large-scale applications.

[0102] Figure 4 shows the current and developmental applications forNH3cracking bycapacity,degree of cracking, and commercial readiness.Transportation Fuel (H2fueling station)

[0103] The key market we intend to target is the H2fuel cell market at H2fueling stations. The global H2 fuel cell vehicle market was valued at $1.5 billion in 2022 and is projected to grow to $58 billion by 2032 at a CAGR of 43% (See “Hydrogen Fuel Cell Vehicle Market Size, Share, Trends, Analysis” (alliedmarketresearch.com)). Major auto players includeMercedes-Benz AG, Hyundai Motor Group, General Motors, AB Volvo, Audi AT, Toyota Motor Corporation, BMW Group, and Honda Motor Co.

[0104] This application requires H2 purity to meet the requirements of PM fuel cells, generally 99.999 mol.%, ISO 14687:Pt2). With a rated power of typically 113-120 kW, assuming an electromotor efficiency of 85% and a PEM fuel cell efficiency of 50%, the H2 requirement for full-rated power continuously is about 200 kg H2 / day. Pipeline H2 for General Industry

[0105] Plans today for producing H2 for the general industry have been primarily for centralized facilities at certain key locations, like at import terminals or existing gas grids. This would allow the H2 to be transported to the industrial customer via centralized pipeline networks. The degree of H2 purity would need to be the same for all customers on the pipeline. If higher purity was required, distributed purification technologies, predominantly pressure swing adsorption (PSA), would be required. Metallurgy

[0106] The metallurgy industry uses forming gas for welding, soldering, brazing, and metal nitriding. These crackers are commercially available. They are operated at atmospheric pressure and temperatures of 850 – 1,000 °C. This provides high NH3conversion with a residual content of less than 30 ppm. The forming gas is not separated. Typically, the gas is sent first to nitriding where the N2is used and then to the heaters which use the H2. Stationary Power (low temperature fuel cells)

[0107] Ammonia-fueled gensets are available commercially combined with alkaline fuel cells. Because the alkaline fuel cells are more tolerant to NH3in the H2feed, the cracker can operate at temperatures of less than 700 °C without requiring further purification. The energy required for the cracking reaction can be supplied by electricity or via burning of the NH3feed or H2 product. Various companies offer these gensets for applications such as off-grid power, electric vehicle charging stations, and backup power supplies. For proton exchange membrane fuel cells (PEMFC) applications, high purity H2 (> 99.5%) is required since a large quantity of NH3leads to catalyst poisoning in fuel cells.Heavy Water Production

[0108] One path for the production of heavy water (deuterium oxide, D2O) is based on the monothermal ammonia-hydrogen exchange process that includes NH3 production and cracking. Cracking of an “enriched’ mixture of ND3 and NH3produces D2, mixed DH, and N2. These crackers operate at pressures of 120-150 bar. The gaseous mixture is combusted in air to produce enriched water (HDO) and D2O. Heavy water is used in some nuclear reactors as a neutron moderator and in some analytic equipment when H2 is a target or the H2O signal interferes with the target molecule. Transportation Fuel (tractors)

[0109] Ammonia cracking can be applied as a transportation fuel to tractors and similar pieces of equipment by feeding the forming as directly to an internal combustion engine (ICE) or purifying the forming gas to provide “pure” H2to a fuel cell as we will be demonstrating in the current project. These ICE applications can take advantage of the engine exhaust to provide the energy needed for cracking. Generally, the cracked gases do not need to be cooled before being fed to the ICE. The exhaust gas from the engine may require emissions control for NOx, N2O, and NH3. Startup may require supplemental fuel until the cracker is online. Transportation Fuel (locomotive)

[0110] Although commonly called “diesels,” the locomotives are electrically driven. The diesel engine drives an alternator, which produces electricity to run electric motors mounted on the locomotive’s axles. The internal combustion engine was a dramatic improvement in efficiency over the steam locomotive, making substantial savings possible in maintenance and the elimination of widespread facilities. A typical DB Class V90 locomotive has a power output of around 1 MW. Development of NH3-fueled and forming gas-fueled locomotives is underway. Transportation Fuel (shipping and the maritime sector)

[0111] Ammonia crackers are effective for four-stroke, spark ignition engines because H2 ignites easily. Ammonia crackers are not effective for two-stroke, compression-ignition engines because both H2 and NH3 are difficult to compress compared to conventional fuels like marine fuel oil. The cracker uses heat from the engine exhaust and the forming gas can be fed to the engine without cooling. The exhaust gas from the engine may require emissions control for NOx,N2O, and NH3. Ammonia crackers are also being developed for use with on-board fuel cells. Here H2 purification may be needed to remove the residual NH3. Transportation Fuel (civil aviation)

[0112] The use of NH3as an aviation fuel is currently being evaluated. There are several research programs, including ARPA-E, examining feeding NH3 and cracked forming gas to the jet engines to improve combustion. The cracker uses heat from the jet engine exhaust and the forming gas can be fed to the engine without cooling. The exhaust gas from the engine may require emissions control for NOx, N2O, and NH3. Power Generation (gas turbines)

[0113] The development of NH3 as a feed to gas turbines continues and blending of the NH3with forming gas research is underway. Ammonia-fed gas turbines are not commercially available; 40 MW turbines are projected to be available in the mid-2020s. Using NH3 crackers for power generation conceptually integrates the cracker to use the heat from the jet engine exhaust and feed the forming gas to the engine without cooling. Industrial Burners

[0114] Industrial burners have a heat capacity ranging from a few MW to hundreds of MW. In some cases, partial cracking of NH3 is sufficient, while in other cases, pure NH3 burners may also find applications. Development is underway for the power and heat generation industrial burners.

[0115] Generally, these burners do not require full cracking of the NH3and can use the forming gas without cooling. The exhaust gas from the burners may require emissions control for NOx, N2O, and NH3. If the burners are not fully enclosed, they may require cracking with H2purification to mitigate NH3 slip and NOx emissions. Stationary Power and Transportation Fuel (high temperature fuel cells)

[0116] Ammonia cracking before the solid oxide fuel cell (SOFC) allows for better energy management inside the fuel cell, as well as higher power densities than NH3-fueled SOFC. The SOFC operate at 700-800°C so cooling of the forming gas is not required. These units can be heated with waste heat from the SOFC with augmentation from electric heating.Due to their elevated operating temperature, they are well-suited to uninterrupted use with minimal startups and shutdowns. While SOFC can tolerate residual NH3, it does affect the power density and energy management inside the fuel cell. Potential Markets Assessment

[0117] In some embodiments, the system under development may be used for applications with a small to medium capacity and a high degree of cracking. This includes transportation fuel (H2 fueling stations), pipeline H2 for general industry, metallurgy, and stationary power (low-temperature fuel cells). Smaller industrial boilers can also be included. First Markets

[0118] The NH3 decomposition technology should focus on transportation fuels and distributed and remote H2for general industry, fuel cells, and boilers. These are market entry opportunities that will be smaller systems that de-risk scale-up issues. The preliminary TEA in preparation for Milestone M6.3.6.2 is scaled to produce 1,500 kg / day of H2. This scale will meet or exceed the early needs of these markets. Product Scale for Stationary and On-board Fuel Cell Applications

[0119] Stationary and on-board fuel cells will require 900-1,500 kg / day of H2per MW of fuel cell power. This demand is too great to be filled for stationary power plant fuel cells where the demand can be 100s of MW. Fuel cell applications where the demand is in the 100s of kW may be served by this technology. Potential Markets for Decomposition Catalyst for Non-Fuel Cell Applications

[0120] The provided system may be used for applications with a small to medium capacity and a high degree of cracking as shown in Figure 4. The preliminary TEA in preparation for Milestone M6.3.6.2 is scaled to produce 1,500 kg / day of H2. This includes transportation fuel (H2 fueling stations), pipeline H2 for general industry, metallurgy, and smaller to medium industrial boilers.

[0121] EXAMPLE 1

[0122] 1. Introduction

[0123] The growing transition towards a low-carbon economy is making hydrogen (H2) production technologies a top priority in the energy sector. However, challenges in H2 storage and transportation are currently limiting its industrial and commercial applications. Indirect H2 carriers, which can be transported to consumption areas and decomposed on-site to generate H2, offer a potential solution. Ammonia (NH3) is a carbon-free H2carrier that contains 17.8 wt.% H2 and is a liquid at 8.6 bar and 20 °C, suggesting safer storage, handling, and distribution through the current established infrastructure and regulations. Moreover, in terms of energy security, NH3 can be locally produced from entirely carbon-free renewable resources. NH3conversion back into H2is commonly conducted by thermal decomposition where energy input is needed for its endothermic cracking reaction (NH3 → 1.5 H2 + 0.5 N2,ΔH∘= +46 KJ / mol). If a source of waste or renewable heat is available to drive the reaction, H2 production from NH3 can be maximized. In rural or emergency areas, or where external waste heat sources are scarce, NH3decomposition systems should operate independently. Further studies are still needed to estimate their operational efficiency.

[0124] The predominant focus in recent NH3 cracking studies revolves around developing catalyst systems with enhanced durability and high activity. These systems aim to efficiently convert NH3 to H2 at the lowest possible temperature, consequently lowering the energy loss and increasing system efficiency for the process. For example, Yavuz and coworkers reported that vertically standing 2D porous supported Ru nano catalysts can decompose NH3with a H2generation rate of 95.17 mmol gRu−1min−1at 450 °C. Advances in solar heating catalysis may provide a potential solution to address the energy consumption associated with NH3 decomposition. In a recent study, Hao and coworkers showed that cobalt single atoms supported on cerium dioxide nanosheets exhibited a H2production rate of 2.7 mmol g−1min−1using solar heating. However, this potential solution is still limited in terms of technical maturity and technological readiness. Solar radiation is intermittent and varies with time of day, weather conditions, and geographical locations. This intermittency can pose challenges for continuous and reliable catalytic processes that depend on a constant energy source.

[0125] An alternative approach for NH3 decomposition without the need for an external energy source was reported by Nagaoka and co-workers using a RuO2 / γ-Al2O3and a Ru / Ce0.5Zr0.5O2−x catalyst. The oxidative decomposition of NH3 was initiated at room temperature immediately after NH3 and oxygen were supplied to the catalyst bed. The catalyticoxidation of NH3 raised the catalyst bed temperature to ∼ 773 K which is sufficient to induce direct NH3decomposition. These findings encouraged other groups to search for catalysts displaying high and stable activity for the oxidative decomposition of NH3. However, a simulation study reported by Jan and Han for a one vessel reactor involving oxidation and decomposition reactions of NH3 indicated that adjusting air flowrate and distribution points along the reactor is crucial to maintain the preset operation temperature and prevent liable NOx formation. Autothermal microreactors, which added NH3 oxidation for heat supply through indirect heat exchange, were also reported. In both studies the NH3required for combustion for > 99 % conversion is high compared to the decomposed NH3, suggesting low system efficiency (0.8 liter per min (LPM) combusted to decompose 0.4 LPM and 4 LPM combusted for 6 LPM ). Most NH3decomposition occurred at the inlet of the reactor, and the formation of NOx compounds could not be avoided. Badakhsh et. al, used boron nitride-coated copper as a composite reactor material for the H2burner for better heat transfer to NH3decomposition catalyst, but the temperature inside the burner was not uniform. Later, they found that using a heat transfer medium, such as gas–liquid organic phase-change material, was necessary to achieve uniform heat distribution in methylcyclohexane dehydrogenation.

[0126] It seems that NH3 decomposition-driven H2 burners are indeed facing challenges related to non-uniform temperature distribution, which can lead to inefficiencies and operational issues. Also, the integration of an H2 burner into cracking reactors is still very uncommon, and self-sustaining operation has not been demonstrated before. Addressing these challenges requires comprehensive research and development using various approaches. Herein, a self-sustained compact reactor for NH3 decomposition was designed and fabricated to provide a continuous production of H2. Once initiated, the system was fed only with pure NH3that was fully decomposed by a Ru-based catalyst to H2 and N2 gas mixture. Part of the decomposition product was mixed with air and fed back to the reactor where H2was catalytically oxidized in an optimized H2 burner, providing the required heat for NH3 decomposition. Experiments were conducted to provide an H2burner with uniform temperature distribution using small diameter perforated tube inside the burner. This innovative technology eliminates the need for external heat sources for NH3decomposition. Using H2as the fuel for heat supply allows for avoiding harmful emissions compared to systems using NH3 combustion. Moreover, stand-alone NH3decomposition systems are highly desirable in rural, remote or emergency areas where external heat is scarce. The objective of this work is to demonstrate the feasibility of a self-sustained and continuous production of H2from NH3moderate temperature thermal decomposition using the heat supplied from the produced H2 using an optimized H2 burner. The developed system is expected to contribute to the development of different applications requiring carbon-free H2 storage and production.

[0127] 2. Experimental Methods

[0128] 2.1. Materials and chemicals

[0129] SBA-200 γ-Al2O3supports were purchased from Sasol Cataltox (99.99 %, 192 ± 20 m2 / g, 30 Å pore radius, 0.45 mL / g pore volume, 0.25–0.5 mm pellet) and used as received. Potassium acetate (KCH3COO, 98 %) was obtained from Fischer Scientific. Tri- ruthenium dodecacarbonyl (Ru3(CO)12, 99.98 %) and lanthanum (III) nitrate (La(NO3)3, 99.9 %) were purchased for catalyst preparation from Sigma Aldrich. Palladium chloride (PdCl2, 99 %), ethylenediaminetetraacetic acid disodium salt dihydrate (>99.5 %), ammonium hydroxide (NH4OH, 25 %), and hydrazine monohydrate (N2H464–65 %, ≥97 %) were purchased from Sigma Aldrich for plating solution preparation.

[0130] 2.2. Catalyst preparation and characterization

[0131] The Ru-based catalyst was developed according to our published procedure. The catalysts were prepared using a sequential impregnation method. An appropriate amount of lanthanum nitrate and potassium acetate were mixed to obtain 1:12 wt ratios of (L:K) and dissolved in DI water. An aliquot of solution was then added to the dried support under magnetic stirring until incipient wetness was achieved. The impregnated support was then dried under air at 120˚C to remove water. This process was repeated until the promoter solution was completely impregnated onto the support. The promoted support was then calcined under air at 550˚C for 1 h. More details on the incipient wetness impregnation of the promoters can be found elsewhere. Following calcination of the impregnated support, an appropriate amount of ruthenium carbonyl was dissolved in tetrahydrofuran (THF) at a ratio of 37 mL THF / gcatwhile stirring to obtain a weight ratio of 2:1:12 Ru:La:K. The promoter impregnated support and the Ru / THF solution were rotated in a rotary evaporator at 60 RPM and atmospheric pressure until the Ru was fully impregnated onto the promoted support. The excess THF was evaporated, and the impregnated support was dried at room temperature. A second calcination was conducted after loading Ru onto the promoter impregnated support at 350˚C for 3 h under air. The finalcatalyst had a particle size of approximately 300 µm and contained 1 % Ru. To test catalyst activity under pure ammonia, catalyst was heated to 450 °C in argon and then subjected to a 10 % H2reduction for 1 h. Ammonia conversion was then tested in the temperature range 250 °C to 450 °C. The catalyst bed temperature was measured using a K-type thermocouple and the space velocity was kept constant for each reaction at 6000 mLNH3hr−1g-cat−1using 500 mg of catalyst.

[0132] 2.3. Fabrication and characterization of H2 burner

[0133] The function of the H2catalytic burner is to provide controlled and uniform heat to the reactor for endothermic NH3 decomposition. This can be achieved through even distribution of the H2 / air mixture along the burner tube. Additionally, the catalyst bed should also ensure efficient water drainage from the burner tube to prevent catalyst deactivation. To accomplish this, the preparation of an H2catalytic burner involved two stages: 1) coating of the burner tube surface with a thin palladium (Pd) layer; 2) using a perforated tube to evenly deliver H2 / air mixture to the active Pd layer. In the first stage, a Pd layer was deposited onto the surface of a stainless-steel (SS 316) tube using electroless plating technique, applying a procedure modified from Tang and Zuo. Prior to Pd deposition. The SS tube was degreased by sonication in acetone, followed by acid etching in 37 % HCl acid at 80 °C for 30 s and then thorough washing in deionized water and immediate drying under a cool compressed air stream. The tube was then immersed in a plating solution prepared by mixing 4.7 g / L PdCl2, 37.2 g / L ethylenediaminetetraacetic acid disodium salt dihydrate and 300 g / L ammonium hydroxide 25 %. Hydrazine monohydrate was added in a dropwise manner to the solution until no reaction took place. Plating was conducted at 50 °C. In the second stage, H2 / air delivery to the Pd coated tube was achieved using 1 / 8-inch perforated stainless steel tube sealed at its end. Evenly spaced tube perforation was made using a micro mill drill machine (MF 70 from Proxxon) and 0.8 mm diameter drill bit as shown in Figure 24. Multiple experiments were made on ¼ SS tube coated internally with Pd layer to test the self-initiation of the catalyst and then to measure the temperature distribution along the burner length before incorporation into the developed NH3 reactor. All experiments were conducted with the H2 / O2 ratio of 2. Moreover, Carl Zeiss AURIGA CrossBeam field-emission scanning electron microscope (FE-SEM) was used to image the Pd-coated layer on SS tube.

[0134] 2.4. NH3decomposition reactor development

[0135] Two NH3decomposition reactors were designed and fabricated and tested in this work. Reactor schematics and photographs are shown in Figure 16 and Figure 25, respectively. Figure 16A shows V1 reactor that was designed on heat exchanger arrangement, where NH3 and the burner fuel were fed in a counter current direction. The burner was ¼ inch SS tube coated internally with Pd layer, and a ⅛ inch tube was used for fine distribution of H2 / air mixture in the lumen side of the reactor. The heat generated from the burner was transferred to NH3catalyst bed in the shell side (½ inch SS tube). The external surface of the reactor was insulated with silica wool to reduce heat loss to the ambient atmosphere.

[0136] The schematic for V2 reactor is given in Figure 16B·NH3entered the reactor through a ⅛ inch spiral coiled SS tube that acted as a preheating section for NH3 feed, and then fed into a ⅜ inch SS tube filled with NH3decomposition catalyst. H2 / air mixture was introduced into the reactor through a perforated ⅛ inch SS tube. The external surface of the NH3decomposition tube in this case was coated with a Pd layer, and the burner perforated tube was positioned just underneath for uniform H2 / air distribution. The generated heat was transferred directly to NH3catalyst bed, and burner exhaust was in contact with NH3feed tube for NH3 preheating. All parts were contained in a long one-inch SS tube that was open at both ends and insulated using silica wool to restrict the heat transfer to the ambient atmosphere.

[0137] 2.5. Testing and analysis of the self-sustained NH3 decomposition reactor system

[0138] The developed NH3 reactor system was initially investigated using an external H2 / air mixture as the fuel for the burner. Variation of NH3flowrate and / or fuel flowrate to the burner was performed to explore optimum conditions that achieve the highest recoverable H2. Self-sustained operation mode using H2generated from NH3decomposition was then investigated and demonstrated for more than 6 h· NH3 conversion, recoverable H2, and reforming efficiency were calculated using Eqns.1–3: NH3conversion, % (1)Recover ble H2, %(2)Reforming efϐiciency, %Where ^^^ ^ is the molar flow rate of component i, NH3_feed and NH3_product represent NH3 feedto the system and NH3in the product, respectively, H2_product and H2_fuel represent H2in the product and fuel, respectively. The molar lower heating value (LHVmol) for NH3 and H2 are assumed to be 316.8 and 241.8 kJ / mol, respectively.

[0139] 2.6. Aspen simulation

[0140] Figure 17A illustrates an overall process flowsheet including the developed reactor. Typically, NH3is pumped from a storage tank and passes through a heat exchanger to capture waste heat from the hot exhaust gases coming of the H2 burner. Then, the preheated NH3 enters the reactor, where NH3 decomposition takes place using the heat from the H2 burner. Part of the decomposition product (75 % H2in N2) is combusted by mixing with air in the reactor to provide the heat required for NH3 decomposition. Aspen HYSYS was used to simulate the current conditions for the process to estimate the feasibility and the maximum H2fraction that can be recovered at steady state conditions applying Peng-Robinson (PR) fluid package. Figure 17B shows the process flowsheet, including the reactor and the burner. The process was designed based on the overall process shown in Figure 17A. A conversion reactor was used to simulate NH3decomposition into H2and N2where conversion is assumed complete at 450 °C and 1 atm for reaction (1) (R1). Part of the outlet was mixed with air and fed to another conversion reactor for H2combustion according to reaction (2) (R2). 2NH3→ N2+ 3H2ΔH∘= +46 kJ / mol NH3(R1) H2+ 1 / 2 O2→ H2O ΔH∘= −285.8kJ / mol H2(R2)

[0141] 3. Results

[0142] 3.1. H2burner development

[0143] The surface SEM images of the SS tube before and after Pd coating are presented in Figure 26. Most H2catalytic studies were conducted using Pt or Pd as reactive metals since low-cost metal oxide catalysts typically showed lower catalytic efficiency and required higher initiation temperatures. To assess the initiation temperature for the coated Pd layer, the coatedtube was fed with 100 sccm of H2mixed with air, and a thermocouple was used to measure the temperature at the front of the external surface of the ¼ inch SS tube, while silica wool insulation was applied. For the first run, to activate the catalyst layer and remove any adsorbed water, the tube was heated to ∼ 100 °C, and then H2 / air mixture was introduced into the burner. In approximately one min the thermocouple read around 350 °C. The system was then allowed to cool down to a starting temperature lower than 100 °C. As noticed in Figure 27, the catalyst was self-initiated at a start temperature down to 35 °C. However, it would take longer time to reach the target temperature if the start temperature was lower than 50 °C.

[0144] Figure 18 shows the temperature profiles along the external surface of the burner tube as a function of time and H2flowrate. In Figure 18A and B, one-centimeter-apart holes were made along a 7-cm section of the tube, while in Figure 18C and D two-centimeter-apart holes were made along a 14-cm section of the tube. In both cases, staggard hole patten was applied and the same number of holes was drilled. The two dotted vertical lines define the start and end of perforation. The results showed significant improvement achieving uniform temperature profile along the burner tube, enabling controlled and uniform heat transfer to NH3 catalyst bed at the desired length of the burner. In an attempt to double the number of holes for the same burner tube length (14 cm), we compared the staggered spiral and straight perforation patterns, as shown in Figure 19A and B. The temperature profile became more uniform using spiral staggard perforation pattern. As noticed the start and end of the burning zone can be well controlled using perforated holes. The increases in H2flowrate to the burner caused a significant temperature increase only in the preset burning zone. The increase in the temperature after the burning zone is attributed to the high temperature exhaust. The technique was found applicable even at longer burner tubes (∼24 cm) and higher flowrates, as indicated in Figure 28. In contrast, in the preliminary experiments conducted using 1 wt.% platinum on γ-Al2O3powder catalyst in short (∼ 3 cm) and long beds (∼24 cm) (Figure 29), the temperature along the short bed was not uniform and longer catalyst beds highlighted the problem of uneven heat distribution and the potential for hot spots forming inside the burner.

[0145] 3.2. NH3 reactor performance analysis

[0146] The catalytic activity of the NH3 decomposition catalyst as a function of temperature is shown in Figure 30. As expected, NH3decomposition depended on the test temperature, with a NH3 conversion of > 99 % achieved at 450 °C, a relatively moderatetemperature compared to other reported NH3decomposition catalysts. This decomposition rate is equivalent to ∼ 6.68 mmol H2 / gcat / min. The decomposition performance of the V1 reactor as a function of NH3 flowrate and burner H2 flowrate is shown in Figure 20. Increasing NH3 flowrate, while keeping H2flowrate constant, caused the reactor to cool down and thus lowered NH3 conversion. Increasing the H2 fuel feeding rate allowed the decomposition of more NH3. When the H2fuel flow rate was 300 sccm, an almost complete conversion of NH3was achieved up to an NH3 flowrate of 250 sccm. At a certain point, the system started generating more H2than that burner consumed, which we define as recoverable H2. The max recoverable H2 percentage attained using this reactor was about 44 %. Mass spectrometer was used to monitor the startup of the burner by analysis of the burner exhaust gases as shown in Figure 31. H2was completely consumed just after two minutes from the start of the burner, and no external heat was required to initiate the reaction.

[0147] The decomposition performance of V2 reactor in terms of NH3 conversion, recoverable H2, and the reforming efficiency is shown in Figure 21· NH3conversion increased significantly compared to V1 reactor when using 300 sccm H2 fuel for the burner. The catalyst bed temperature is initially high enough to accommodate the increase in NH3flow rate, maintaining a conversion rate close to 100 %, and resulting in a steady increase in the recovered H2. However, as the NH3flow rate continued to rise, it eventually caused a reduction in the bed temperature to below 450 °C. This temperature drop led to a decrease in the catalyst's activity, causing an immediate decline in NH3conversion efficiency.

[0148] The recoverable H2 reached 66.33 % while maintaining a relatively high NH3conversion (∼99 %) at NH3flowrate of 600 sccm. Since the NH3decomposition reaction yields 1.5 mol of H2per mole of NH3, the immediate impact on H2recovery may initially appear minimal despite the reduced conversion. The system can generate more H2 by increasing NH3flowrate, while sacrificing some NH3conversion. At 800 sccm, for example, the reactor showed 90 % NH3 conversion and ∼ 72.2 % recoverable H2. At the point when NH3 conversion dropped significantly, H2production declined and eventually fell below the H2consumption rate of the burner. However, the reactor operation was maintained within acceptable conversion limits for NH3, ensuring that the drop in the recoverable H2was not observed.

[0149] The reforming efficiency increased as more NH3was decomposed, maintaining a similar conversion while producing more H2. The reforming efficiency reached a maximum of82 % based on a balance between the input energy to the system and NH3flowrate (600 sccm) and then dropped with more NH3 feed. At this moment the decrease in NH3 conversion with the increase in NH3flowrate is more significant compared to the little increase in the recoverable H2. The reforming efficiency accomplished in this work was higher than that of the microreactors relying on NH3for combustion (75.9 %) and 57.3 %. Moreover, the efficiency was higher than that reported using H2 for combustion (70.95 %). Increasing the reforming efficiency of V2 reactor in comparison to previously reported microreactors indicates that the reactor is effectively using energy, which reflects the improvements in reactor design, such as optimized heat distribution and management coupled with the highly active Ru-catalyst. These enhancements can lead to more consistent and stable reactor operation. With enhanced reforming efficiency, less feedstock is needed to produce the same amount of H2. This can also reduce operating costs and improve the economic feasibility of the process.

[0150] In the V1 reactor configuration, NH3is introduced directly at ambient temperature, which results in cooling down of the catalyst bed, particularly under conditions of high NH3flowrates. This thermal effect leads to partial deactivation of the catalyst, thereby reducing its catalytic activity and efficiency in NH3 decomposition. The V2 reactor incorporates a critical design enhancement—a preheating zone. In this configuration, the high-temperature exhaust gases (>400 °C) from the catalytic burner are utilized to preheat the incoming NH3 feed stream. This preheating process elevates the temperature of NH3before it contacts the catalyst bed, ensuring the catalyst remains within its optimal operating temperature range, thereby maintaining high conversion efficiency for NH3decomposition.

[0151] Heat transfer between the heat source and the reaction zone took place mainly via conduction through the SS tube, which is primarily governed by material type and wall thickness. The same material and wall thickness were applied in both reactors. The reduction in NH3conversion in V1 reactor compared to V2 reactor was attributed to the thermal energy loss through the hot exhaust gas exiting the burner, which is not effectively transferred via conduction to the outer side of the burner wall. Badakhsh et. al, provided a comprehensive analysis of the heat losses from the reforming reactor, when an H2 burner was used as the heat source. Their results showed that the loss of energy through the burner exhaust ranged from 20.4 to 43.3 % of the H2 input energy. Furthermore, the residence time of the hot gases varies between the two reactor configurations. Reactor 1 employs a ¼-inch SS tube for the burner, while in the V2 reactor, the ammonia feed tube and the burner tube are concentrically housedwithin a 1-inch tube. This configuration in the V2 reactor results in a increased residence time for the gases compared to the V1 reactor, affecting thermal and heat transfer dynamics.

[0152] 3.2.1. V2 reactor: Self-sustained long-term operation

[0153] The data shown in Figure 21 were obtained using external H2as the fuel. Figure 22 shows the mass spectroscopy signals for V2 reactor during cracking of 600 sccm NH3 while being fed with 300 sccm H2fuel. External H2fuel (30 % H2in N2) was fed into the burner in the beginning 25 min. After 25 min on stream, NH3 decomposition product was split into two streams. One stream containing 300 sccm H2(∼75 % H2in N2) was returned to the reactor as a feed to the burner, defined as H2from NH3decomposition. No change in the performance of the burner was noticed for 6 h on stream using H2 from NH3 decomposition. Using a fuel with less amount of inert N2(25 %) enhanced the operation of the burner giving an almost complete decomposition of NH3 feed and showing a little rise in H2 signal (Figure 22). NOx was detected in the burner exhaust only at the transition of the burner to the H2from NH3decomposition. We also tried another procedure for starting up the reactor, where no external H2 was used. The reactor was first heated to 450 °C using a heating tape, and then NH3 was directly introduced. This electric power for the heating tape could be supplied using an external battery. These results proved the concept that our reactor can sustain itself while being fed only with NH3 and producing H2mixture that can be later separated as high-purity H2using H2selective membranes or pressure swing adsorption. Our technology, therefore, might be appropriate for generating H2on site as needed for fuel cells and H2stations.

[0154] 3.3. Process simulation and comparison with our system efficiency

[0155] The developed reformer considers coupling H2 combustion reaction with NH3decomposition reaction. Simple analysis of the previous reactions (1), (2) indicates that the combustion of one mole of H2 is sufficient to decompose approximately 6 mol of NH3, assuming perfect adiabatic process. In other words, 1 / 6 mol of H2is required to decompose one mole of NH3. In the meanwhile, the decomposition of one mole of NH3 will release 1.5 mol of H2. This means 1.34 mol (89 %) of H2is recoverable. Heat losses from the system are, however, expected to reduce the number of moles of NH3 that can be cracked per mole of H2 combusted.

[0156] Figure 17B illustrates an overall process flowsheet including the developed reactor and the burner. Typically, NH3 is pumped from a storage tank and passes through a crossflow heat exchanger to capture waste heat from the hot exhaust gases coming of the H2burner, preheating NH3 stream to 200 °C. Then, the preheated NH3 enters the reactor, where NH3decomposition takes place using the heat from the H2burner. The simulation was conducted using 450 sccm NH3, and all the decomposition products were initially directed to the H2combustion reactor. This means there is no recovered H2from the system, and exhaust temperature is very high. Figure 23 shows the temperature of the burner exhaust as a function of the recoverable H2percentage. The data extracted from the simulation model showed that the exhaust temperature decreased with the increase of the recoverable H2. The point at which the heat exchanger can no longer preheat the inlet NH3represents the maximum amount of H2that can be recovered from the system which corresponds to 78.8 % recoverable H2.

[0157] Moreover, we also considered a loss of 5–15 % of the H2sent to the burner to account for the heat loss from the system. The assumption of 15 % H2 loss leads to a maximum recoverable H2of 63.9 %, which is close to what we have already achieved by our reactor. In the theoretical process NH3 decomposition product is split into two streams; one is recovered and the other will be recycled back to the burner at the decomposition temperature, and heat loss to the atmosphere is neglected. The maximum H2 that can be recovered while maintaining the operation of the reactor is estimated to be 80 % of the H2generated (Figure 32B).

[0158] 4. Conclusions

[0159] A novel self-sustained NH3 cracker that combines Ru-based catalyst, and an H2burner was developed and demonstrated, eliminating the need for an external heat supply. The reactor design incorporated an H2 / air distribution system, ensuring a well-controlled and uniform heat distribution simply using a perforated SS tube inside the burner. Improved performance was accomplished when the high-temperature exhaust of the H2 burner was utilized to preheat the incoming NH3feed stream. The developed module was operated for more than 6 h using 600 sccm NH3, showing > 99 % NH3 conversion and ∼ 82 % reforming efficiency, higher than previously reported self-sustained NH3decomposition reactors. This reflects the improvements in the reactor design and energy management and will lead to improved process economic feasibility. Steady state operation revealed negligible NOxcontent in the exhaust. The fabricated reactor and its operational strategy present a scalable viable solution for on-demand H2production, making it suitable for portable and on-board applications.

[0160] EXAMPLE 2

[0161] 1. Introduction

[0162] Hydrogen (H2) production and purification technologies are becoming a top priority in the energy sector. It is well accepted that H2 is a viable alternative to fossil fuels across various industries and energy sectors, offering high efficiency and reduced environmental impact. However, challenges and safety concerns in H2 storage and transportation are still restricting its industrial and commercial utilization. This makes indirect H2carriers, which can be safely transported over long distances, stored for long time in designated areas, and later decomposed to generate H2, a potential solution. Ammonia (NH3) is a carbon-free H2 carrier having a high H2content (17.8 wt%) and is a liquid at 0.86 MPa and 20 °C, enabling safer and more convenient storage and transportation using the current infrastructure and standards. Although direct combustion is a straightforward way of utilizing NH3as a fuel, this process releases significant amounts of toxic NOx, generating high burden for exhaust gas treatment. In addition, NH3-fueled systems face challenges such as a narrow combustion range and very low laminar burning velocity. In contrast, NH3 thermal decomposition back into H2 and N2 is a cleaner technology, but it is endothermic and needs a continuous supply of energy (2NH3 → 3H2 + N2, ΔH∘= 46 KJ / mol) and the product (75 % H2) is not suitable for efficient operation of fuel cell-based applications. An efficient NH3 cracking system that can sustain its operation without external energy input, while producing high purity H2, therefore, is highly desirable.

[0163] Proton exchange membrane (PEM) fuel cells are highly sensitive to impurities in their H2 fuel supply. The presence of trace NH3 in the anode fuel stream clearly degrades the PEM cell performance to impractical levels by deactivating the catalyst and damaging fuel cell components. The International Organization for Standardization set the permissible NH3 and N2concentrations in PEM feed at 0.1 and 100 ppm. This imposes integrating NH3cracking systems with high-selectivity membranes for effective H2 separation. The membrane unit may simply follow the cracking unit. For example, Dolan and coworkers coupled a two-stage method for NH3 decomposition using 1 wt% Ru-based catalyst and H2 separation using palladium-coated tubular vanadium membrane; NH3decomposition approaching equilibrium and >90 % H2 recovery was achieved. The two-stage system seems flexible where catalyst temperature can be changed without impacting the membrane, but the membrane is not shifting the thermodynamic equilibrium conversion and thus does not play its additional role of pushing the reaction to a complete conversion.

[0164] Alternatively, the use of catalytic membrane reactors (CMRs) is frequently reported in NH3 decomposition literature as recently reviewed by Gapp and Pfeifera. CMRs integrate an H2-selective membrane with NH3decomposition catalyst in a single unit, enabling in-situ H2 separation. This design accelerates reaction kinetics and overcomes the thermodynamic equilibrium barrier, resulting in an enhanced NH3conversion and potentially lower thermal budget compared to conventional reactors. Ru-based catalysts are the most active metal catalyst for NH3decomposition. The membranes must be compatible with the reactants, catalysts, and reaction conditions and must have high H2 selectivity and permeance. Pd-based dense membranes showed remarkable performance in terms of H2separation as well as good stabilities. For example, the high selectivity of Pd / Ag membrane, leading to >99.99 % H2at ≥99.1 % NH3conversion in a CMR, could not be achieved using a modified MFI zeolite or carbon molecular sieve membranes. Gascon and coworkers used a Pd-Au based membrane to improve H2permeation at low temperature and offer higher resistance to embrittlement and demonstrated stable performance during ∼41 days on stream and >99.97 % H2 at >90 % H2recovery. Other studies were focused on different CMR configurations to enhance the performance, such as impregnating the catalyst into the membrane support or using a tube-wall CMR. These designs also help protect the membrane surface from abrasion caused by contact with the catalyst bed.

[0165] The reviewed literature indicates that H2 production by NH3 decomposition in a CMR has been experimentally explored. However, to the best of our knowledge, none of these studies have addressed or demonstrated the self-sustaining operation mode for these reactors. This capability would be especially valuable in rural, remote, or emergency areas where external heat sources are limited or unavailable. Recently, a great interest was devoted to self-sustained NH3 decomposition systems where the energy is provided by the combustion of part of NH3 feed or part of the decomposition product. Nagaoka and co-workers showed that oxidative decomposition of NH3 could be initiated at room temperature after NH3 and oxygen were fed to RuO2 / γ-Al2O3 and a Ru / Ce0.5Zr0.5O2−x catalyst beds. The bed temperature reached ∼773 K which is sufficient to induce direct NH3decomposition. This approach appears promising, but it faces significant challenges in optimizing oxygen flow rate and distribution along the reactor to maintain a uniform operating temperature and minimize the risk of NOxformation. Indirect NH3 combustion was also reported using autothermal microchannel reactors. However, peak NH3combustion occurring at the inlet of the reactor significantly affected the reformingefficiency and NOxemissions could not be avoided. The integration of an H2burner in NH3 decomposition reactor was reported by Badakhsh et al., but the temperature inside the burner was not uniform. It seems that NH3decomposition-driven H2burners are in fact facing challenges related to non-uniform temperature distribution, which can lead to inefficiencies and operational issues.

[0166] Recently, we reported a unique, self-sustained NH3decomposition reactor that coupled Ru-based catalyst and an H2 burner, eliminating the need for an external heat supply. Uniform heat distribution was achieved using a perforated stainless-steel tube in the burner. The fabricated reactor was operated for >6h using 600 sccm NH3, showing >99 % NH3conversion and ∼82 % reforming efficiency, greater than previously reported self-sustained NH3 decomposition reactors. As a follow-up work, herein, for the first time, a self-sustained compact CMR was designed and fabricated by the integration of a Ru-based catalyst, Pd / Ag membrane, and a H2 burner in one unit for continuous production of high purity H2 from NH3decomposition. Reported H2recovery ranged from 90 to 78.6 % in MR with external heat sources, because of the limited driving force for H2 permeation and thus some H2 remained in the retentate side. Our approach is to make use of the retentate stream with low H2concentration for combustion and providing heat for the reactor, therefore the system can sustain its operation, while continuously producing H2, and eliminate the need for an additional H2purification unit. To the best of our knowledge, the integration of an H2 burner into NH3 cracking CMR is still uncommon, and self-sustaining operation has not been demonstrated. This challenge requires comprehensive research to optimize their design and estimate their operational efficiency. Achieving efficient NH3 conversion, combined with the production of high purity H2 within a compact module, could facilitate the use of NH3as a H2carrier between production and utilization sites. The objective of this work is to design, fabricate and report the performance results of a self-sustained NH3CMR for continuous production of high purity H2from NH3 thermal decomposition using the heat supplied by combustion of the retentate stream in an integrated H2burner.

[0167] 2. Experimental methods

[0168] 2.1. Design of the CMR and CMR components

[0169] Figure 33 presents a schematic for the initial CMR fabricated in this work, and photographs for the reactor parts are given in Figure 41 The reactor integrates a Ru-basedcatalyst for NH3decompostion, high quality Pd / Ag membranes for H2purification, and a H2 catalytic burner for combustion of the retentate to provide the required energy for NH3decomposition. Pd / Ag membranes were purchased from Media and Process Technology Incorporation. The membranes are prepared on the external surface of 5.7 mm OD α-alumina hollow fiber supports. The 1 wt.% Ru-based catalyst was prepared using a sequential impregnation procedure. The performance characteristics of this catalyst were reported previously. It is important to note that catalyst was reduced before charging into the reactor by heating at 500 °C for 2 h in 200 sccm 30 % H2 in N2. The H2 catalytic burner was fabricated by coating the inside surface of the burner tube with a thin Pd layer, and a perforated ⅛ inch tube was used for air / H2 mixture distribution in the burner tube. Details and characteristics of the burner can be found in our previous work.

[0170] 2.2. Fabrication and optimization of the CMR

[0171] To enhance the heat transfer from the H2 burner to NH3 catalyst and to make it easier and practical for welding and operation, the design for the CMR was slightly modified. The reactor housing is basically a 2-inch, square stainless-steel (SS) box. At one end, it has two openings for NH3feed inlet and an outlet for the retentate (H2, N2, and undecomposed NH3). The other end is sealed with graphite sealing and flange bolts. This end is known as the H2 collecting chamber as it is connected to four 24-cm long Pd / Ag membranes using graphite sealing and Swagelok fittings. The membranes were tested with ethanol to detect potential leakages from both the sealing and the Pd / Ag membrane surface. At the center of the housing, a ½ inch SS tube was coated with a Pd layer and sealed at its end as the H2burner. A ⅛ inch perforated SS tube inserted in the burner tube was used to finely divide the air / retentate mixture for combustion in the burner (Figure 42). The outcoming gases and reaction products, exhaust, will heat the inlet H2 / air mixture and thus enhance the burner efficiency. The heat generated in the burner is transferred to a surrounding 1-inch2channel for NH3catalyst bed. A small amount of catalyst was placed just after the permeate to fully decompose any residual NH3 in the produced H2stream. Catalyst was tightly fixed at the end of the square channel using silica wool·NH3 preheating greatly affects the reforming efficiency. A ⅛ −inch SS tube was used to preheat NH3by making use of the heat escaping from the surface of the external SS housing. The tube was wrapped around the reactor channel and connected to NH3 inlet, as shown in Figure 43. Heating tape was then wrapped above the preheating tube to initiate the reactor and heat it to a temperature of 350–450 °C. To limit the heat transfer to the surrounding atmosphere, the reactorwas sealed tightly with silica wool and silicon tape and then enclosed in a silica foam package, as shown in Figure 43.

[0172] 2.3. Testing and analysis of the CMR

[0173] Figure 34 shows a schematic of the experimental setup for testing the developed CMR, and a photograph is given in Figure 44. Leak test with N2 gas was conducted before the initiation of the CMR at elevated pressure (up to 6 bar) at room temperature and confirmed no leakage. The reactor was initiated by slow heating using a wrapped heating tape, while monitoring the reactor surface temperature and the temperature at the center of the burner tube using thermocouple probes attached to the surface and placed inside the burner tube. As shown in Figure 45, the temperature at the surface reached 450 °C in 30 min and stayed constant at that temperature, while the temperature inside the burner tube increased at a slower rate reaching 350 °C in 60 min. At this moment, NH3was fed to the reactor at the desired flowrates controlled using a high-capacity flow meter in the range 1-3 SLPM. Initially the back pressure regulator was fully open, so that the decomposition product (H2and N2) was completely directed to the retentate where it was mixed with sufficient air for combustion in the burner. After 30 min, the back pressure regulator was slightly closed to raise the reactor pressure, while keeping the permeate pressure at 1 atm, to control the fraction of the generated H2 permeating through the membranes. Air flowrate was then controlled for combustion of the remaining H2 in the retentate. The fabricated system was tested at varying NH3 flowrates in a self-sustaining mode to explore the optimum conditions for the operation of the reactor to maximize the recoverable H2. Air was supplied to the burner to maintain a H2 / O2ratio of 2, and then the effect of slight variations in air flowrate in the burner feed was also investigated. Additionally, the reactor was running continually for 100 h with 13 cycles, excluding the time for initiation and shutdown of the system, to determine the effect of heating and cooling on the system efficiency. Cooling down and preheating the system was done under nitrogen atmosphere to avoid embrittlement of the membranes.

[0174] For each data point, the system was operated continuously for 2 h, and average was calculated and reported, showing variation of <1 %. The retentate was continuously analyzed using a mass spectrometer, Omnistar® GSD 350 from Pfeiffer Vacuum Inc., using a needle valve to direct a tiny fraction for gas analysis. The permeate was also frequently analyzed using a mass spectrometer, and small content of NH3in the permeate was monitored usingAutomation Technologies Online, ATO, handheld NH3gas detector (model: GD200-NH3,0– 50 ppm) with a detection limit of 0.01 ppm. The produced H2 is estimated based on the measured permeate stream using a bubble flow meter and H2composition in the permeate. The exhaust was also monitored for NOx content using nitric oxide detector (model: HNAG900-NO-T, 0– 50 ppm portable detector). NH3conversion (XNH3), recovered H2, and reforming efficiency were calculated using Eqns. (1), (2), (3):Where ^^^ denotes the molar flow rate, NH3Inrepresents NH3 supplied to the system, and NH3Outrepresents the total NH3present in the permeate and the retentate sides of the membrane. H2Permeateand H2Retentatecorrespond to H2 in the permeate / product stream and H2 in the retentate as the fuel, respectively. The molar lower heating value (LHVmol) for NH3and H2are considered to be 316.8 and 241.8 kJ / mol, respectively.

[0175] 2.4. Energy and mass balance of the CMR

[0176] Figure 35A shows the process flow diagram for the self-sustained NH3 decomposition system. NH3 is provided from a storage tank and passes along two crossflow heat exchangers to recover the waste heat from the hot permeate stream and the exhaust gas out of the H2 burner. The preheated NH3 is fed to the reactor, where NH3 cracking takes place and the energy is provided from the H2burner. The permeate is collected and compressed as a product, and the retentate is mixed with air for combustion in the burner. Aspen Plus was used to model the process and investigate its feasibility and estimate the maximum recoverable H2 at steady state conditions applying Peng-Robinson (PR) fluid package. Figure 35B shows the process flowsheet, showing separate blocks for the cracker, the burner, and the membrane. Theprocess was designed based on the overall process shown in Figure 35A. A conversion reactor is used to simulate NH3 cracking into H2 and N2 where conversion is assumed complete at 450 °C and 4 bar for reaction (1) (R1). A component splitter is used to separate H2from the cracker product to the permeate, and remaining H2 in N2 is mixed with air and fed to another conversion reactor for H2combustion according to reaction (2) (R2). Crossflow heat exchangers are utilized for preheating NH3 feed using the waste heat from the permeate and burner exhaust. Minimum outlet temperature difference approach is assumed 14–15 °C. 2NH3→ N2+ 3H2ΔH∘= +46 kJ / mol NH3(R1) H2+ 1 / 2 O2→ H2O ΔH∘= −285.8kJ / mol H2(R2)

[0177] 3. Results

[0178] 3.1. NH3 CMR performance characteristics

[0179] Figure 36 shows the performance of the developed CMR when tested with different NH3flowrates, as a function of the recovered H2. The recovered H2and reforming efficiency were low when 1SLPM NH3 was fed to the reactor; although NH3 conversion increased with the decreasing H2recovery, the reforming efficiency was still low (<65 %). This is due to the size of the reactor that needed more H2 in the retentate to maintain a temperature high enough to ensure almost complete NH3decomposition. For NH3flowrates of 2-3SLPM, there was a trade-off performance line between NH3 conversion and recovered H2, as clearly noticed in Figure 36A. At low recovered H2(<50 %), NH3conversion was almost complete (>99 %). At higher recovered H2, the conversion started to decrease gradually with the increase of the recovered H2. Maximum recovered H2was ∼52 % for the tested flowrates (2-3SLPM), while maintaining NH3 conversion of >98 %. More H2 exiting the reactor through the membrane rather than to the retentate came at the cost of significant reduction in NH3conversion. This trade-off between H2 recovery and NH3 conversion is essential when determining the optimum operation conditions for the reactor.

[0180] The H2recovery and NH3conversion are expected to be related directly to the system pressure and temperature. The H2 permeability of Pd-based membranes is significantly influenced by reactor pressure and temperature, which can be described by Sieverts' Law and Arrhenius equation (Eqns. (R1), (R2)). H2 diffusion through Pd-based membranes follows asolution-diffusion mechanism, where the H2flux through the membrane, J, is described by Eq. (4): (4)where and stand for the membrane permeability and thickness, respectively, and and are the H2 partial pressure at the feed and permeate sides of the membrane, respectively. n is the pressure exponent, conventionally 0.5 for a bulk diffusion dominant transport, owing to the rate-limiting step when H2atoms diffuse through the dense metal layer.

[0181] Moreover, H2transport through Pd-based membranes is an activated process. The relationship between H2 permeability and temperature is explained by Eq. (5): (5)where ^^0 is a material specific cy high temperature, ^^^is the activation energy for H2diffusion, ^^ is the universal gas constant, and ^^ is the absolute temperature.

[0182] Figure 37. shows the temperature at the surface of the reactor inside the burner tube and the absolute pressure in the reactor at different operation conditions. The temperature inside the burner tube is basically dependent on the balance between the H2 in the retentate and the NH3fed to the system. Initially, the reactor pressure was atmospheric, and all decomposition products were directed for combustion in the burner and thus NH3 conversion was almost complete (data not shown in Figure 35A). Because NH3decomposition is endothermic, NH3 decomposition is favored at high temperatures and low pressures. As noticed, the increase in the recovered H2was attained by increasing the reactor pressure for each flowrate tested. In the meantime, the temperature of the system decreased, due to less H2 combustion in the burner. Both parameters affected NH3conversion and the H2permeation through the membrane. By comparison with the performance data in Figure 35A, a good performance was obtained when the temperature inside the burner tube was above 600 °C and the surface temperature of the reactor was above 400 °C, regardless of the pressure. This temperature range seems to ensure the catalyst bed was about 450 °C for better performance of the catalyst. The increase in the reactor pressure with sufficiently high temperature produced a higher H2 partial pressure in the reactor.This will in turn enhance the activation processes and the driving force for H2permeation through the membrane, based on Eqns. (R1), (R2).

[0183] While a pressurized feed is beneficial for H2 production in a CMR, it adversely affects the NH3cracking in the catalytic reaction. It is well documented that NH3conversion is not favorable at higher pressure according to Le Chatelier’s principle, and kinetics will be decreased as well with the decrease of the reactor temperature. H2permeation through the membrane will improve the kinetics and shift the thermodynamic equilibrium, therefore the negative effect of pressure build-up is equilibrated by the positive effect of H2 removal. H2removal from the reaction zone also decreases inhibitory effect of the high H2partial pressure on the forward kinetics of NH3 cracking reaction as per the Temkin-Pyzhev type rate mechanism. Thus, the effect of increasing the system pressure will not be significant if the system temperature is still high enough to ensure high NH3 conversion (>98 %). Only when the system temperature decreased significantly due to less H2directed to the burner, did NH3 conversion exhibit a noticeable reduction, as shown in Figure 35A. The reduction of NH3conversion increased with the increase of the NH3flowrate at high recovery, which is a common phenomenon in CMR. This may be attributed to the fact that the residence time of NH3in the catalyst bed decreases, resulting in slightly lower NH3conversion. Also, the time for the produced H2 to get in contact with the separation membrane also decreases so that a relatively smaller amount of H2permeates through the membrane, leading to increased H2 production rate but decreased H2 recovery.

[0184] The reforming efficiency is based on a balance among the recovered H2, energy provided to the system by retentate combustion, and the energy extracted from the system to decompose NH3feed. A linear relationship was found between the reactor reforming efficiency and the recovered H2, as shown in Figure 36B, with the increase of the recovered H2 up to ∼52 %. At constant NH3 flowrate, increasing the recovered H2 led to an increase in the reforming efficiency according to Eq. (3). However, an additional increase in the recovered H2 led to a decrease in the temperature of the reactor. At optimum conditions, the reforming efficiency of the system reached 73 %. The reactor operation was maintained within acceptable conversion limits for NH3, ensuring that the drop in reforming efficiency was not significant. The reforming efficiency accomplished in this work was higher, compared to self-sustained NH3 reactors with no H2 separation using NH3 or H2 as a fuel, where they reported 57.3 % and 70.95 %, respectively. However, it was lower compared to our previous self-sustainedNH3 reactor, achieving ∼82 % reforming efficiency at the optimum operating conditions. The reason behind this is the larger scale reactor and larger NH3flowrate (3 SLPM, compared to 0.6 SLPM) and thus more energy escaping the system either in the permeate or the retentate in addition to the energy escaping in the reactor exhaust. Efficiency is expected to increase with improved energy recovery.

[0185] 3.2. NH3CMR product purity

[0186] Figure 38 shows the purity of the produced H2and the NOxemissions for the burner exhaust of the developed CMR as a function of the recovered H2 and NH3 flowrate. The H2product permeating through the membrane showed a H2purity of 99.7–99.3 % and NH3 content below the minimum NH3 detector readability level (0.01 ppm) under all the testing conditions. This reflects the quality of the used Pd / Ag membranes and absence of major membrane defects or sealing issues. The permeate purity reached 99.7 % at low operation pressure and high reactor temperature, attributing to the increased NH3conversion and activated H2 diffusion and thus higher flux at higher reactor temperature. Although H2 purity remained very high in the entire investigated pressure range at different flowrates, N2 concentration in the H2stream increased with increasing H2recovery or system pressure. This indicates that at high pressure there is a tiny leak through the membrane. For dense metallic membrane, H2 permeance increases with the square root of the pressure difference, while N2permeance linearly increases with the increase of the pressure difference in case of defects. As a result, pressure increase may affect N2permeance more, compared to H2. Therefore, the increase in the operating pressure caused a slight decrease in the purity of the produced H2, although larger amounts of H2 were recovered. This is quite frequent in membrane separation and can hardly be avoided unless thicker membranes are used.

[0187] NH3, a lighter molecule than N2, is expected to have a larger flux through the defects in the membrane (by a factor of 1.28), assuming the governing transport mechanism is Knudsen diffusion. Some studies reported ultrahigh purity H2using CMRs, >99.99 % H2and >99.97 % H2, and very low NH3 concentration (<0.1 ppm), based on the detection limits of their gas composition analysis equipment. Other studies showed that H2purity was not always sufficiently high for directly feeding to a PEM fuel cell, with NH3 concentration of 25– 1,000 ppm. Various strategies have been investigated to increase the purity of the produced H2 and lower the residual NH3 concentration to meet the feed specification of PEM fuel cells.One effective strategy is the addition of an H2purification step downstream of the membrane reactor to remove the undesired NH3 using a 13X zeolite or clinoptilolite adsorption bed. Although this solution adds more complexity to the system, it could be an economically viable option. Our solution of a small dose of the catalyst just after the membrane helped decompose residual NH3. This explains the absence of NH3in the permeate, while ppm level of N2existed. This finding suggests that the reactor product could be used directly for PEM fuel cell applications as it is, with no extra processing.

[0188] In the meantime, NOx concentration in the burner exhaust was less than 25 ppm at the worst condition, as shown in Figure 38. At 1SLPM, NOxcontent was always below 2 ppm. It increased with the increase of the recovered H2 when NH3 feed was constant and above 2SLPM. This could be due to the undecomposed NH3in the retentate stream that increased with the increase of the recovered H2. NOx concentration also increased with the increase of NH3flowrate due to less residence time of NH3in the reactor.

[0189] 3.3. Effect of mixing air in H2burner on NH3CRM performance

[0190] The CMR was tested using 2.6 SLPM NH3 flowrate and keeping the recovered H2constant, while varying the air flowrate to the burner. As shown in Figure 39, H2purity did not change with the change of the air flowrate, because H2 purity depends primarily on the membrane quality. NH3conversion slightly decreased with the decrease of the air flowrate. The reason behind this change is the change in the reactor temperature manifested as a noticeable reduction in the temperature inside the burner tube, as shown in Figure 39. This is attributed to less available oxygen (H2 / O2 > 2) for H2 burning. The increase of air flowrate, on one hand, is beneficial initially for combustion of the small content of undecomposed NH3in the retentate, leading to increased burner temperature. On the other hand, the NOx content in the exhaust increased. More increase in air flowrate will eventually lead to a decrease in the burner temperature due to the increase in the inert gas flowrate flowing in the burner tube. Therefore, it seems reasonable to maintain the H2 / O2ratio at 2 or slightly lower.

[0191] 3.4. CMR cyclic performance

[0192] The major challenges of Pd-based membranes include H2 embrittlement at lower temperatures due to increased H2solubility and durability issues at higher temperatures caused by intermetallic diffusion and Pd layer morphological changes. However, recent studies suggestthat these problems can be mitigated by optimizing operating conditions. The CMR was tested continually for 100 h, excluding initiation and cooling down times, with 13 startup / shutdown cycles using a 2.6 SLPM NH3flowrate, and the performance data are summarized in Figure 40. The membrane module was purged with N2 during heating and cooling to protect the Pd / Ag membranes and to prevent the reoxidation of the Ru-based catalyst. The data demonstrates stable performance with negligible variations in NH3 conversion and H2 recovery. Consistent H2recovery of ∼52 % and H2purity of >99.5 % were achieved at NH3conversion of >99 %. NH3concentration in the permeate was intermittently measured several times, and the reading on the analyzer showed 0 ppm, suggesting NH3 concentration is less than 0.01 ppm, the detection limit of the meter. The detection of very low NOx concentrations in the burner exhaust highlights the efficiency of the catalytic process and suggests minimal environmental impact.

[0193] 3.5. Process simulation and energy evaluation

[0194] The model in Figure 35A shows the proposed flowsheet for the self-sustained NH3 cracking system. As shown, the energy needed for cracking is provided directly from the burner. The burner is operated based on combustion of the fraction of H2 in the retentate and the provided air. Since NH3feed is constant at 2.6 SLPM, the amount of energy needed for preheating and conversion of NH3 at 450 °C is constant. The energy required for preheating and full cracking of NH3at 450 °C is 137.6 W. Part of this energy can be provided by extraction of the waste heat from the permeate and the burner exhaust using heat exchangers. So, the model was used to determine the minimum fraction of H2in the retentate that upon combustion can provide the required heat and the heat exchangers still work satisfactorily. The data extracted from Aspen Plus model showed that the maximum H2that can be recovered from the system, while maintaining stable operation, is 80 %. Also, the simulation revealed the importance of the heat exchange process by saving ∼25.6 % of the required heat for full cracking of NH3feed. Unfortunately, this part of energy was lost in our experiments, since no heat exchange was utilized. The flowsheet also assumes ideal steady state conditions and perfect insulation, and the retentate maintains its temperature before mixing with air, which is reflected as improved burner operation. Adjusting the model to ignore the role of heat exchangers and cooling down of the retentate as in our experimental system revealed that about 69 % of the generated H2 can be recovered. This indicates that a major part of energy loss is due to imperfect insulation of the reactor, and this aspect will be considered in future work.

[0195] 4. Conclusions

[0196] A self-sustained NH3decomposition CMR, integrated with Ru-based catalyst, Pd / Ag membranes, and an H2 burner, was successfully designed and fabricated, eliminating the need for an external heating source. Self-sustained operation was demonstrated using NH3 flowrates of 1–3 SLPM, and optimal operating conditions were determined based on a trade-off between the recovered H2and NH3conversion. ∼52 % of the generated H2could be recovered in the permeate stream at NH3 conversion of >98 %. The produced H2exhibited >99.3–99.7 % purity and NH3levels below 0.01 ppm, the detecting limit of the meter, regardless of the reactor pressure. Negligible NOx emissions were observed, suggesting that the system could meet stringent environmental regulations, making it suitable for sustainable power generation solutions. The developed CMR demonstrated stable performance for over 100 h and endured 13 cycles of cooling and heating, while producing ∼0.25 kg / day H2·NH3conversion, H2purity, and H2flux remained nearly constant during this period, indicating the stability of both Pd membranes and Ru-based catalyst. The stable production of high-purity H2 presents a promising opportunity for integrating this CMR into fuel cell systems for electricity generation and will contribute to the progress of applications requiring carbon-free H2 storage and production.

[0197] 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.

Claims

We claim:

1. A portable device for generating hydrogen from ammonia, comprising: a first reactor layer having an ammonia inlet, a retentate port, and a chamber containing an ammonia decomposition catalyst; a first gas-collecting layer having a manifold with a hydrogen outlet; a first hydrogen-selective membrane disposed between the first reactor layer and the first gas-collecting layer and configured such that hydrogen gas generated in the chamber of the first reactor layer will permeate through the hydrogen-selective membrane into the manifold of the first gas-collecting layer; and a burner layer adjacent to the first reactor layer and separated from the chamber by a first conduction plate having a palladium coating on a burner side, the burner layer having an intake port and an exhaust port, and wherein the intake port is in fluid connection with the retentate port of the first reactor layer.

2. The portable device of claim 1, further comprising a sealing member at each of one or more of (a) an interface of the first reactor layer and the first hydrogen-selective membrane and (b) an interface of the first gas-collecting layer and the first hydrogen-selective membrane.

3. The portable device of claim 2, wherein the sealing member is a gasket, such as a graphite gasket.

4. The portable device of claim 1, wherein the chamber contains the ammonia decomposition catalyst.

5. The portable device of claim 4, wherein the ammonia decomposition catalyst comprises ruthenium.

6. The portable device of claim 4, wherein the chamber further contains an inert material configured to confine movement of the ammonia decomposition catalyst.

7. The portable device of claim 6, wherein the inert material is disposed between the first hydrogen-selective membrane and the ammonia decomposition catalyst.

8. The portable device of claim 6, wherein the inert material is silica wool.

9. The portable device of claim 1, wherein the first hydrogen-selective membrane provides near- absolute selectivity for hydrogen with near complete exclusion of nitrogen and ammonia.

10. The device of claim 1, wherein the first hydrogen-selective membrane comprises palladium or a palladium silver alloy.

11. The device of claim 1, wherein the chamber of the first reactor layer is configured to operate at a temperature of 400 °C to 450 °C.

12. The device of claim 1, wherein the chamber of the first reactor layer is configured to operate at a pressure of 1-4 bar, inclusive.

13. The device of claim 1, further comprising a heat exchanger having a first channel in fluid communication with the exhaust port of the burner layer and a second channel in fluid communication with the inlet of the first reactor layer, wherein the heat exchanger is configured to use an exhaust gas from the burner layer to pre-heat ammonia gas supplied to the first reactor layer.

14. The device of claim 1, further comprising: a second reactor layer adjacent to the burner layer, the second reactor layer having an ammonia inlet, a retentate port, and a chamber configured to contain a ammonia decomposition catalyst; a second gas-collecting layer having a manifold with a hydrogen outlet; a second hydrogen-selective membrane disposed between the second reactor layer and the second gas-collecting layer and configured such that hydrogen gas generated in the chamber of the second reactor layer will pass through the second hydrogen-selective membrane into the manifold of the second gas-collecting layer; and wherein a second conduction plate separates the burner layer from the chamber of the second reactor layer.

15. The portable device of claim 14, further comprising additional sealing members at each of one or more of (a) an interface of the reactor layer and the hydrogen- selective membrane and (b) an interface of the gas-collecting layer and the hydrogen- selective membrane.

16. The portable device of claim 15, wherein the additional sealing member is a gasket, such as a graphite gasket.

17. The portable device of claim 14, wherein the chamber of the second reactor layer contains an ammonia decomposition catalyst.

18. The portable device of claim 17, wherein the ammonia decomposition catalyst comprises ruthenium.

19. The portable device of claim 17, wherein the chamber of the second reactor layer further contains an inert material configured to confine movement of the ammonia decomposition catalyst.

20. The portable device of claim 19, wherein the inert material is disposed between the second hydrogen-selective membrane and the ammonia decomposition catalyst.

21. The portable device of claim 19, wherein the inert material is silica wool.

22. The portable device of claim 14, wherein the second hydrogen-selective membrane provides near-absolute selectivity for hydrogen with near complete exclusion of nitrogen and ammonia.

23. The device of claim 14, wherein the second hydrogen-selective membrane comprises palladium or a palladium silver alloy.

24. The device of claim 14, wherein the chamber of the second reactor layer is configured to operate at a temperature of 400 °C to 450 °C.

25. The device of claim 14, wherein the chamber of the second reactor layer is configured to operate at a pressure of 1-4 bar, inclusive.

26. The device of claim 14, further comprising a heat exchanger having a first channel in fluid communication with the exhaust port of the burner layer and a second channel in fluid communication with the inlets of the first reactor layer and the second reactor layer, wherein the heat exchanger is configured to use an exhaust gas from the burner layer to pre-heat ammonia gas supplied to the first reactor layer and the second reactor layer.

27. A hydrogen fuel cell comprising a device according to any one of claims 1-26.

Images