Frea
Van Steenweghen
,
Lander
Hollevoet
and
Johan A.
Martens
*
Surface Chemistry and Catalysis: Characterisation and Application Team (COK-KAT), KU Leuven, Leuven BE-3001, Belgium. E-mail: johan.martens@kuleuven.be
First published on 6th February 2024
A process for producing ammonia out of NOx from hydrogen engine flue gases is proposed. NOx is captured on a lean NOx trap (LNT) and catalytically reduced with hydrogen to ammonia (NOCCRA). The energy requirement is similar to that of Haber–Bosch processes. NOCCRA is attractive for decentralised green NH3 production.
The natural gas-based HB process in which N2 from air is reduced with H2 from steam methane reforming (SMR) is very energy-efficient, with an energy cost as low as 0.48 MJ molNH3−1.2,5,6 The use of natural gas as an energy and H-atom source entails a CO2 emission of ca. 1.6 tonne per tonne of NH3 produced.5,6 An obvious way of rendering NH3 synthesis more sustainable is by using green instead of grey hydrogen. An HB plant, running with hydrogen from water electrolysis using renewable electricity, has an estimated energy cost of ca. 0.65–0.70 MJ molNH3−1.2,10
Powering the HB process with renewable energy is challenging because of the large scale at which the process is cost-effective. The economy of scale of HB results from the need for a reaction pressure of 10–40 MPa at 400–650 °C. Such reaction conditions impose a requirement of continuous operation3,11,12 which does not align well with variable renewable electricity supply.2,6 Furthermore, the application on land of N-fertiliser of ammonia and ammonium nitrate (NH4NO3) derived from it is highly decentralised.13 This makes small-scale ammonia production under mild reaction conditions with flexible production schemes, despite a slightly higher energy cost, a viable alternative to the centralised HB process.1,3,14 Local production complementing centralised production could solve supply chain problems and price volatility, especially for remote farming areas.15
Several concepts for the synthesis of ammonia from atmospheric N2 using renewable energy sources have been proposed, like (i) direct electrocatalytic reduction,16,17 (ii) plasma-enabled synthesis,18 and (iii) chemical looping,19,20 as documented in reviews.4,6,21 Electrocatalytic N2 reduction suffers from very low yield.22 Plasma processes have an energy cost many times higher than those of HB processes.23 Chemical looping is facing some challenges related to mass transfer, cyclability, material volumes and cost.21,24 An additional option for producing ammonia is to extract and convert N-sources contained in side products and waste streams from the agro-industry, contributing in this way to N-circularity.25–27
The N2 molecule is very difficult to activate for chemical reaction. Oxidation of N2 molecules to nitrogen oxides (NOx) or nitrates (NO3−) leads to more reactive N-species. The dissociation energy of the N–O bond in the NO molecule of 204 kJ mol−1 (ref. 28) is much lower than the energy needed for cleaving the triple NN bond in the N2 molecule (942 kJ mol−1 (ref. 29)).30 In an approach proposed earlier30,31 atmospheric N2 is oxidised using a plasma process.27,30 The downside of oxidising first and performing the reduction to ammonia in a second step is the need for more hydrogen molecules per ammonia molecule (eqn (1) compared to 2). Nevertheless, producing ammonia according to the reaction of eqn (1) could be attractive because of the less severe reaction conditions.11,25,30,31
2NO + 5H2 → 2NH3 + 2H2O | (1) |
N2 + 3H2 → 2NH3 | (2) |
4NO + 4NH3 + O2 → 4N2 + 6H2O | (3) |
2NO2 + 4NH3 + O2 → 3N2 + 6H2O | (4) |
NO + NO2 + 2NH3 → 2N2 + 3H2O | (5) |
NH2CO2NH2 + H2O → 2NH3 + CO2 | (6) |
One of the commercial NOx abatement technologies for lean-burn automobile combustion engines is the lean NOx trap (LNT) also known as the NOx storage and reduction (NSR) system or NOx adsorber.41 The washcoat of an LNT monolith typically consists of platinum and barium oxide supported on alumina (Pt–Ba/Al2O3).42,43 An internal combustion engine equipped with LNT is alternatingly run on lean and rich fuel mixtures. During the lean phase, NOx is trapped on the LNT and converted to barium nitrate (Ba(NO3)2) (eqn (7)). During the rich phase, the LNT is regenerated by chemical reduction of the trapped nitrates with reductants produced in the engine (eqn (8)).42 Among H2, CO and hydrocarbon reductants in the exhaust from rich fuel mixtures, at low temperatures H2 is most reactive.44
In the pollution abatement application, N2 is the desired reaction product (eqn (8)). NH3 is a reaction intermediate in the two-step formation of N2 (eqn (9) and (10)), while N2O is an unwanted by-product (eqn (11)).43,45–48 The reaction of eqn (10) is slow and rate determining for the formation of N2 at temperatures below 150 °C.48 Limiting the conversion of Ba(NO3)2 to the first step (eqn (9)) is a way to produce ammonia. High NH3-selectivity of nitrate reduction on LNT catalysts in the range of 75–90% has been reported.47–52 Provided green hydrogen is available, the ammonia produced in this way on an LNT can be qualified as “green”.
BaO + 2NO2 + ½O2 → Ba(NO3)2 | (7) |
Ba(NO3)2 + 5H2 → N2 + BaO + 5H2O | (8) |
Ba(NO3)2 + 8H2 → 2NH3 + BaO + 5H2O | (9) |
3Ba(NO3)2 + 10NH3 → 8N2 + BaO + 15H2O | (10) |
Ba(NO3)2 + 4H2 → N2O + BaO + 4H2O | (11) |
Flue gases of stationary emission sources are considered as a NOx source for a NOx capture and catalytic reduction to ammonia process, NOCCRA. The use of NOx emission sources for ammonia synthesis has three major advantages: (i) the use of an inexpensive source of NOx which otherwise would entail a cost of elimination, (ii) resolving the NOx emission issue of flue gases, and (iii) in contrast to state-of-the-art NOx abatement technologies, NH3 is now produced instead of consumed for reducing NOx to N2.
NOx-containing flue gas is the feed of the NOCCRA process (Fig. 1). It is sent over an LNT to selectively chemisorb the NOx. During the subsequent reduction phase, nitrate molecules stored on the LNT are reduced with green H2 from an electrolyser to ammonia which desorbs spontaneously. NH3 is separated from the outlet gas stream by phase transfer to an aqueous solution in a washing column. In this way, residual H2 remaining in the gas phase can be recovered and reused. The LNT alternates between phases of NOx trapping and chemical reduction. For achieving continuous ammonia production with NOCCRA at least two LNT units, alternating between NOx adsorption and NH3 formation, are needed.
Experimental details of the laboratory setup are provided in the ESI (sections 1 and 2†). Pt/Ba/Al2O3 catalyst pellets were prepared as described earlier.31 The Pt/Ba/Al2O3 weight ratios were 1/20/100. The performance of Pt/Ba/Al2O3 pellets in a NOCCRA cycle was investigated in a tubular flow reactor described earlier.31 More information about the automated reactor set-up can be found in ref. 53. A flue gas mimic composed of 200 ppm NO, 5% O2 and 1.5% H2O in N2 carrier gas was fed to the reactor at a gas flow rate of 0.1 mL h−1 g−1 for 250 s. Chemical reduction of adsorbed NOx was performed using a gas mixture of 5% H2 with 1.5% H2O in N2 carrier gas at the same gas feed rate and temperature during 1800 s. The temperature was varied in the range of 75–200 °C. The operating pressure was always atmospheric.
At temperatures above 100 °C all NOx in the feed was captured on the Pt/Ba/Al2O3 LNT material (Fig. 2A). The NH3-selectivity showed a maximum of 84% at 125 °C (Fig. 2B). The N2O-selectivity showed a minimum of 1.7% at 150 °C. Over the investigated reaction temperature range, the N2O-concentration in the product attained an average of ca. 0.4 ppm, which is four orders of magnitudes lower than the typical N2O concentration in the flue gas of nitric acid plants.54,55 These results confirm the feasibility of producing ammonia from flue gas on an LNT.
NOx adsorption and its reduction to NH3 using hydrogen are exothermic processes.42 It is envisioned to use hot flue gas as a feed such that NOCCRA will be self-sufficient in terms of heating. Energy is consumed in the production of hydrogen gas, needed for reducing the adsorbed NOx. The NOCCRA energy consumption essentially depends on the energy efficiency of the electrolyser and the NH3-selectivity of nitrate reduction reached on the LNT. As excess, unconverted H2 is separated from the produced NH3 in the washing column and recirculated in the NOCCRA process, near-complete utilization of H2 can be achieved. Stoichiometrically 4 moles of H2 are needed for producing 1 mol of NH3 (eqn (8)). Hence, assuming 100% NH3-selectivity of the barium nitrate reduction reaction, the hydrogen production in an electrolyser with an efficiency of 70%56 would need 1.37 MJ per mol NH3 (estimations detailed in the ESI, section 3†). At an NH3-selectivity of 84% reached experimentally at 125 °C (Fig. 2B), the energy consumption amounts to 1.50 MJ molNH3−1. In earlier work, a related process called PNOCRA (“plasma nitrogen oxidation coupled with catalytic reduction to ammonia”) was proposed, in which the feed consists of NOx generated from air by a plasma reactor.30,31 The plasma process producing NOx requires energy, lifting the total energy cost of ammonia production with PNOCRA to 2.10 MJ molNH3−1.31
Natural gas-based HB processes are run at an energy cost of 0.47–0.71 MJ molNH3−1.2,5 Green HB processes in which H2 is generated by H2O electrolysis have energy costs of 0.65–0.70 MJ molNH3−1.2,10 While NOCCRA is more energy demanding, its benefits are rather indirect. Besides producing ammonia, NOCCRA can be considered to be a depollution technique for NOx emissions. State-of-the-art NOx emission abatement techniques, such as NH3-SCR and SNCR, consume NH3, while NOCCRA produces it. NH3-SCR and SNCR require 1 mol of NH3 to eliminate 1 mol of NOx.37,38 In this way, the NOCCRA process saves 1 mol of NH3 and generates 1 mol of NH3. Taking this benefit into consideration, the energy requirement of NOCCRA is reduced to 0.75 MJ molNH3−1, making it more competitive with HB processes (Fig. 3).
![]() | ||
Fig. 3 Energy requirement [MJ molNH3−1] of the ammonia synthesis processes: natural gas-based Haber–Bosch,2 electrified Haber–Bosch,2 PNOCRA31 and NOCCRA with and without accounting for the savings of NH3 by avoiding the need for an S(N)CR process for NOx abatement. |
Besides the operating cost of which energy is the largest share, the installation cost of NOCCRA determines the economic viability. In this early stage of research and development, a detailed techno-economic analysis would be inaccurate. Qualitative comparison of NOCCRA with competing processes hints at its potential for practical application. NOCCRA, being an integration of green ammonia production and NOx abatement, intrinsically has the potential to lower the equipment cost. Compared to the HB process for ammonia synthesis, NOCCRA operates at much lower pressures and temperatures. Long lifetimes are predicted for the lean NOx trap of NOCCRA based on experience in the automotive industry, but the downside of this benefit is the use of a noble metal-based catalyst (Pt).
In the NOCCRA process, ammonia is recovered at the outlet of the LNT using a gas scrubber (Fig. 1). The scrubber can be run using water, sulphuric acid (H2SO4) or nitric acid (HNO3) to produce ammonium sulphate ((NH4)2SO4) or ammonium nitrate (NH4NO3), respectively.3 The product is directly applicable as fertiliser. Concentration of the ammonia is not needed for that application, implying no additional energy cost.57 In this manner, NOCCRA also avoids corrosive substances like anhydrous ammonia, which enables the use of low-cost materials for reactors and conducts.
Further potential advantages of NOCCRA are related to the intermittent character of renewable energy supply. An electrified HB process will produce green NH3 in large-scale, highly centralised production facilities,58 which require a steady supply of vast amounts of green electricity. However, when green electricity is intermittently available, and when NOx emissions are present, NOCCRA becomes attractive for the distributed small- and medium-scale production of NH3-based fertilisers.16,17 Intermittent green hydrogen production and storage are easy to accommodate in small plants running a NOCCRA process. Given that locally produced NH3 with NOCCRA is not intended to be stored over long periods nor transported over long distances, it can be applied directly to fields via fertigation, or irrigation with fertiliser solutions.11,59 The commercial use of LNT technology in the automotive industry provides evidence for scalability.60
One limitation of NOCCRA is the need for a local NOx source. Stationary sources are qualified, while mobile applications such as automobiles would be impractical due to the need for on-board ammonia storage. Another limitation is that agricultural activities need to be present within the vicinity of the NOx source to provide a market for the ammonia fertiliser product.
The exhaust from fossil fuel-based combustion processes contains sulphur oxides (SOx) which are poisons of LNT catalysts so that they require periodic regeneration.61,62 The transition towards green energy carriers, such as hydrogen gas, offers a futureproof perspective overcoming this sulphur problem. Combustion of hydrogen causes thermal NOx formation, but the absence of CO2 and SOx in the exhaust gas makes it an ideal inlet gas feed for the NOCCRA process.63,64
The potential of NOCCRA for distributed production of ammonia is illustrated by the following case of a hydrogen-fuelled combustion engine for electricity generation.65 Small-scale internal engines, such as gas turbines or combined heat and power systems, with a capacity of 100 kW and powered by hydrogen gas at 100%, have a flue gas flow rate of 22400 m3 h−1.66 Flue gases of such hydrogen engines typically contain NOx at an average concentration of 500 ppm.65 Such a NOx source represents a potential NH3 production capacity of 62.5 tonnes per year, assuming an NH3-selectivity of 84% (Fig. 2B). This local NH3 production by one industrial NOCCRA plant would meet the yearly fertiliser demand of ca. 730 ha of cropland,13 assuming an average nitrogen fertiliser demand of 86 kgN ha−1 in the European Union67 (estimation provided in the ESI, section 4†).
The energy cost of ammonia production with NOCCRA is estimated at 1.50 MJ molNH3−1. Considering that NOCCRA avoids NH3 consumption for abating NOx emissions the net energy cost becomes competitive with electrified HB processes. The NOCCRA process serves as both a green ammonia production method and a NOx abatement technique, consolidating two distinct processes into a single one, thereby offering potential advantages in terms of reduced installation cost.
NOCCRA enables the small-scale decentralised production of green ammonium nitrate fertiliser in remote farming areas. NOCCRA uses intermittently available renewable energy sources. A convincing example of a hydrogen engine from which the NOx emission is used to produce green ammonium nitrate is presented.
Some scientific challenges remain to make NOCCRA fully competitive, including better matching of the two phases of the NOCCRA cycle and catalyst development to enhance the NH3-selectivity and productivity of the precious metal catalyst. Platinum is a high-cost precious metal, and therefore, replacing it with Earth abundant metals would reduce the cost of the NOCCRA process.68
Footnote |
† Electronic supplementary information (ESI) available: Experimental method and additional results. See DOI: https://doi.org/10.1039/d3gc04432g |
This journal is © The Royal Society of Chemistry 2024 |