Ammonia as a Marine Fuel towards Decarbonization: Emission Control Challenges

Abstract

Decarbonization of the maritime sector to achieve ambitious IMO targets requires the combination of various technologies. Among alternative fuels, ammonia (NH₃), a carbon-free fuel, is a good candidate; however, its combustion produces NO_x, unburnt NH₃ and N₂O—a strong greenhouse gas (GHG). This work conducts a preliminary assessment of the emission control challenges of NH₃ application as fuel in the maritime sector. Commercial catalytic technologies are applied in simulated NH₃ engine exhaust to mitigate NH₃ and NO_x while monitoring N₂O production during the reduction processes. Small-scale experiments on a synthetic gas bench (SGB) with a selective-catalytic reduction (SCR) catalyst and an ammonia oxidation catalyst (AOC) provide reaction kinetics information, which are then integrated into physico-chemical models. The latter are used for the examination of two scenarios concerning the relative engine-out concentrations of NO_x and NH₃ in the exhaust gas: (a) shortage and (b) excess of NH₃. The simulation results indicate that NO_x conversion can be optimized to meet the IMO limits with minimal NH₃ slip in both cases. Excess of NH₃ promotes N₂O formation, particularly at higher NH₃ concentrations. Engine-out N₂O emissions are expected to increase the total N₂O emissions; hence, both sources need to be considered for their successful control.

1. Introduction

Maritime transport, mainly powered by diesel engines, is responsible for almost 3% of global greenhouse gas (GHG) emissions, which is expected to further increase until 2050 [1]. Apart from GHGs, the maritime sector accounts for 24% of nitrogen oxides (NO_x), 24% of sulfur oxides (SO_x) and 9% of particulate matter (PM) emissions in the European Union (EU) [2].

According to the initial IMO strategy, GHG emissions shall be reduced by at least 50% by 2050 and carbon intensity by 40% by 2030 compared to 2008, aiming at complete decarbonization of maritime transport by 2100 [3]. The latest meetings of the IMO Marine Environment Protection Committee (MEPC) recently adopted a revised strategy that aims at net-zero GHG emissions by 2050 [4]. This is a notable acceleration in the emission reduction efforts compared to the initial IMO strategy. In parallel, NO_x emissions shall comply with Tier III limits (3.4 g/kWh for vessel propelled by low-speed two-stroke engines) in Emission Control Areas (ECAs) and Tier II (14.4 g/kWh for vessel propelled by low-speed two-stroke engines) globally. Concerning SO_x emissions, IMO has introduced the global sulfur cap, which imposes an upper limit of 0.50% in the fuel sulfur content globally, dropping to 0.10% in Sulfur Emission Control Areas (SECAs) [5].

Moving towards this direction, several technologies and strategies have already been implemented or are currently being developed, such as [6]:

• direct reduction in fuel consumption, e.g., operating strategies and route optimization as well as slow steaming;
• direct reduction in vessel resistance, e.g., air lubrication, optimized hull design and coating, lightweight materials;
• alternative propulsion system and power sources, e.g., wind-assisted propulsion, fuel cells, cold ironing;
• improvement of energy efficiency, e.g., propulsion system hybridization, waste heat recovery;
• post-combustion gas treatment, e.g., CO₂ capture

Another way to reduce carbon intensity of shipping is the application of alternative fuels with low or zero carbon content, produced using sustainable sources and feedstock and renewable energy (often referred as e-fuels or green fuels). Some alternative fuels (e.g., biofuels) can be used directly on existing engines (drop-in fuels), while others (e.g., ammonia, hydrogen) require significant developments and modifications before becoming the main energy source on board the vessel. Although the combination of various technology packages and practices can reach significant reductions of GHG emissions, complete decarbonization can be achieved only when using carbon-neutral fuels [6,7].

Among other alternative fuels with low or zero carbon content (LNG, LPG, methanol, hydrogen, etc.), ammonia (NH₃) is a promising solution to limit carbon (C) and sulfur (S) emissions from the maritime sector due to several advantages, such as absence of C and S atoms from its molecule, high energy density and relatively easy storage. However, the poor combustion properties of NH₃ (low flammability, high autoignition temperature, low flame speed, etc.) create the need of a pilot fuel quantity, usually carbonaceous, for initiating combustion [8,9,10,11]. Depending on the pilot fuel used, there may be carbon dioxide (CO₂), SO_x and PM emissions, but these levels should be very low (or almost negligible) compared to conventional fuels (particularly heavy fuel oil (HFO)).

Moreover, NH₃ combustion produces three main emission species that have a significant impact on human health and climate: unburned NH₃ that is highly toxic and can cause several health issues when found in high concentrations; NO_x, which is one of the main air pollutants and nitrous oxide (N₂O) [9], a GHG with global warming potential (GWP) almost 300 times higher than that of CO₂, over a 100-year period [12]. Consequently, even small concentrations of N₂O potentially decrease the benefit from CO₂ reduction [13]. Nitrous oxide (N₂O) is a potential byproduct of both NH₃ in-cylinder combustion and chemical reactions in the exhaust gas aftertreatment system. Unburned NH₃ and NO_x emissions can be expected from an engine running on NH₃ and can be reduced using catalytic devices that are commercially available. In marine applications, a vanadium-based selective catalytic reduction (V-SCR) system is commonly used to reduce NO_x with NH₃ or urea as reducing agent [14,15]. Ammonia slip (i.e., unreacted NH₃ downstream of the SCR system) exceeding the need of the deNO_x process can be minimized with an ammonia slip catalyst (ASC). Although that reduction of NH₃ can promote NO_x and N₂O formation [16], N₂O may also be formed in lower concentrations through the SCR reactions [17,18]. Catalytic N₂O reduction technologies are already available but are customized for other applications such as chemical industries and stationary combustion [19,20]. As reduction of a specific species can promote the formation of another, the design of the exhaust after-treatment system (EATS) for an engine running on NH₃ is expected to consist of multiple emission control technologies.

Based on the above, it is clear that commercial technologies must be developed and adopted to effectively reduce NH₃ slip, NO_x and N₂O at the same time in the NH₃ engine exhaust. However, NH₃ engines (particularly large two-stroke ones used in the maritime sector) are not yet commercially available; therefore, the exact exhaust gas conditions in terms of composition, temperature and flow rate needed for the design of an emission control system are not precisely known. Even when the exhaust gas of the NH₃ engine is known from measurements, the design of emission control via trial and error is prohibitive in view of the huge testing costs on a large two-stroke marine engine. It is therefore imperative to develop accurate and predictive models of the aftertreatment system that will be applicable in a wide range of conditions to ensure the coverage of all possible scenarios to be expected in a real NH₃ engine exhaust. The development of such a model is actually the main target of the present work.

The current study presents the development of simulation models of the aftertreatment system of NH₃ application as a fuel in the maritime sector. The primary aim is to determine the viability of current catalytic technologies, identify emission control challenges and ultimately guide the optimum design at an early phase. Exhaust aftertreatment models rely on kinetic mechanisms and rate expressions that describe the intrinsic chemical properties of the active materials. In this work, experiments are performed to derive the respective kinetic information for two technologies of interest and introduce them in an integrated physico-chemical model of the transient transport and reaction processes in monolithic catalytic reactors. The model is then used to study two possible scenarios concerning the proportion of NH₃ and NO_x emissions from NH₃ combustion (engine-out conditions). The first scenario assumes that engine-out NH₃ is less than NO_x (NH₃/NO_x < 1), so additional NH₃ has to be injected upstream of a V-SCR catalyst to achieve NO_x levels below Tier III limits. The second scenario examines the case of excess engine-out NH₃ (NH₃/NO_x > 1) where a dual layer ASC (SCR on top of an ammonia oxidation catalyst (AOC) layer) is integrated to the aftertreatment system to handle the NH₃ slip. Particular emphasis is given to the formation of N₂O through NO_x and NH₃ catalytic reduction.

2. Materials and Methods

2.1. Experimental

Two small-scale samples of commercial catalysts are used in the experimental part of this study: (1) a V-SCR (commonly used in marine applications) with a diameter of 28 mm and a length of 90 mm and (2) a platinum-based AOC with the same dimensions. Their catalytic activity is evaluated with measurements on a synthetic gas bench (SGB), presented in Figure 1. The flow and composition of the mixture is controlled by the programmable mass flow controllers (MFCs). Moisture can be added to the mixture through an H₂O feed, which is heated beforehand to prevent condensation of the flue gas. The mixture is then heated to the required temperature through a pre-heater system before passing through the catalyst sample. The bypass line gives the flexibility to conduct operational and calibration checks of the analyzers, as well as to determine the exhaust gas composition without exposing the catalyst to the gas mixture. An FTIR gas analyzer (AVL Sesam i60 FT SII Small) measures the concentrations of all species in the outlet gas.

In the present work, the SCR and AOC reaction mechanisms are studied by running targeted experimental protocols. For the SCR, steady-state measurements are performed at temperatures between 150 °C and 500 °C and at atmospheric pressure. In the case of AOC, its activity is tested by a temperature ramp (light-off test) from 150 °C to 600 °C under atmospheric pressure. The test conditions for the SCR and AOC testing are summarized in

Table 1. SCR experimental conditions.

Phenomena	Inlet Feed Gas	Temperature [°C]	Space Velocity [h−1]
NO oxidation	2000 ppm NO, 6% O2, 15% H2O, 15 ppm SO2, N2 balance	150 200 250 300 400 500	20,000
NH3 oxidation	1000 ppm NH3, 6% O2, 15% H2O, 15 ppm SO2, N2 balance
Standard SCR	2000 ppm NH3, NH3/NOx = 0.8, 1, 1.5, 6% O2, 15% H2O, 15 ppm SO2, N2 balance
Fast SCR	2000 ppm NH3, 2000 ppm NOx (NO2/NOx = 0.2), 6% O2, 15% H2O, 15 ppm SO2, N2 balance

and

Table 2. AOC experimental conditions.

Phenomena	Inlet Feed Gas	Temperature [°C]	Space Velocity [h−1]
NH3 oxidation	250 ppm NH3, 50 ppm NO, 6% O2, 15% H2O, 15 ppm SO2, N2 balance	150 → 600	20,000

, respectively.

2.2. Modeling

2.2.1. Main Assumptions and Governing Equations

The kinetic mechanisms of the V-SCR and Pt-AOC are implemented into a model of the ExothermiaSuite^® simulation platform [21]. The monolith is simulated as a single representative channel (1D simulation approach), assuming that the inlet flow distribution is uniform and heat losses are negligible. Temperature and species concentrations are computed by solving the quasi-steady state balance equations for heat (Equation (1)) and mass (Equation (2)) transfer:

$${\mathsf{\rho }}_{\mathrm{g}}{\mathrm{C}}_{\mathrm{p},\mathrm{g}}{\mathrm{v}}_{\mathrm{g}}\frac{\partial {\mathrm{T}}_{\mathrm{g}}}{\partial \mathrm{z}}=-\mathrm{h}\times \left(\frac{{\mathrm{S}}_{\mathrm{F}}}{\mathsf{\epsilon }}\right)\times \left({\mathrm{T}}_{\mathrm{g}}-{\mathrm{T}}_{\mathrm{s}}\right)$$

(1)

$$\frac{\partial \left({\mathrm{v}}_{\mathrm{g}}{\mathrm{y}}_{\mathrm{g},\mathrm{j}}\right)}{\partial \mathrm{z}}={-\mathrm{k}}_{\mathrm{j}}\times \left(\frac{{\mathrm{S}}_{\mathrm{F}}}{\mathsf{\epsilon }}\right)\times \left({\mathrm{y}}_{\mathrm{g},\mathrm{j}}-{\mathrm{y}}_{\mathrm{s},\mathrm{j}}\right)$$

(2)

The wall surface temperature is calculated using the transient energy balance in the solid phase (Equation (3)):

$${\mathsf{\rho }}_{\mathrm{s}}{\mathrm{C}}_{\mathrm{p},\mathrm{s}}\frac{\partial {\mathrm{T}}_{\mathrm{s}}}{\partial \mathrm{t}}={\mathsf{\lambda }}_{\mathrm{s},\mathrm{z}}\frac{{\partial }^2{\mathrm{T}}_{\mathrm{s}}}{{\partial \mathrm{z}}^2}+\mathrm{S}$$

(3)

The surface concentrations are obtained by solving the concentration field inside the washcoat layer (Equation (4)):

$$-{\mathrm{D}}_{\mathrm{w},\mathrm{j}}\frac{{\partial }^2{\mathrm{y}}_{\mathrm{s},\mathrm{j}}}{\partial {\mathrm{w}}^2}=\sum _{\mathrm{k}}{\mathrm{n}}_{\mathrm{j},\mathrm{k}}{\mathrm{R}}_{\mathrm{k}}$$

(4)

The convective mass transfer from the gas to the washcoat surface is formulated as

$$\frac{\partial \left({\mathrm{v}}_{\mathrm{g}}{\mathrm{y}}_{\mathrm{g},\mathrm{j}}\right)}{\partial \mathrm{z}}={\mathrm{k}}_{\mathrm{j}}\left(\frac{{\mathrm{S}}_{\mathrm{F}}}{\mathsf{\epsilon }}\right)\left({\left.{\mathrm{y}}_{\mathrm{s},\mathrm{j}}\right|}_{\mathrm{w}=-{\mathrm{w}}_{\mathrm{c}}}-{\mathrm{y}}_{\mathrm{g},\mathrm{j}}\right)$$

(5)

Supplementary equations regarding the model of the flow through catalyst are provided in Appendix A.

The solution of the concentration field in the washcoat layer is of particular importance for the case of technologies with multiple catalytic layers (1D + 1D model). In fact, this is the case with ASCs that usually contain both a precious metal (PGM) layer, particularly an AOC layer for the oxidation of NH₃, as well as an SCR layer on top (Figure 2). This combination comes with advantages concerning NH₃ reduction and selectivity properties of the ASC, as NO_x formed in the oxidation layer diffuses through the SCR layer where it can be reduced [14].

2.2.2. Reaction Mechanisms

In order to examine the potential of the existing catalytic devices to treat NH₃ combustion products, thoroughly calibrated and validated SCR and AOC models are necessary. To describe the SCR reactivity over the vanadium-based catalyst, commonly used SCR reactions are adopted [16,22], as listed in

Table 3. SCR reaction scheme.

Type	Reaction
NH3 storage/release	NH3 ↔ NH3 *
Standard SCR	4 NH3 * + 4 NO + O2 → 4 N2 + 6 H2O
Fast SCR	4 NH3 * + 2 NO + 2 NO2 → 4 N2 + 6 H2O
NO2 SCR	NH3 * + 3/4 NO2 → 7/8 N2 + 3/2 H2O
N2O formation	2 NH3 * + 2 NO + O2 → N2 + N2O + 3 H2O 4 NH3 * + 4 NO2 → 2 Ν2 + 2 N2O + 6 H2O
NO oxidation	NO + 1/2 O2 ↔ NO2
NH3 oxidation	4 NH3 * + 5 O2 → 4 NO + 5 H2O 2 NH3 * + 3/2 O2 → N2 + 3 H2O 4 NH3 * + 4 O2 → 2 N2O +6 H2O

* Stored NH₃ on the catalyst sites.

. The standard, fast and NO₂ SCR reactions are considered the principal reactions between NO_x and NH₃ (depending on the proportion of NO₂/NO_x). While NH₃ is primarily oxidized to N₂, oxidation reactions to NO and N₂O are also considered. The formation of N₂O has also been attributed to the oxidation of NH₃ and NO [23], as well as to the direct reaction between NH₃ and NO₂ [24,25].

Ammonia oxidation on the platinum-based AOC is approached with a simple kinetic model that can give good representation of the overall reactions. The oxidation reactions used are listed in

Table 4. AOC reaction scheme.

Type	Reaction
NO oxidation	NO + 1/2 O2 ↔ NO2
NH3 oxidation	4 NH3 + 5 O2 → 4 NO + 5 H2O 2 NH3 + 3/2 O2 → N2 + 3 H2O
NH3 and NO oxidation to N2O	2 NH3 + 2 NO + 3/2 O2 → 2 N2O + 3 H2O

. These include the oxidation of NH₃ to N₂ and NO, the simultaneous oxidation of NH₃ and NO to N₂O and the oxidation of NO to NO₂ (including the reverse reaction of NO₂ decomposition) [26,27].

2.3. Full-Scale Application of the Model

Assumptions and Inlet and Boundary Conditions

The target of this section is to demonstrate the use of modeling in the design phase of the NH₃ marine engine exhaust aftertreatment system. In the early design phase, many parameters that influence the catalyst selection and optimization are not known, including NO_x, NH₃ and N₂O engine-out emission levels. Here, it is assumed that the engine will use small amounts of pilot fuel [9,28,29]; therefore, CO₂ emissions can be neglected, at least at the preliminary design of the EATS. Although engine-out N₂O produced by NH₃ combustion is a topic of high concern [28,30], for the purposes of the present work it will also be considered negligible; nevertheless, the N₂O that is potentially produced in the EATS as an unwanted byproduct will be examined, and its greenhouse effect potential will be evaluated.

Ammonia combustion is likely to produce high levels of unburnt NH₃ [29,30], resulting in concentrations comparable to the respective NO_x emissions in terms of mole fraction. It is well known that the molar ratio of NH₃/NO_x in the exhaust gas is very critical for the operation of the SCR. Ratios below 1 would probably necessitate extra NH₃ in the exhaust gas stream, eventually via an additional NH₃ injection system. On the other hand, if the ratio is above 1, then the excess NH₃ escaping the SCR reactions will have to be treated by dedicated catalysis.

In this preliminary concept study, both cases described above will be studied. In the first case, NH₃ injection upstream of a V-SCR is required as shown in Figure 3a. In the second case, a dual layer/dual function ASC is placed downstream from the SCR (Figure 3b) to treat the unreacted NH₃ of the deNO_x process. The ASC is assumed to be a combination of V-based (SCR) and precious-metal based catalytic layers (AOC) (as shown indicatively in Figure 2) in order to attain a desired NH₃ oxidation selectivity to N₂.

Since NH₃ engines are currently under development and their real exhaust gas conditions and emission concentrations are still not known precisely, the current study is based on real-world engine-out conditions of low-speed diesel engines used in marine applications, assuming a high-pressure (pre-turbo) SCR system [31,32]. The simulated catalyst inlet conditions are summarized in

Table 5. Exhaust gas conditions assumed in this study.

Engine Load	%	100	75	50	25
Exhaust gas temperature	°C	410	350	310	290
Exhaust gas pressure	bar	4.0	3.1	2.1	1.4
SCR space velocity	h−1	40,000	32,000	25,000	10,000
ASC space velocity	h−1	140,000	115,000	85,000	40,000
NOx concentration	ppm	1500–2000	1500–2000	1500–2000	1500–2000

. Pre-turbo SCR configurations have the advantage of higher pressures and temperatures that prevail right after the engine and expand the active range of SCR operation, especially in low loads [33,34]. It is worth noting that the physico-chemical model can be applied in the entire operating envelope and is sensitive to the effect of pressure on reaction rates and species diffusivity.

3. Results

3.1. Reaction Model Calibration

The reaction kinetic parameters of the two catalysts are calibrated to fit the experimentally determined NO_x, NH₃ and N₂O concentrations. The results of the NO and NH₃ oxidation tests for the V-SCR catalyst (see Table 1) are presented in Figure 4 with markers. Oxidation of NO to NO₂ is hardly detected even at high temperatures (Figure 4a). NH₃ is mainly oxidized to N₂ above 300 °C and is almost fully oxidized at 500 °C (Figure 4b), while NO and N₂O formation is observed only at very high temperatures (500 °C). The same figures contain the results of the simulation model after fitting of the reaction kinetic rate parameters. The model achieves a good agreement with the test results in the whole temperature range and is able to predict the reaction selectivity towards NO and N₂O.

The results of the SCR activity tests presented in Figure 5 show that the catalyst exceeds 80% NO_x conversion above 300 °C. Obviously, the SCR process is highly dependent on the amount of NH₃ in the feed gas (Figure 5a). When the NH₃/NO_x ratio is greater than 1, NO_x is almost fully converted at high temperatures, although this leads to unreacted ammonia. When the NH₃/NO_x ratio is less than 1, only partial NO_x conversion is achieved as expected from the reaction stoichiometry (Standard SCR reaction (Table 3)). Addition of NO₂ in the feed gas (Figure 5b) enhances NO_x conversion rates, especially at low temperatures. Low selectivity to N₂O (below 20 ppm) is observed in all conditions with a significant increase of up to 120 ppm at 500 °C.

The results of the calibrated simulation model presented in the above figures with lines clearly show a good agreement with the respective measured data in the whole range of temperature, NH₃/NO_x ratio and NO₂/NO_x ratio conditions.

Figure 6 shows the axial profiles of NO_x, NH₃ and N₂O along the V-SCR catalyst as well as the reaction rates governing N₂O formation pathways. One can observe that N₂O attains stabilization approximately midway through the catalyst length, whereas NO_x and NH₃ concentrations reach a state of equilibrium near the catalyst outlet. This is attributed to the constrained availability of NH₃ and NO_x in the feed gas; hence, the reaction rates that favor N₂O formation are minimized midway through the catalyst.

The results of the Pt-based AOC tests are summarized in Figure 7. Here, the focus is not only on the conversion rate of NH₃ as a function of temperature, but also on the unwanted NO_x and N₂O produced by the NH₃ oxidation reactions. It is worth noting that the calibrated model is capable of capturing these complex trends with respect to NO_x byproducts in the whole temperature range with good accuracy. This provides the basis for using this physico-chemical model in the conditions expected in a real marine engine.

The concentration of NH₃ shows a steep decrease from 200 °C to 250 °C and is fully oxidized around 300 °C. Above 200 °C, N₂O selectivity increases significantly with maximum concentration at 250 °C. Selectivity to NO and NO₂ is favored at temperatures above 250 °C, while N₂O selectivity is simultaneously decreasing. It is important to highlight the temperature range of N₂O formation (in the aftertreatment system) between 200 °C and 400 °C, which is crucial for low-speed marine engines since their exhaust gas temperature falls within this range (see Table 5). The trends can be interpreted by referring to the reaction rates depicted in Figure 8, highlighting the competition between the reactions. Between 200 °C and 400 °C, the simultaneous oxidation of NH₃ and NO to N₂O is favored, while above 250 °C the oxidation of NH₃ to NO becomes dominant; hence, the availability of NH₃ towards N₂O is limited.

Figure 9 presents the axial distribution of NH₃, NO and N₂O along the AOC. At notably low temperatures (i.e., 150 °C), oxidation reactions governing the AOC are not activated; therefore, no alteration in emission levels is observed. Conversely, at elevated temperatures (i.e., 350 °C and 500 °C), NH₃ experiences complete oxidation in close proximity to the catalyst inlet, precluding its availability for subsequent oxidation pathways towards NO and N₂O. Consequently, NO and N₂O levels exhibit an early stabilization along the catalyst length.

3.2. Model Application in Marine Engine Exhaust

Based on the assumed marine engine exhaust gas conditions of Table 5 and the weighting factors of the legislated E3 test cycle [35], it can be estimated that the NO_x conversion efficiency required to reduce large two-stroke engine-out emissions below the Tier III limit of 3.4 gNO_x/kWh is in the order of 90%.

In the case of lack of NH₃ (engine-out NH₃/NO_x < 1), where only the SCR catalyst is used (Figure 3a), the minimum deNO_x requirement may be achieved provided that NH₃ is injected with a target ratio of NH₃/NO_x equal to 0.9.

Applying the simulation model at the four loads of the E3 cycle, using the conditions shown in Table 5, NH₃ slip and N₂O formation after the SCR are calculated, as presented in Figure 10a. Almost all NH₃ is predicted to be consumed during NO_x reduction, leading to limited NH₃ slip of less than 5 ppm. Low levels of N₂O (below 8 ppm) are expected to be formed at all loads with increased selectivity at full load as N₂O formation is favored at elevated temperatures. Despite its low selectivity, N₂O is a strong GHG with 100-year GWP almost 300 times higher than CO₂. Hence, even small concentrations of N₂O can be equivalent to significant CO₂ emissions. In this case, the CO₂-equivalent emissions over a 100-year period reach almost 25 g/kWh at 100% load, which are decreased at lower loads (Figure 10b), resulting in an average value of 14.1 gCO₂-eq/kWh (taking into account the weighting factors of the E3 test cycle [28]).

Figure 11 presents the average NH₃, NO_x, N₂O and CO₂-equivalent emissions for different NH₃/NO_x ratios in the case of excess engine-out NH₃ (NH₃/NO_x > 1). All concentrations are estimated taking into account the weighting factors of each load according to the E3 test cycle [35]. According to the standard SCR reaction, NH₃ and NO react on a 1:1 molar ratio. Thus, increased NH₃/NO_x values lead to elevated unreacted NH₃ at the SCR outlet. Unreacted NH₃ of the deNO_x process is then oxidized in the ASC (Figure 11a). The activity of the ASC is decreased at higher NH₃/NO_x ratios, leading to increased NH₃ emissions at the outlet. Despite the strong NH₃ oxidation, the ASC is characterized by high selectivity to NO_x and N₂O that becomes more important when unreacted NH₃ in the SCR is higher. NO_x conversion is maximized in the SCR due to the abundant concentrations of NH₃, while the SCR layer of the ASC catalyst counterbalances the high selectivity of NH₃ oxidation to NO, keeping NO_x concentrations at acceptable levels, compliant with Tier III limits (<3.4 g/kWh) (Figure 11b). Concerning N₂O, limited formation is observed during SCR (minimally affected by NH₃/NO_x); however, the simultaneous oxidation of NH₃ and NO in the ASC results in important N₂O formation, especially at higher NH₃ engine-out concentrations (Figure 11c). According to Figure 11d, N₂O produced in the AOC layer corresponds to significant levels of CO₂-equivalent emissions that reach almost 500 g/kWh at high NH₃ concentrations. These levels are comparable to other low-carbon solutions, such as LNG combustion, where CO₂-equivalent emissions for low-speed two-stroke engines vary between 400 g/kWh (high-pressure dual-fuel mode) and 500 g/kWh (low-pressure dual-fuel mode) [36].

In addition, N₂O emissions from NH₃ in-cylinder combustion are expected to further increase the total GHG emissions. Therefore, both sources need to be considered for the successful control of N₂O emissions, eventually via a targeted additional catalyst. A different SCR layer composition could be beneficial for N₂O abatement. For example, iron-based catalysts might be a better option compared to the V-based catalyst considered here as they have the ability of simultaneous reduction of NO_x and N₂O [18]. Another possible solution is the direct reduction of N₂O through thermal decomposition [17,18,37].

4. Summary and Conclusions

The testing and simulation results of the marine engine aftertreatment models highlighted the following:

• In the case where engine-out NH₃ levels are lower than the ones required in the deNO_x process (i.e., NH₃ injection upstream of the SCR), NO_x conversion can be optimized to comply with the strictest IMO limits with minimal levels of NH₃ slip and N₂O formation.
• In the case where NH₃/NO_x is greater than 1, unreacted NH₃ of the deNO_x process can be efficiently handled with an ASC, while NO_x concentrations can be kept at acceptable levels. Concerning N₂O, NH₃ oxidation in the ASC is highly selective to N₂O formation, which is enhanced at higher NH₃ concentrations. In this case, the CO₂-equivalent emissions over a 100-year period are comparable to LNG marine engines.

Considering these indications, it is preferable to tune NH₃ combustion to ensure that NH₃/NO_x is less than 1, so as to minimize unreacted ammonia in the aftertreatment system and thus keep N₂O formed there at low levels. Except from the part produced in the catalytic aftertreatment devices, N₂O levels in the exhaust gas are expected to further increase when engine-out quantity from NH₃ combustion is considered. Therefore, the potential CO₂ benefit of NH₃ combustion may be counterbalanced, to a certain extent, due to the strong GWP of N₂O. Hence the use of NH₃ as a fuel to decarbonize the maritime sector will be beneficial only if these levels can be kept at low levels. Based on the above, an appropriate control strategy and optimization of the exhaust aftertreatment system of the NH₃ engine are of high importance as NO_x reduction should be accompanied by limited NH₃ slip and N₂O formation. For this reason, the activities of this work are further expanded with future steps including the following:

• Integration in the catalyst model of N₂O chemistry and the relevant catalytic processes in a dedicated deN₂O catalyst.
• Experimental small-scale investigation of the performance of new catalyst technologies, followed by calibration and validation of the model using the test data.
• Application of the new catalyst models in the exhaust gas stream of NH₃ engines.
• Development and optimization of the complete exhaust aftertreatment system and controls for NH₃ marine engine applications.