Ammonia combustion does not produce carbon dioxide, and has been recognized as one of the promising approaches to achieving carbon neutrality in transportation sector. Due to the slow flame propagation speed and high ignition energy requirements, reactive fuel pilot ignition is essential for ammonia combustion in compression ignition engines. The reduction of pilot fuel drives CO2 mitigation by curtailing carbon input, but this has a demand of advanced combustion modulation techniques to sustain engine efficiency. Designed pilot fuel stratification enables an activated in-cylinder environment, overcoming the difficulty in ammonia ignition and combustion, and allowing for a minimal pilot fuel amount to trigger premixed ammonia combustion. The minimum pilot fuel permits 99.1% ammonia energy substitution, accounting for only 1.3% of the CO2 emissions from diesel combustion at the same load condition. Optimized intake organization coupled with improved reactivity stratification also achieves over 46% brake thermal efficiency, and reduces unburned ammonia by over 80% compared to baseline operation.
The growing international concern regarding carbon emissions reduction in the energy sector prompts an imminent need for the transformation of traditional internal combustion engines (ICEs) [1]. Ammonia, as a zero-carbon fuel, produces no CO2 during the combustion process, thus being a promising approach to the decarbonization of ICEs [2]. The combustion organization and the emissions control of ammonia-fueled engines have attracted widespread attention.
Compared to hydrogen, ammonia exhibits a higher volumetric energy density in the liquid state, which is advantageous for storage and transport. Ammonia production can utilize various pathways, one of which is green power generation, allowing for zero-carbon emissions across the entire life cycle [[3], [4], [5], [6]]. However, ammonia combustion is characterized by high ignition energy and slow combustion rate, which leads to the delayed combustion phase and the increased unburned ammonia emissions [7,8]. The auto compression-ignition (CI) of pure ammonia typically requires an ultra-high compression ratio over 35:1 [9,10], which is difficult to realize on engines. To circumvent the challenges associated with engine structure modifications, several approaches, such as the application of high-reactivity fuels and the implementation of novel ignition modes, are employed to enhance ammonia combustion [11,12].
The common method to improve the ignition performance of ammonia is to blend it with high-reactivity fuels for combustion. Hydrogen, as a representative high-reactivity fuel, has been recognized as a potential combustion promoter for ammonia [[13], [14], [15], [16]]. Wang et al. [17] investigated the ignition process of the ammonia-hydrogen mixture under high-pressure direct-injection engine conditions. The co-combustion of ammonia-hydrogen is effective to enhance combustion, yet challenges such as high flame temperature, high NOx emissions, and narrow combustion limits still need to be addressed [[18], [19], [20]]. Other high-reactivity fuels, such as methane [[21], [22], [23]] and gasoline [[24], [25], [26]], are also commonly employed to improve the combustion of ammonia, serving as the ignition fuel. The application of ignition fuels has been validated as one of the effective approaches in spark-ignition (SI) ammonia engines [27]. Meanwhile, the challenges of emissions control and engine stability maintenance further promote the development of novel combustion modes, such as spark-assisted compression ignition [28,29] and reactivity-controlled turbulent jet ignition [26,30,31].
While in CI engines, the absence of an ignition source makes direct compression ignition of ammonia challenging. Typically, high cetane number fuels are employed as pilot ignition fuels, such as diesel [[32], [33], [34]]. In the ammonia-diesel dual-fuel combustion mode, achieving both high ammonia energy content and high efficiency is often challenging [35]. Yousefi et al. [36,37] found that the optimization of multi-stage injection can enhance engine efficiency. Furthermore, Niki [38] found that the early diesel injection promotes homogeneous mixture formation, resulting in the optimized combustion process. Meanwhile, the reactivity-controlled compression ignition aids in ammonia combustion. Unfortunately, the ammonia energy fractions (AEFs) in most of the existing works were limited, diminishing the carbon reduction potential.
Several experimental studies demonstrated that increased AEF adversely impacts combustion stability and emission profiles [[39], [40], [41]]. A significantly reduced peak flame area and decreased flame natural lightness were found by Sun et al. [42] when AEF came to 80%, which indicated a deterioration in-cylinder combustion. As a result, the large amount of unburned ammonia emissions and the drastically reduced efficiency can be observed as the AEF increased. In the face of these challenges, Pei et al. [43] found that the extended mixing time of ammonia and diesel could promote the formation of the highly reactive mixture, resulting the improved combustion efficiency when the premixed charge compression ignition mode is employed. The importance of the mixture reactivity distribution is equally argued by other researchers [44,45], where AEF was usually maintained at approximately 70%. Meanwhile, some scholars have focused on the effects of liquid ammonia direct injection on the transition of in-cylinder combustion [46,47], in which the direct ammonia injection improves the fuel blending mode, accelerates the combustion, and enhances the efficiency. Optimizing in-cylinder combustion modes in ammonia engines using diesel pilot ignition can be an important approach to achieving high AEF with acceptable efficiency and emissions levels. However, research on the comparison of different combustion modes in engines with high AEF is relatively limited. The substitution rate boundary for diesel-ignited ammonia engines is still unclear.
This study systematically investigated the operational performance of an ammonia-fueled engine with diesel pilot ignition under high-load conditions. First, the impact of intake air mass on combustion and emissions was examined at AEF exceeding 80%. Subsequently, the underlying mechanism was elucidated by analyzing the transition of in-cylinder combustion modes driven by changes in reactivity stratification. Building on these insights, the clean and efficient combustion strategy optimal for high-AEF operation was identified, ultimately achieving a new upper limit for AEF.
2. Methodology
2.1. Experimental methods
A four-stroke compression ignition engine with a displacement of 12.8 L was modified to meet the experimental requirements of an ammonia engine, the detailed parameters of which are exhibited in Table 1. Meanwhile, Fig. 1 illustrates the schematic diagram of the experimental setup components. Ammonia is injected into the intake manifold. To maintain a stable supply pressure of ammonia, multiple-stage heat exchangers are employed. Subsequently, a precise flow rate of ammonia injection is achieved through electronically controlled gas rail nozzles, with the injection pressure controlled at (6 ± 0.5) bar (1 bar = 105 Pa). The control systems for ammonia and diesel injection are independent of each other. Diesel is injected into the cylinder through a high-pressure common rail system, allowing for multi-stage direct injection. Various parameters of the injection, such as injection timing, events, and intervals, can be adjusted in real-time through the controller. The dynamometer (DYBAS3 HD 460 STD, Horiba) controls the engine operation condition to the target power.
To accurately measure the emission parameters of the ammonia engine, multiple detection devices are employed to measure the concentration of substances in the exhaust gas. Major gaseous emissions (CO, CO2, and THC, THC refers to incompletely burned or unburned hydrocarbons) were measured by a gas analyzer (MEX-ONE-D1, Horiba). Furthermore, given the consistently high ammonia substitution rates tested, unburned ammonia emissions usually remain elevated. In the experiment, an exhaust gas dilution is employed, and the concentrations of NH3, NOx, and N2O in the exhaust are accurately measured through a Fourier-transform infrared spectroscopy (FTIR, BOB-2000FT-G, A&D). The test facilities and their measurement uncertainty are presented in Table 2.
2.2. Data analysis and experimental conditions
The AEF can be calculated through the percentage of the calorific value of ammonia in the overall fuels, which is exhibited in Eq. (1), where ${{m}_{\text{N}{{\text{H}}_{3}}}}$ and ${{m}_{\text{Diesel}}}$ are the mass flow rate of ammonia and diesel (kg·h−1), respectively; $\text{LH}{{\text{V}}_{\text{N}{{\text{H}}_{3}}}}$ and $\text{LH}{{\text{V}}_{\text{Diesel}}}$ are the lower heating value (LHV) of ammonia and diesel (kJ·kg−1), respectively.
The excess air coefficient λ can intuitively reflect the proportion of fuel and oxidizer in the combustion chamber, which can be calculated by Eq. (2), where mair is the mass flow rate of intake air (kg·h−1); $\text{A}{{\text{F}}_{\text{st},\text{N}{{\text{H}}_{3}}}}$ and $\text{A}{{\text{F}}_{\text{st},\text{Diesel}}}$ are the stoichiometric air–fuel ratios of NH3 and diesel, respectively.
The brake thermal efficiency (BTE) typically represents the efficiency of fuel chemical energy conversion to work in the engine. The BTE can be calculated by Eq. (3), where Pb is the brake power (kW).
Heat release rate (HRR) is derived based on the first law of thermodynamics. Several key combustion parameters could then be calculated from the cumulative in-cylinder heat release. CA10, CA50, and CA90 represent the crank angles corresponding to the accumulated heat release reaching 10%, 50%, and 90% of the total heat release, respectively.
During the experiment, the AEF was maintained to be no less than 80% to investigate the effects of different intake parameters and different combustion modes on combustion and emissions. The engine was typically operated at the brake mean effective pressure (BMEP) of 1.87 MPa. Since the advantages of multiple injections of diesel have been pointed out in several studies [38,48,49]. The diesel injection was first fixed to a two-stage injection before top dead center (BTDC) when investigating the impact of the λ. The pre-injected 5 mg of diesel entered the cylinder for combustion at 20 °CA BTDC, followed by the main-injected diesel at 10 °CA BTDC. The intake air mass and temperature were controlled by adjusting the operating conditions of the intercooler and throttle. While investigating the transition of in-cylinder combustion modes, emphasis is placed on examining the effect of diesel injection strategy on the reactivity stratification in the cylinder. Therefore, different masses and phases of diesel injection were considered, in combination with the investigation of several combustion modes. The detailed experiment conditions are exhibited in Table 3.
3. Results and discussion
3.1. The impact of excess air coefficient on combustion and emissions of ammonia engines
This section focuses on the effects of intake air mass on the combustion and emissions of the diesel-ignited ammonia engine. As a result, the two-stage diesel injection is maintained to operate the engine, as exhibited in Table 3. To compare the effect of intake air conditions in the cases of ammonia-dominated combustion, the AEF was maintained at 80% and 90% in the experiments, respectively.
Fig. 2 illustrates the in-cylinder pressure and HRR of the ammonia engine under different intake air masses. In all the conditions, there is a certain pressure increase during the pure compression phase with the increase in the intake air mass, where the diesel has not been injected yet. The increase in compression pressure contributes to the compression ignition of diesel, which leads to an advanced HRR initiation phase under high λ conditions. As shown in Fig. 2(a), the shorter ignition delay under the same injection conditions can be observed at the higher λ, which is represented by the start phase of the HRR curves. Since the second-stage diesel injection occurs at 10 °CA BTDC, combustion of the pre-injected diesel has already been initiated before this timing. The high λ enhances the diesel ignition, resulting in combustion phase separation between ammonia and diesel. This phenomenon is most pronounced at λ = 1.59, exhibiting the maximum phase interval between two distinct HRR peaks. Under this condition, the first heat release peak predominantly originates from diesel combustion and ammonia ignition near the injection, while the subsequent peak arises from the remaining ammonia combustion. Experimental observations reveal that decreasing λ delays the heat release initiation phase, whereas it advances the peak phase and increases peak magnitude. This results in more concentrated in-cylinder combustion processes. The changed λ affects the diesel ignition delay, which in turn causes changes in the combustion phase of ammonia and diesel.
At 90% AEF, pressure sensitivity to the λ is similar to that observed at 80% AEF. Increased intake air mass shortens the ignition delay of pre-injected diesel and reduces the secondary heat release peak intensity. Notably, the heat release peak phase ceases to exhibit monotonic advancement with decreasing λ, as evidenced by the curves marked in the red area. Consistent with previous analysis, the dominant heat release peak is governed by the ammonia combustion. Consequently, it would be difficult for diesel with a 10% energy contribution to reliably initiate combustion of the ammonia-dominant mixture. The inherently slower flame propagation of ammonia combustion under these conditions directly leads to the heat release peak retardation.
Fig. 3 exhibits the evolution of CA10, CA50, and CA90 at varied intake air conditions in high-AEF combustion regimes. Increased λ consistently advances the CA10 phase at both 80% and 90% AEF conditions, demonstrating a λ-dependent behavior. This advancement can be attributed to enhanced pre-injected diesel ignition under elevated λ conditions, as previously described. At 80% AEF, the delayed CA50 phase under high λ conditions is mainly caused by the separated combustion, manifested as distinct phase differentiation between two heat release peaks. This phenomenon is no longer pronounced when the AEF is further increased to 90%, as exhibited in Fig. 2(b). When λ decreases to approximately 1.0, suppressed diesel combustion hardly ignite ammonia, and as such the flame propagation of ammonia dominates the combustion process. Consequently, the CA50 phase initially advances through the concentrated combustion but subsequently delays due to the slow ammonia combustion when λ decreases. The CA50-CA90 interval, an important metric for late-stage combustion evaluation, demonstrates a gradual increase with decreasing λ at 80% AEF. This may be because the late-stage combustion is dominated by ammonia. At elevated AEF = 90%, diesel combustion impact becomes localized to ignition initiation and early combustion phases. Therefore, λ predominantly governs ignition delay characteristics and early combustion development at high-AEF diesel-pilot ammonia conditions. An elevated λ could intensify the ammonia-diesel combustion separation, but accelerate pre-injected diesel heat release.
Engine efficiency variations are governed by the effect of λ on the combustion process. The analysis reveals that elevated λ delays CA50 phase despite promoting diesel ignition. The sweet point of BTE emerges within the λ range of 1.3-1.4, which is close to 44.5%. Deviations from the optimal λ (both elevated and reduced) decrease the thermal efficiency, as the delayed combustion phase due to reduced λ exerts more pronounced efficiency degradation. The engine efficiency drops by over 2% from its sweet point as λ approaches unity at AEF = 90%.
Ammonia-fueled engines exhibit higher NOx and unburned ammonia emissions compared to conventional diesel engines. Fig. 4(a) quantifies an approximate 35% reduction in NH3 when λ decrease from around 1.6 to around 1.1 at both 80% and 90% AEF conditions. As discussed above, reduced λ delays the CA10 phase by about 3 °CA, inducing concentration combustion that enhances ammonia oxidation completeness. Oh et al. [50] similarly found that a low λ contributes to the reduction of unburned ammonia emissions in an ammonia-natural gas dual-fuel engine. This suggests that the control strategy of λ is one of the important ways to reduce ammonia emissions. The elevated burning velocity of ammonia at low λ also plays a role in enhancing combustion completeness [51]. Although minor deviations are observed at certain operation points, such as the marginally lower NH3 when AEF = 90% and λ = 1.45, the ammonia emissions under high AEF conditions follow a similar trend. When it comes to NOx emissions, prior studies confirm that ammonia combustion dominates NOx production, differing from the thermal NOx generation mechanisms in conventional diesel engines [48]. From Fig. 4(a), the NOx emissions at AEF = 90% are always higher than those at AEF = 80%. The difference in NOx emissions under the two AEF conditions at similar λ primarily stems from the varied extent of ammonia involvement in combustion, as evidenced by comparable levels of unburned ammonia emissions. Consequently, NOx emissions exhibit a positive correlation with the quantity of ammonia consumed in combustion, further indicating that NOx formation is predominantly governed by the combustion rather than the thermal mechanism.
Meanwhile, the CO and THC emissions can characterize the diesel combustion, as exhibited in Fig. 4(b). THC refers to incompletely burned or unburned hydrocarbons, which change little as intake air mass decreases, especially at 80% AEF. This indicates that the pyrolysis process of diesel is not sensitive to the λ in this condition. Unburned hydrocarbons slightly increase as λ decreases when AEF = 90% due to that the high concentrations of ammonia, and reduced oxygen levels limit further reaction of hydrocarbons. In contrast, CO emissions demonstrate a distinct nonlinear response, maintaining trace concentrations (< 50 parts per million (ppm)) under high λ conditions, but exhibiting rapid growth as λ decreases below 1.3. As the energy share of diesel increases, the critical λ for the abrupt rise in CO emissions also increases. As mentioned earlier, the simultaneous combustion of diesel and ammonia during this phase creates a competitive interaction between the two fuels for the available oxidant. A reduction in the λ intensifies this competitive interaction, resulting in the direct release of substantial quantities of incomplete combustion products. Meanwhile, this phenomenon stems from the fuel distribution in the cylinder, in which ammonia disperses homogeneously while diesel is directly injected. Experimental results demonstrate that maintaining a combustion regime (1.3 < λ < 1.4) achieves optimal NH3 emissions mitigation. When λ is less than 1.2, the incomplete diesel combustion represented by high CO emissions reduces engine efficiency.
Due to the strong greenhouse effect, N2O emissions are equally noteworthy, which constitute the main greenhouse gas (GHG) emissions together with CO2 [[52], [53], [54]]. As exhibited in Fig. 5, it is obvious that the CO2 emissions are correlated with the diesel energy share in the fuel, as diesel constitutes the exclusive carbon source. The CO2 concentration increases slightly with decreased intake air mass, which is mainly caused by the corresponding reduced exhaust volume as λ decreases. Meanwhile, the N2O emissions decrease as λ decreases, paralleling the trend observed in unburned ammonia emissions across the tested λ range. This signifies that the optimized λ facilitates GHG emission reductions. The N2O can be consumed through reactions to form N2 [55]. Meanwhile, it can also produce N2 by thermal dissociation. When the λ decreases, the dehydrogenation of ammonia becomes difficult, and reactions NO + HO2 = NO2 + OH and NH2 + NO2 = N2O + H2O are suppressed, resulting in reduced N2O emissions [56]. Furthermore, when AEF is 90%, the N2O emissions are always less than those of 80% AEF. In summary, the high AEF contribute to the direct reduction of carbon content in the fuel, which maintains CO2 emissions at a consistently low level compared with conventional diesel engines. Although N2O constitutes a minor fraction of total emissions (totally < 100 ppm), its strong global warming potential (GWP-100 = 298 CO2-equivalent) substantially offsets the climate benefits achieved through CO2 mitigation. While intake air optimization achieves > 35% N2O reduction, the development of advanced aftertreatment systems represents another viable solution for comprehensive emission control, achieving simultaneous reductions in N2O, NH3, and NOx [57].
3.2. The impact of injection strategies on combustion transition
This section discusses the effect of different diesel injection strategies on the in-cylinder combustion pattern. As a result of the above analyses, appropriate intake air conditions (λ = 1.3 ± 0.1; intake air temperature = 50 °C) were chosen as boundaries for comparing the combustion modes. Meanwhile, AEF was maintained at 80% to ensure stable engine operation when the injection pattern changed dramatically. In this study, a novel pilot-ignition reactivity-stratification (PIRS) combustion mode is employed, in which a large percentage of diesel is firstly injected into the cylinder substantially earlier (at 60 °CA BTDC) to mix fully with ammonia and intake air. The remaining amount of diesel is injected directly for compression ignition near the top dead center, providing the ignition energy for the highly active mixtures. Meanwhile, the single diesel direct injection compression ignition (DICI) mode was tested as a comparison. The test conditions are exhibited in Table 3, where the single injection timing of diesel was increased 6-16 °CA BTDC in the DICI mode, while a certain amount of diesel (20, 25, and 30 mg) is injected at 60 °CA BTDC, with the second diesel injection occurring at 10-14 °CA BTDC in the PIRS mode.
Fig. 6 exhibits the pressure and HRR at different diesel injection timings in the two combustion modes. When diesel is injected in a single stage, the rapidly rising heat release indicates compression ignition of the diesel after a certain delay, developing the first heat release peak, as shown in Fig. 6(a). The pre-mixed ammonia is then ignited, the combustion of which results in the second heat release peak. The earlier injection timing advances the combustion phase, which in turn leads to a rapid increase in the maximum pressure. Under these conditions, ammonia combustion phase is governed by the diesel ignition event, whose HRR peak increases with advanced injection timing. The enhanced heat release peak primarily results from the thermodynamic conditions near the top dead center, the pressure and temperature of which are more favorable for ammonia combustion. Consequently, ammonia combustion can be reliably initiated through a single diesel pilot injection, with the combustion phase being regulated through diesel injection timing modulation.
In the PIRS mode, the pre-injected diesel leads to a unique HRR compared with that of the DICI mode. The early-stage injection of 25 mg diesel (about 70% of the total diesel mass per cycle) at 60 °CA BTDC enables complete mixing between the pre-mixed ammonia charge and the atomized diesel. The remaining injected diesel is rapidly compressed and ignited near the top dead center, which drives the combustion within the chamber. Different from the DICI mode, the subsequent combustion of the ignition is no longer controlled solely by the ammonia flame speed. The dispersed distribution of highly reactive fuels improves the ignition properties of ammonia. Consequently, the second heat release peak has a higher intensity than that of DICI under similar conditions. Meanwhile, an early second injection timing advances the combustion phase, leading to the elevated heat release peak and maximum pressure, which is similar to previous research findings [37]. Unfortunately, more aggressive diesel injection tends to cause the maximum pressure to exceed the designed pressure limit (approximately 20 MPa). The pre-injected diesel fuel (25 mg, 60 °CA BTDC) enhances the global reactivity of the in-cylinder mixture, enabling rapid flame propagation following localized ignition, and enhancing the combustion efficiency.
Analysis of the combustion process further validates the finding of the cylinder pressure curves, as exhibited in Fig. 7. Across both combustion modes, advancement of the diesel injection timing reduces the overall combustion duration, represented by the decreased intervals between CA10 and CA90. In the PIRS mode, the combustion duration is more sensitive to the advanced diesel injection (indicated by the blue arrow) compared with that of DICI (indicated by the purple arrow). The shorter combustion duration is mainly attributed to the shorter intervals between CA50 and CA90. Although the CA10-CA50 interval is slightly reduced, diesel combustion remains the dominant factor. The reason lies in the boosting effect of pre-injected diesel on in-cylinder reactivity, as discussed above. In the PIRS mode, the combustion of the dispersed distribution of diesel creates favorable environment for ammonia. Consequently, the advanced diesel ignition establishes an optimal thermodynamic environment for ammonia combustion, further accelerating the late-stage combustion. The advantages of the PIRS mode can also be reflected in the engine efficiency. Both modes have some improvement in efficiency as the injection timing is advanced. When the second diesel injection timing is held constant, PIRS demonstrates an efficiency benefit exceeding 1.5% over DICI. Partially pre-mixed diesel creates a favorable combustion environment for ammonia combustion, and promotes high thermal efficiency. When the second injection timing comes to 14 °CA BTDC, the BTE achieves 46.1%.
The representative emissions are exhibited in Fig. 8. High ammonia emissions can be observed in the DICI mode (above 8000 ppm), which decrease slightly with advanced diesel injection timing. The difference in thermal efficiency is visualized in the combustion efficiency of ammonia. The improved ammonia combustion results in a rapid decline (more than 50%) in unburned emissions in the PIRS mode. Furthermore, when the second diesel injection timing is optimized at 14 °CA BTDC, the ammonia concentration in the exhaust in the PIRS mode was only 26.8% of that in the DICI mode, confirming the improvement in ammonia combustion efficiency. The experimental results reveal an inverse correlation between unburned NH3 and NOx, which is one of the combustion products. A similar phenomenon has been observed in other scholars [38,49]. NOx emissions are consistently maintained above 5000 ppm as combustion efficiency improves, which is higher than conventional diesel operation under equivalent load conditions, putting pressure on aftertreatment systems. Ammonia in the exhaust can replace the urea injected in conventional selective catalytic reduction (SCR). Therefore, the modulation of the ratio of NH3 to NOx by in-cylinder combustion to achieve equivalent reaction in the aftertreatment will be a viable option. However, the exhaust gas typically does not contain enough ammonia to reduce all the NOx to N2 in the SCR at the sweet efficiency point. It means that the additional reductant has to be used in the reactor, which may increase the complexity of the aftertreatment system. How to balance nitrogen-containing emissions with engine thermal efficiency will be one of the priorities for ammonia engine aftertreatment design. An alternative strategy involves reducing engine efficiency to maintain enough ammonia in the SCR to consume the fuel-based NOx. The small amount of residual ammonia can be further collected and disposed of by the ammonia slip catalyst (ASC).
Furthermore, the CO and THC emissions also differ between the two modes, as exhibited in Fig. 8(b). The CO emissions are maintained at around 100 ppm due to the λ being controlled at approximately 1.3. In the PIRS mode, the concentrated combustion caused by advanced injection timing contributes to the CO emissions reduction. The THC emissions in the PIRS mode are always higher than those of the DICI mode, which indicates the incomplete combustion of diesel. This phenomenon is a result of the peripheral diesel distribution in the combustion chamber, where the high ammonia concentration inhibits the diesel reactions. The pre-injected diesel and pre-mixed ammonia compete with each other during the combustion process. Furthermore, the CO2 emissions remain virtually unchanged due to the near-fixed carbon input. N2O emissions likewise decline as ammonia combustion efficiency increases. When injection timing is advanced to 14 °CA BTDC in the PIRS mode, the N2O emissions are reduced to 16 ppm, which weakens the emission greenhouse effect compared to that of the DICI mode (above 20 ppm in all operating conditions). Consequently, the PIRS mode significantly improves ammonia combustion efficiency and reduces unburned ammonia emissions, accompanied by rapidly rising NOx and somewhat elevated THC emissions. The advantages of the PIRS mode in reducing N2O emissions also make it more environmentally friendly.
When focusing on the PIRS mode, the effect of different pre-injection masses on emissions is further investigated. Fig. 9 exhibits the change of NH3 and NOx emissions as the injection timing advanced at different pre-injection masses. When the first injection mass is 20 or 25 mg, an advanced combustion phase usually reduces unburned ammonia emissions. Elevated unburned NH3 concentration (> 6000 ppm) is observed as the pre-injection fuel mass is 20 mg. This indicates that the amount of pre-injected diesel needs to reach a certain level to improve the overall in-cylinder combustion. In this condition, the advanced second injection timing can accelerate the total combustion process. This is beneficial for ammonia combustion and reduces the unburned emissions. In comparison, the lowest unburned ammonia emissions can be observed when the pre-injection mass is 30 mg, in which the pre-injected ratio of diesel is more than 80%. Though the amount of fuel injected in the second stage is very small, it is still necessary as the ignition source. The increase in the proportion of pre-injected diesel contributes to the complete combustion of ammonia [37,58]. Meanwhile, the NOx emissions are always at a high level when the first injection mass is 25 or 30 mg. Lower NOx emissions can be observed when the first injection mass is 20 mg, echoing the high ammonia emissions. The emission pattern of N2O is similar to that of ammonia. Reduced N2O emissions occur at 30 mg pre-injection, with a minimum of 11.3 ppm N2O when the combustion phase is advanced. Therefore, it is important to ensure that a sufficient amount of pre-injected diesel is adequately mixed with ammonia to increase in-cylinder reactivity. The concentration of unburned ammonia decreases as the percentage of pre-injection increases.
From the above discussion, a small amount of pilot ignition fuel is required for combustion initiation in the PIRS mode. Therefore, when the limits of ammonia substitution rates are explored, the micro-pilot diesel ignition is employed to promote the combustion of high-concentration ammonia. It is found to be difficult to maintain stable engine operation as the amount of diesel gradually decreases, especially when the injection mass is less than 5 mg per cycle. Advancing the timing of single diesel injection contributes to the in-cylinder ignition. Finally, the highest AEF can reach 99.1%, at a coefficient of variation of the indicated mean effective pressure of 9.3%. At the maximum AEF point, the carbon reduction benefit is significant, resulting in CO2 emissions equivalent to only 1.3% of those under diesel operation at the same load. Furthermore, when accounting for the greenhouse effect of N2O using its CO2 equivalence factor (298:1) [59], the total equivalent GHG emissions are also reduced to 13.2% compared to the diesel baseline. However, the unburned ammonia emissions would exceed 12 000 ppm with high NOx emissions over 4000 ppm. Thus, the combination of multi-stage SCR and ASC would present an effective approach for emission treatment in this scenario.
As mentioned above, in the process of developing the performance of ammonia-fueled engines, several indicators are in a delicate balance. The incomplete combustion of ammonia and diesel, represented by the nitrogen- and carbon-containing emissions, is in dynamic equilibrium when λ is adjusted. When striving for an efficient combustion mode, high thermal efficiency can result in low unburned ammonia emissions, but it often causes fuel-type NOx to rise rapidly, requiring the addition of a reducing agent in the aftertreatment. The carbon emission reduction benefits resulting from the ultra-high AEF can also be diluted by its deteriorated efficiency and increased nitrogen emissions. The increase in load also helps improve the efficiency and emissions of ammonia engines [48]. After comprehensively evaluating various indicators, a relatively reasonable operation condition is found when AEF is 90% and λ is controlled at 1.3. Under this condition, BMEP is maintained as 2.26 MPa, and the PIRS mode is employed to improve the in-cylinder combustion, where 15 mg diesel is pre-mixed at 60 °CA BTDC and a micro-diesel of around 5 mg is injected for ignition at 15 °CA BTDC. As a result, the combustion of ammonia and diesel, emissions of unburned ammonia and NOx, and the balance between AEF and BTE have all been taken into consideration. The unburned NH3 and NOx emissions are maintained at similar levels, which facilitates the treatment in aftertreatment processes. Furthermore, the low GHG emissions can be achieved, in which CO2 emissions contribute to the 1.2% and N2O emissions contribute to 23.6 ppm, as exhibited in Fig. 10.
4. Conclusions
The performance of an ammonia-fueled engine was investigated in this study. The combustion modes, primarily due to reactivity stratification and the interaction with the λ, were discussed in relation to combustion and emissions. The clean and efficient combustion mode suitable for high AEF in ammonia engines was found. The conclusions are as follows.
(1) The PIRS mode characterized by designed premixed fuel stratification is attainable at high AEF conditions. This mode allows part of the diesel being premixed in advance, enhancing the overall reactivity of the in-cylinder fuel-air mixture. Subsequently, ignition initiated by a small quantity of direct-injected diesel utilizes the dispersed premixed diesel as a medium to promote extensive ammonia ignition, effectively overcoming the challenge of ammonia ignition difficulty. The enhanced ignition reactivity and improved combustion rate shorten the combustion duration, resulting in a significant efficiency increase and an elevated ammonia complete combustion.
(2) The effect of the λ on combustion is centered on the ignition delay. The pre-injected diesel tends to ignite and release heat at earlier moment under high intake air mass conditions, which advances the combustion phase and leads to the combustion separation. The interval between CA50 and CA90 is mainly controlled by ammonia combustion, which does not vary significantly as the λ changes. The unburned ammonia emissions decrease as the λ decreases. Furthermore, the reduced air mass deteriorates the diesel combustion when the λ is lower than 1.2, which is represented by the rapidly rising CO emissions, signifying an increase in the proportion of incomplete combustion.
(3) Changed diesel injection strategy results in different fuel-air mixing modes, which influence the combustion process while alter the ignition method. The improved ammonia combustion in PISR results in a considerable reduction in ammonia emissions compared to those of DICI. Rapid ignition and concentrated combustion improve the engine efficiency, allowing the BTE to exceed 46%, which is higher than that of diesel conditions. The ammonia emissions can be eventually controlled to below 2000 ppm along with ultra-low N2O emissions of 11.3 ppm, when the pre-injection mass is increased to 30 mg. Meanwhile, the NOx and ammonia emissions exhibit a negative correlation trend, in which the NOx emissions approach over 6000 ppm when ammonia combustion is improved.
(4) A minimum amount of direct diesel injection to ensure stable operation is obtained through the tests of the limiting injection diesel mass, and the highest AEF of 99.1% is achieved. In the maximum AEF condition, the carbon reduction benefits account for 1.3% of the CO2 emissions in diesel combustion at the same load condition. The globally optimized condition can be achieved when AEF is 90% in the PIRS mode, where ammonia and NOx emissions are regulated at similar concentrations and a high efficiency is maintained. It allows both nitrogenous pollutants to be adequately treated by SCR without additional reductant additions.
CRediT authorship contribution statement
Yuxiao Qiu: Writing - original draft, Investigation, Data curation, Conceptualization. Yanyuan Zhang: Writing - original draft, Investigation. Yingnan Yang: Investigation, Data curation. You Zhang: Investigation, Data curation. Dong Han: Writing - review & editing, Supervision, Resources, Conceptualization. Zhen Huang: Supervision, Resources, Conceptualization.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
This research work was supported by the National Key Research and Development Program of China (2022YFE0209000), the Science and Technology Commission of Shanghai Municipality (24120742400 and 24120750400), the Science and Technology Cooperation Program of Shanghai Jiao Tong University in Inner Mongolia Autonomous Region–Action Plan of Shanghai Jiao Tong University for Revitalizing Inner Mongolia through Science and Technology, and SJTU Wuxi Carbon Neutral Energy & Power Innovation Center.
S.Chu, Q.Wang. Climate change and innovative paths to a more sustainable future. Front Energy, 18 (6) (2024), pp. 717-726.
[2]
W.S.Chai, Y.Bao, P.Jin, G.Tang, L.Zhou. A review on ammonia, ammonia-hydrogen and ammonia-methane fuels. Renew Sustain Energy Rev, 147 (2021), Article 111254.
[3]
G.Wang, Z.Chen, J.Xie, L.Ding, J.Zhu, W.Wei, et al. Recent trends and prospects in electrochemical nitrate reduction to ammonia with an emphasis on cobalt catalysts. Coord Chem Rev, 539 (2025), Article 216751.
[4]
C.Kurien, M.Mittal. Review on the production and utilization of green ammonia as an alternate fuel in dual-fuel compression ignition engines. Energ Conver Manage, 251 (2022), Article 114990.
[5]
J.Cao, Y.Hu, Y.Zheng, W.Zhang, B.Yu. Recent advances and challenges of nitrogen/nitrate electro catalytic reduction to ammonia synthesis. Front Energy, 18 (2) (2024), pp. 128-140.
[6]
F.Gong, Y.Jing, R.Xiao. Plasma-assisted ammonia synthesis under mild conditions for hydrogen and electricity storage: mechanisms, pathways, and application prospects. Front Energy, 18 (4) (2024), pp. 418-435.
[7]
J.H.Lee, J.H.Kim, J.H.Park, O.C.Kwon. Studies on properties of laminar premixed hydrogen-added ammonia/air flames for hydrogen production. Int J Hydrogen Energy, 35 (3) (2010), pp. 1054-1064.
[8]
Y.Ding, G.Pan, D.Han, Z.Huang. Combustion and emissions of an ammonia-gasoline dual-fuel spark ignition engine: effects of ammonia substitution rate and spark ignition timing. Int J Hydrogen Energy, 122 (2025), pp. 348-358.
[9]
W.Avery. A role for ammonia in the hydrogen economy. Int J Hydrogen Energy, 13 (12) (1988), pp. 761-773.
[10]
P.Dimitriou, R.Javaid. A review of ammonia as a compression ignition engine fuel. Int J Hydrogen Energy, 45 (11) (2020), pp. 7098-7118.
[11]
Q.Cheng, A.Muhammad, O.Kaario, Z.Ahmad, L.Martti. Ammonia as a sustainable fuel: review and novel strategies. Renew Sustain Energy Rev, 207 (2025), Article 114995.
[12]
Y.Fan, T.Zhu, Z.Li, J.Li, F.Ren, Z.Gao, et al. Non-equilibrium plasma cracking assisted ammonia marine engine for zero carbon emissions. Int J Hydrogen Energy, 112 (2025), pp. 433-445.
[13]
M.H.Dinesh, J.K.Pandey, G.N.Kumar. Study of performance, combustion, and NOx emission behavior of an SI engine fuelled with ammonia/hydrogen blends at various compression ratio. Int J Hydrogen Energy, 47 (60) (2022), pp. 25391-25403.
[14]
A.Mercier, C.Mounaïm-Rousselle, P.Brequigny, J.Bouriot, C.Dumand. Improvement of SI engine combustion with ammonia as fuel: effect of ammonia dissociation prior to combustion. Fuel Commun, 11 (2022), Article 100058.
[15]
J.Lei, C.Shi, G.Tian, J.Niu, J.Bao, X.Yang. Comparative analysis and amelioration of different ammonia/hydrogen injection methods in a carbon-free rotary engine. Energ Conver Manage, 344 (2025), Article 120286.
[16]
H.Guo, B.Liko, D.Stevenson, K.Austin. An experimental investigation on combustion and emissions of a hydrogen enriched ammonia-diesel dual fuel engine at a medium load condition. J Eng Gas Turbine Power, 147 (9) (2025), Article 091010.
[17]
Y.Wang, X.Zhou, L.Liu. Study on the mechanism of the ignition process of ammonia/hydrogen mixture under high-pressure direct-injection engine conditions. Int J Hydrogen Energy, 46 (78) (2021), pp. 38871-38886.
[18]
Z.Yu, H.Zhang. End-gas autoignition and knocking combustion of ammonia/hydrogen/air mixtures in a confined reactor. Int J Hydrogen Energy, 47 (13) (2022), pp. 8585-8602.
[19]
G.Tang, P.Jin, Y.Bao, W.S.Chai, L.Zhou. Experimental investigation of premixed combustion limits of hydrogen and methane additives in ammonia. Int J Hydrogen Energy, 46 (39) (2021), pp. 20765-20776.
[20]
K.P.Shrestha, C.Lhuillier, A.A.Barbosa, P.Brequigny, F.Contino, C.Mounaïm-Rousselle, et al. An experimental and modeling study of ammonia with enriched oxygen content and ammonia/hydrogen laminar flame speed at elevated pressure and temperature. Proc Combust Inst, 38 (2) (2021), pp. 2163-2174.
[21]
G.B.Ariemma, G.Sorrentino, R.Ragucci, M. DeJoannon, P.Sabia. Ammonia/methane combustion: stability and NOx emissions. Combust Flame, 241 (2022), Article 112071.
[22]
M.Zhang, Z.An, X.Wei, J.Wang, Z.Huang, H.Tan. Emission analysis of the CH4/NH3/air co-firing fuels in a model combustor. Fuel, 291 (2021), Article 120135.
[23]
A.Yousefi, H.Guo, M.Birouk. Split diesel injection effect on knocking of natural gas/diesel dual-fuel engine at high load conditions. Appl Energy, 279 (2020), Article 115828.
[24]
B.Guo, M.Ichiyanagi, K.Kajiki, N.Aratake, Q.Zheng, M.Kodaka, et al. Combustion analysis of ammonia fueled high compression ratio SI engine with glow plug and sub-chamber: effects of ammonia content under condition of CO-combustion with gasoline/ammonia/air. Int J Automot Eng, 13 (1) (2022), pp. 1-8.
[25]
F.Salek, M.Babaie, A.Shakeri, S.V.Hosseini, T.Bodisco, A.Zare. Numerical study of engine performance and emissions for port injection of ammonia into a gasoline/ethanol dual-fuel spark ignition engine. Appl Sci, 11 (4) (2021), p. 1441.
[26]
Z.Liu, L.Zhou, L.Zhong, H.Wei. Enhanced combustion of ammonia engine based on novel air-assisted pre-chamber turbulent jet ignition. Energ Conver Manage, 276 (2023), Article 116526.
[27]
M.C.Chiong, C.T.Chong, J.H.Ng, S.Mashruk, W.W.F.Chong, N.A.Samiran, et al. Advancements of combustion technologies in the ammonia-fuelled engines. Energ Conver Manage, 244 (2021), Article 114460.
[28]
C.Mounaïm-Rousselle, A.Mercier, P.Brequigny, C. Dumand, J.Bouriot, S.Houillé. Performance of ammonia fuel in a spark assisted compression ignition engine. Int J Engine Res, 23 (5) (2022), pp. 781-792.
[29]
T.Zhu, X.Yan, Z.Gao, Y.Qiu, L.Zhu, Z.Huang. Combustion and emission characteristics of ammonia-hydrogen fueled SI engine with high compression ratio. Int J Hydrogen Energy, 62 (2024), pp. 579-590.
[30]
Z.Liu, L.Zhou, H.Wei. Experimental investigation on the performance of pure ammonia engine based on reactivity controlled turbulent jet ignition. Fuel, 335 (2023), Article 127116.
[31]
X.Wu, Y.Feng, G.Xu, Y.Zhu, P.Ming, L.Dai. Numerical investigations on charge motion and combustion of natural gas-enhanced ammonia in marine pre-chamber lean-burn engine with dual-fuel combustion system. Int J Hydrogen Energy, 48 (30) (2023), pp. 11476-11492.
[32]
N.Førby, T.B.Thomsen, R.F.Cordtz, F.Bræstrup, J.Schramm. Ignition and combustion study of premixed ammonia using GDI pilot injection in CI engine. Fuel, 331 (2023), Article 125768.
[33]
Y.Qu, Y.Qiu, Y.Zhang, Y.Zhang, C.Mounaïm-Rousselle, D.Han, et al. Review of combustion technologies for ammonia-diesel compression ignition engines: Approaches to achieving high ammonia substitution. Front Energy, 19 (2025), pp. 907-924.
[34]
C.Yang, Z.Wang, J.Li, X.Cheng. Effects of ammonia energy fractions, diesel injection timings, and loads on combustion and emission characteristics of PFI-DI ammonia-diesel engines. Int J Engine Res, 25 (4) (2024), pp. 743-757.
[35]
C.Shi, Z.Zhang, H.Wang, J.Wang, T.Cheng, L.Zhang. Parametric analysis and optimization of the combustion process and pollutant performance for ammonia-diesel dual-fuel engines. Energy, 296 (2024), Article 131171.
[36]
A.Yousefi, H.Guo, S.Dev, B.Liko, S.Lafrance. Effects of ammonia energy fraction and diesel injection timing on combustion and emissions of an ammonia/diesel dual-fuel engine. Fuel, 314 (2022), Article 122723.
[37]
A.Yousefi, H.Guo, S.Dev, S.Lafrance, B.Liko. A study on split diesel injection on thermal efficiency and emissions of an ammonia/diesel dual-fuel engine. Fuel, 316 (2022), Article 123412.
[38]
Y.Niki. Reductions in unburned ammonia and nitrous oxide emissions from an ammonia-assisted diesel engine with early timing diesel pilot injection. J Eng Gas Turbine Power, 143 (9) (2021), p. 91014.
[39]
A.J.Reiter, S.C.Kong. Demonstration of compression-ignition engine combustion using ammonia in reducing greenhouse gas emissions. Energy Fuels, 22 (5) (2008), pp. 2963-2971.
[40]
Z.Lin, Y.Liu, Q.Chen, Q.Sun, W.Zhu, Y.Qi, et al. Experimental study on the combustion pattern in an ammonia engine using micro diesel ignition. Energy, 320 (2025), Article 135480.
[41]
X.Zhou, T.Li, N.Wang, X.Wang, R.Chen, S.Li. Pilot diesel-ignited ammonia dual fuel low-speed marine engines: a comparative analysis of ammonia premixed and high-pressure spray combustion modes with CFD simulation. Renew Sustain Energy Rev, 173 (2023), Article 113108.
[42]
W.Sun, W.Zeng, L.Guo, H.Zhang, Y.Yan, S.Lin, et al. An optical study of the combustion and flame development of ammonia-diesel dual-fuel engine based on flame chemiluminescence. Fuel, 349 (2023), Article 128507.
[43]
Y.Pei, D.Wang, S.Jin, Y.Gu, C.Wu, B.Wu. A quantitative study on the combustion and emission characteristics of an ammonia-diesel dual-fuel (ADDF) engine. Fuel Process Technol, 250 (2023), Article 107906.
[44]
B.Wu, Z.Zi, S.Jin, Y.Pei, D.Wang. Effect of diesel injection strategy and ammonia energy fraction on ammonia-diesel premixed-charge compression ignition combustion and emissions. Fuel, 357 (2024), Article 129785.
[45]
Powertrain system for a sustainable future. In: Proceedings of the International Conference on Powertrain Systems for a Sustainable Future 2023; 2023 Nov 29-30; London, UK. CRC Press; 2023.
[46]
C.Shi, R.Duan, T.Cheng, F.Nie, X.Yan, J.Zhu. Understanding the role of ammonia combined injection in improving combustion and emissions characteristics for heavy-duty CI engine. Energy, 332 (2025), Article 135959.
[47]
Z.Zhang, W.Long, P.Dong, H.Tian, J.Tian, B.Li, et al. Performance characteristics of a two-stroke low speed engine applying ammonia/diesel dual direct injection strategy. Fuel, 332 (2023), Article 126086.
[48]
Y.Qiu, Y.Zhang, Y.Shi, Y.Zhang, Z.Wang, H.Lin, et al. Ammonia fueled engine with diesel pilot ignition: approach to achieve ultra-high ammonia substitution. Int J Engine Res, 25 (9) (2024), pp. 1751-1763.
[49]
X.Wang, T.Li, X.Zhou, S.Huang, R.Chen, P.Yi, et al. Reductions in GHG and unburned ammonia of the pilot diesel-ignited ammonia engines by diesel injection strategies. Appl Therm Eng, 260 (2025), Article 124967.
[50]
S.Oh, C.Park, S.Kim, Y.Kim, Y.Choi, C.Kim. Natural gas-ammonia dual-fuel combustion in spark-ignited engine with various air-fuel ratios and split ratios of ammonia under part load condition. Fuel, 290 (2021), Article 120095.
[51]
Y.Sun, Y.Qian, Z.Sun, Z.Gong, X.Gu. Experimental and simulation study on the combustion characteristics of ammonia/n-dodecane mixtures. Fuel, 358 (2024), Article 130143.
[52]
X.Wang, T.Li, R.Chen, S.Li, M.Kuang, Y.Lv, et al. Exploring the GHG reduction potential of pilot diesel-ignited ammonia engines—effects of diesel injection timing and ammonia energetic ratio. Appl Energy, 357 (2024), Article 122437.
[53]
R.Chen, T.Li, X.Wang, S.Huang, X.Zhou, S.Li, et al. Engine-out emissions from an ammonia/diesel dual-fuel engine—the characteristics of nitro-compounds and GHG emissions. Fuel, 362 (2024), Article 130740.
[54]
J.Liu, X.Wang, W.Zhao, P.Sun, Q.Ji. Effects of ammonia energy fraction and diesel injection parameters on combustion stability and GHG emissions characteristics in a low-loaded ammonia/diesel dual-fuel engine. Fuel, 360 (2024), Article 130544.
[55]
O.Mathieu, E.L.Petersen. Experimental and modeling study on the high-temperature oxidation of ammonia and related NOx chemistry. Combust Flame, 162 (3) (2015), pp. 554-570.
[56]
R.Tang, Q.Xu, J.Pan, J.Gao, Z.Wang, H.Wei, et al. An experimental and modeling study of ammonia oxidation in a jet stirred reactor. Combust Flame, 240 (2022), Article 112007.
[57]
M.Qian, B.Guan, Z.Zhuang, J.Chen, L.Zhu, Z.Ma, et al. Progress in catalytic strategies for mitigating NOx/N2O emissions: from mechanisms to applications. Cat Sci Technol, 15 (7) (2025), pp. 2061-2103.
[58]
S.Mi, H.Wu, X.Pei, C.Liu, L.Zheng, W.Zhao, et al. Potential of ammonia energy fraction and diesel pilot-injection strategy on improving combustion and emission performance in an ammonia-diesel dual fuel engine. Fuel, 343 (2023), Article 127889.
[59]
X.Zhou, T.Li, R.Chen, Y.Wei, X.Wang, N.Wang, et al. Ammonia marine engine design for enhanced efficiency and reduced greenhouse gas emissions. Nat Commun, 15 (1) (2024), p. 2110.
AI Summary 中Eng×
Note: Please be aware that the following content is generated by artificial intelligence. This website is not responsible for any consequences arising from the use of this content.
PDF
(2506KB)
