Ammonia Emissions in SI Engines Fueled with LPG

: Ammonia is a toxic exhaust component emitted from internal combustion engines. Both pure ammonia and the products of its reaction with nitrogen and sulfur compounds, being the source of particulate matter (PM) emissions, are dangerous for human health and life. The aim of the article was to demonstrate that NH 3 can be produced in exhaust gas after-treatment systems of spark-ignition (SI) engines used in light-duty vehicles. In some cases, NH 3 occurs in high enough concentrations that can be harmful and dangerous. It would be reasonable to collect research data regarding this problem and consider the advisability of limiting these pollutant emissions in future regulations. The article presents the results of the spark-ignition engine testing on an engine test bench and discusses the impact of the air–fuel ratio regulation and some engine operating parameters on the concentration of NH 3 . It has been proven that in certain engine operating conditions and a combination of circumstances like the three-way catalytic reactor (TWC) temperature and periodic enrichment of the air–fuel mixture may lead to excessive NH 3 emissions resulting from the NO conversion in the catalytic reactor. This is a clear disadvantage due to the lack of limitation of these pollutant emissions by the relevant type-approval regulations. This article should be a contribution to discussion among emissions researchers whether future emission regulations (e.g., Euro 7 or Euro VII) should include a provision to reduce NH 3 emissions from all vehicles.


Introduction
Ammonia is a toxic substance classified by international regulations [1]. US regulations allow for the NH 3 concentration limit of 25 ppm during 8-h human exposure [2]. NH 3 is not a product forming itself inside the engine combustion chamber during the combustion process. In spark ignition (SI) engines, NH 3 is formed in exhaust after-treatment systems during NO conversion in the presence of CO. Moreover, NH 3 may react with the acids contained in the exhaust gas, forming particulate matter (PM) and harmful aerosols. Exposure to higher NH 3 concentration could cause irritation of the skin, eyes, nose, or throat due to direct contact. Ammonia, even at low concentrations, has an unpleasant odor when released into the air but most notably harms vegetation, particularly at high concentrations. In water bodies, ammonia causes more serious harm due to the toxicity to organisms living in water [3]. Regulation No. 595/2009 for the value of NH 3 concentration in Euro VI has set 10 ppm as the limit for NH 3 concentration in the exhaust gas emitted from a heavy-duty vehicle [4]. So far, there are no NH 3 limits for light-duty vehicles. NH 3 emissions have been associated mainly with the emissions from heavy-duty vehicles using selective catalytic reduction (SCR) as an after-treatment device. This type of catalytic reactor allows us to reduce NO x emissions in the exhaust of the engines fueled with a lean air-fuel mixture. In a stoichiometric engine, carbon monoxide is used to reduce nitrogen oxides inside three-way catalytic reactors (TWCs). In compression ignition engines, fueled with a lean air-fuel mixture with an excess of oxygen, most of the CO contained in the exhaust gases is oxidized to CO 2 . This process slows down the NO x conversion and prevents the use of such reactors in stoichiometric spark-ignition engines. The operation principle of SCR reactors is based on the use of chemical reactions between Energies 2021, 14, 691 2 of 12 nitrogen oxides and ammonia formed from urea injected into the reactor [4]. Overdosing of the amount of injected urea results in increased NH 3 emissions from the exhaust system. However, urea overdosing is not the only factor increasing NH 3 emissions from the internal combustion engine. NH 3 can also be emitted from a spark ignition engine operating as a stoichiometric engine. NH 3 emissions are mainly produced from NO emissions by reactions (1) and (2) The greater the concentrations of NO, CO, and H 2 in the exhaust gas are, the greater the mass of the NH 3 is. These gases are the basic components of the exhaust gas on which the mass of the NH 3 depends. CO is a natural component of exhaust gases and has higher concentrations when the air-fuel mixture is rich. H 2 can be obtained in the exhaust gas in two ways-during the combustion stroke, where, under high pressures and temperature, the hydrogen contained in the fuel is released or by means one of the below reactions: the water gas shift (3) or the steam reforming (4).
C n H m + n H 2 O → n CO + (m/2 + n) H 2 (4) Thus, higher engine NO, CO, HC, and H 2 emissions could bring about higher NH 3 emissions. It can be seen that the type of used fuel may have an influence on the amount of H 2 produced in the combustion chamber according to reaction (4). In particular, the value of n/m fuel ratio is important. For lighter gaseous fuels, such as methane or propane, a higher value of the n/m ratio means that the fuel contains more hydrogen, and therefore more water will be formed in the products of its combustion, which, according to reactions (3) and (4), may change the quantity of hydrogen participating in the process of NH 3 formation. In a stoichiometric engine, lambda sensor periodically and alternately changes the air-fuel ratio between rich-to-lean stages, but the average air-fuel equivalent ratio oscillates around a value equal to 1.00. During engine operation, it is possible to identify the operating conditions in which the engine has a rich (λ < 1.00) and a lean (λ > 1.00) air-fuel ratio. When the air-fuel mixture is rich, conditions for the formation of NH 3 occur. Due to the limited access to oxygen, the raw exhaust gas will contain more CO and unburned hydrocarbons (C n H m ), because these gases cannot be fully oxidized. This situation leads to the generation of NH 3 emissions in accordance with reactions (1)- (4).
The greater and deeper the changes of the air-fuel ratio occur, the greater tendency the engine shows for increasing NH 3 emissions. Therefore, by analyzing the influence of air-fuel ratio on NH 3 emissions, it can be assumed that the engine will have the lowest NH 3 emission during operation in stationary, steady-state conditions with constant speed and torque. In the event of a sudden change of operating parameters, e.g., rapid acceleration of the vehicle during aggressive driving and temporary enrichment of air-fuel ratio, or only with frequent changes of the engine operating parameters caused by the vehicle's driving, conditions for higher NH 3 emissions will be created.
Apart from the concentration of gases participating in the reactions (1) and (2) (NO, CO, and H 2 ), the mass of NH 3 emission will be determined by the kinetics of these reactions depending on the temperature at which the NO converts to NH 3 . According to [5], reaction (3) takes place at lower TWC temperatures (300-500 • C), whereas reaction (4) occurs at higher temperatures (>500 • C).
Gaseous ammonia has been shown to contribute to the formation of airborne particulate matter. Ammonia emitted into ambient air neutralizes nitric and sulfuric acid to form ammonia nitrate (NH 4 NO 3 ) and sulfate (NH 4 ) 2 SO 4 , two important components contributing to airborne fine particles or PM2. 5 is a major constituent of outdoor air pollution and can cause cancer in humans (IARC 2013). In addition to human health effects, ammonia also impacts terrestrial vegetation.
Three-way catalysts have been used in modern light-duty vehicles driven by sparkignition engines for over three decades since the introduction of low-emission vehicles to the automotive market. Today, this is a typical part of the exhaust after-treatment system. The time that has elapsed since the start of using SCR technology in heavy-duty vehicles is a little shorter. TWC and SCR are identified as the main causes of NH 3 emissions in the exhaust gas. Although agriculture is still the main source of NH 3 emissions to the atmosphere, some researchers have reported an increase in NH 3 emissions from vehicles in recent years.
Ammonia emitted into the ambient air has the ability to react with nitric acid to form solid ammonium nitrate, which is the secondary source of pollutant emissions in the form of PM emissions. For this reason, an increase in NH 3 emissions to the atmosphere can be dangerous and undesirable. Therefore, it is time to take a closer look at the problem of NH 3 emissions from vehicles, including also those powered by SI engines.

Test Bed and Tested Object
The aim of the testing was to determine the influence of selected engine parameters on NH 3 emissions. The experimental testing was carried out on a spark-ignition engine fueled with propane. It was a V-type engine with a displacement of 7.2 dm 3 , equipped with two TWCs. The diameter of the TWC's metal substrate was 148 mm, catalyst body volume 4.22 dm 3 , cell density 400 cpsi, and precious metal charge 70 g/ft 3  The used experimental equipment allowed to ensure the accuracy of emission measurements of 2%. Figure 1 shows the NH 3 concentration in the exhaust gas and the exhaust temperature during the hot World Harmonized Transient Cycle (WHTC). In accordance with UN Regulation No. 49.06, the WHTC cycle is a commonly used pollutant emissions test in Euro VI standard. The figure shows that NH 3 reaches relatively low concentrations at both low and high exhaust gas temperatures, which roughly corresponds to the urban and highway phases of the WHTC cycle. The relatively high concentration of NH 3 is achieved during rural driving, where the exhaust gas temperature is approx. 550-600 • C. • AVL AMA i60 LDD NH3/H2O type analyzer by AVL being a diode laser gas analyzer with a measuring principle based on the specific light absorption of different gas components; • airflow meter Sensyflow P-type to measure the air consumption of by the engine, operating on the principle of measuring the resistance of the heated wire; • mass flowmeter CMF 025M 300 NQFZGZZZ by Emerson for measurement of gaseous fuel consumption, operating on Coriolis principle.

Result of the Testing
The used experimental equipment allowed to ensure the accuracy of emission measurements of 2%. Figure 1 shows the NH3 concentration in the exhaust gas and the exhaust temperature during the hot World Harmonized Transient Cycle (WHTC). In accordance with UN Regulation No. 49.06, the WHTC cycle is a commonly used pollutant emissions test in Euro VI standard. The figure shows that NH3 reaches relatively low concentrations at both low and high exhaust gas temperatures, which roughly corresponds to the urban and highway phases of the WHTC cycle. The relatively high concentration of NH3 is achieved during rural driving, where the exhaust gas temperature is approx. 550-600 °C.  Figure 2 shows the dependence of NH3 concentration and the exhaust gas temperature in the WHTC cycle. The total duration of the WHTC cycle is 1800 s. The cycle covers three successive different types of driving conditions-urban, rural, and highway, 600 s each. Urban driving mode is characterized by the greatest frequency of changes in engine speed and loading. In highway driving mode, the changes of engine speed and loading are relatively the smallest.

Result of the Testing
The highest concentrations of NH3 correspond to the exhaust gas temperature range of 570-625 °C. Outside this temperature range, the NH3 concentration decreases with both increasing and decreasing temperature. The distribution of NH3 concentrations ( Figure 2) indicates that the NH3 concentration increases along with the increase in exhaust gas temperature from the lowest values to approx. 570 °C. This means that one of the reactions responsible for the NH3 formation (i.e., (1) or (2)) is accelerated. It is claimed [3] that this is an effect of water gas shift reaction (reaction (3)). At higher temperatures, this reaction slows down and the hydrogen needed for NH3 formation is produced as a result of hydrocarbons steam reforming (reaction (4)).  Figure 2 shows the dependence of NH 3 concentration and the exhaust gas temperature in the WHTC cycle. The total duration of the WHTC cycle is 1800 s. The cycle covers three successive different types of driving conditions-urban, rural, and highway, 600 s each. Urban driving mode is characterized by the greatest frequency of changes in engine speed and loading. In highway driving mode, the changes of engine speed and loading are relatively the smallest.
The highest concentrations of NH 3 correspond to the exhaust gas temperature range of 570-625 • C. Outside this temperature range, the NH 3 concentration decreases with both increasing and decreasing temperature. The distribution of NH 3 concentrations ( Figure 2) indicates that the NH 3 concentration increases along with the increase in exhaust gas temperature from the lowest values to approx. 570 • C. This means that one of the reactions responsible for the NH 3 formation (i.e., (1) or (2)) is accelerated. It is claimed [3] that this is an effect of water gas shift reaction (reaction (3)). At higher temperatures, this reaction slows down and the hydrogen needed for NH 3 formation is produced as a result of hydrocarbons steam reforming (reaction (4)). Most of the registered points in the WHTC cycle having the highest concentrations of NH3 (Figure 2) correspond to the exhaust gas temperatures with the highest frequency of occurrence in the cycle ( Figure 3); hence, they have significant importance and influence on total NH3 emissions in the whole emission cycle. The concentrations of the gases participating in the reactions (1)-(4) should be correlated with each other. However, only the concentration of CH4 from the concentrations of NO, CO, particulate number emissions (PN), and CH4 presented in Figure 4, i.e., all exhaust gas components related to the formation of NH3, is similar to the course of NH3 concentrations ( Figure 2). This observation is confirmed by the values of the correlation coefficients presented in Table 1. Most of the registered points in the WHTC cycle having the highest concentrations of NH3 ( Figure 2) correspond to the exhaust gas temperatures with the highest frequency of occurrence in the cycle ( Figure 3); hence, they have significant importance and influence on total NH3 emissions in the whole emission cycle. The concentrations of the gases participating in the reactions (1)-(4) should be correlated with each other. However, only the concentration of CH4 from the concentrations of NO, CO, particulate number emissions (PN), and CH4 presented in Figure 4, i.e., all exhaust gas components related to the formation of NH3, is similar to the course of NH3 concentrations ( Figure 2). This observation is confirmed by the values of the correlation coefficients presented in Table 1.  Figure 4, i.e., all exhaust gas components related to the formation of NH 3 , is similar to the course of NH 3 concentrations (Figure 2). This observation is confirmed by the values of the correlation coefficients presented in Table 1.
The correlation coefficient (r) between x and y parameters were calculated in accordance with Formula (5). This is a measure of the relationship between two measured data values.
where x i , y i are the current values of x and y parameters as a function of time in the WHTC cycle, x, y are the mean values of parameters x and y in the WHTC cycle, and n = 18,000 is the number of registered points in the test.  The correlation coefficient (r) between x and y parameters were calculated in accordance with Formula (5). This is a measure of the relationship between two measured data values.
where xi, yi are the current values of x and y parameters as a function of time in the WHTC cycle, ̅ , are the mean values of parameters x and y in the WHTC cycle, and n = 18,000 is the number of registered points in the test.   Small values of the correlation coefficient for NO, CO, and PN (Table 1) indicate that used TWC had proper conversion efficiency for these pollutants, and its efficiency was independent of NH 3 concentration. A much stronger correlation between NH 3 and CH 4 means that TWC had rather poor conversion efficiency for CH 4 , and as a result, a linear relationship can be found between concentrations of these two exhaust gas components.
In SI engines, NO and CH 4 concentrations depend on engine loading, and for this reason, the initial increase of NO and CH 4 concentrations in Figure 4 is due to the increase of the engine loading. It can be observed that the maximum of NO and CH 4 concentrations occur at T EXH = 600-625 • C and then decrease with increasing exhaust gas temperature. This means that the TWC's conversion efficiency and the decrease of NO and CH 4 concentrations caused by it become greater in this temperature than the increase of the concentration caused by the engine loading changes. TWC used in the tests has sufficient efficiency of CO and NMHC oxidation in the whole operating range of the exhaust gas temperature shown in Figure 4. However, CH 4 conversion efficiency apparently depends on the temperature. It may be an effect of the commonly known fact that methane oxidizes poorly in platinum-dominant catalytic converters used in gasoline-powered engines. The effect of exhaust gas temperature on the PN number concentration is difficult to notice.
The strong correlation between NH 3 and CH 4 , especially for CH 4 emitted in the highway phase of the emission cycle, confirms an important role that reaction (4) plays in the creation of conditions for the NH 3 formation. In particular, it means that when the exhaust gas temperature increases (transition from the urban phase to the highway phase), the share of reaction (4) in the NH 4 formation process increases, which proves that the share of reaction (3) in this process is simultaneously decreasing. Figure 5 shows the relative rate of mass emissions (NO, CH 4 , CO, and NH 3 ) and particulate number emissions (PN) in the WHTC cycle. In this figure, it can be seen that each pollutant is emitted with different mass rates expressed as a percentage of the total mass emissions of a pollutant. In the initial phase of the test (urban driving), the relative rate of emission is approximately constant for each pollutant. This part of the WHTC cycle is characterized by relatively low exhaust gas temperatures, relatively low average engine load, and frequent changes in engine loading and speed. In the rural phase of the WHTC cycle (600-1200 s), where the exhaust gas temperature, engine loading, and its speed increase, and the frequency of gear changes and braking decreases, the TWC reactor operates with greater efficiency, which significantly reduces the mass rate of CO emissions. At the same time, CH 4 emissions increase as a result of the poor efficiency of catalytic converter for CH 4 conversion and more exhaust mass flow through the exhaust system. of the engine loading. It can be observed that the maximum of NO and CH4 concentrations occur at TEXH = 600-625 °C and then decrease with increasing exhaust gas temperature. This means that the TWC's conversion efficiency and the decrease of NO and CH4 concentrations caused by it become greater in this temperature than the increase of the concentration caused by the engine loading changes. TWC used in the tests has sufficient efficiency of CO and NMHC oxidation in the whole operating range of the exhaust gas temperature shown in Figure 4. However, CH4 conversion efficiency apparently depends on the temperature. It may be an effect of the commonly known fact that methane oxidizes poorly in platinum-dominant catalytic converters used in gasoline-powered engines. The effect of exhaust gas temperature on the PN number concentration is difficult to notice.
The strong correlation between NH3 and CH4, especially for CH4 emitted in the highway phase of the emission cycle, confirms an important role that reaction (4) plays in the creation of conditions for the NH3 formation. In particular, it means that when the exhaust gas temperature increases (transition from the urban phase to the highway phase), the share of reaction (4) in the NH4 formation process increases, which proves that the share of reaction (3) in this process is simultaneously decreasing. Figure 5 shows the relative rate of mass emissions (NO, CH4, CO, and NH3) and particulate number emissions (PN) in the WHTC cycle. In this figure, it can be seen that each pollutant is emitted with different mass rates expressed as a percentage of the total mass emissions of a pollutant. In the initial phase of the test (urban driving), the relative rate of emission is approximately constant for each pollutant. This part of the WHTC cycle is characterized by relatively low exhaust gas temperatures, relatively low average engine load, and frequent changes in engine loading and speed. In the rural phase of the WHTC cycle (600-1200 s), where the exhaust gas temperature, engine loading, and its speed increase, and the frequency of gear changes and braking decreases, the TWC reactor operates with greater efficiency, which significantly reduces the mass rate of CO emissions. At the same time, CH4 emissions increase as a result of the poor efficiency of catalytic converter for CH4 conversion and more exhaust mass flow through the exhaust system. Throughout the WHTC cycle, the mass rate of CH4 and NH3 emissions are similar to each other. In the rural phase of the WHTC cycle, the rate of particle number (PN) emissions has increased compared to other pollutants. This may be the result of a higher concentration of NH3 in this part of the test cycle and, as a result, acceleration of the reactions (6) and (7), whose product is particulate matter. Comparing the relative mass emission of Throughout the WHTC cycle, the mass rate of CH 4 and NH 3 emissions are similar to each other. In the rural phase of the WHTC cycle, the rate of particle number (PN) emissions has increased compared to other pollutants. This may be the result of a higher concentration of NH 3 in this part of the test cycle and, as a result, acceleration of the reactions (6) and (7), whose product is particulate matter. Comparing the relative mass emission of different pollutants in the rural fragment of the WHTC cycle, it can be observed that the relative mass rate of NH 3 increases faster than the NO and CO emissions. Reactions (1)-(3) in which CO and NO are involved directly or indirectly may be responsible for this phenomenon.
In the final phase of the WHTC cycle (highway driving), the exhaust gas temperature increases to its maximum value. It reduces the NH 3 concentration as an effect of more stable driving. This results in the decrease of the NH 3 emission rate in the final part of the WHTC cycle, and the equalization of the rate of NH 3 , CH 4 , and PN of emissions changes as a consequence. Table 2 shows the masses of emitted pollutants and their specific emissions. Based on the stoichiometry of reactions (1) and (2), it can be determined that 0.56 kg of NH 3 was produced from 1 kg of NO, which means that 1.41 g of NO was used to produce 0.79 g of NH 3 stated in Table 2. Similarly, it can be determined that an additional 3.9 g of CO was consumed in reaction (1) to make 0.79 g of NH 3 . The above stoichiometric   Figures 6 and 7 show cumulative masses of emissions and their ratios in the WHTC cycle determined by dividing the mass of NH 3 emissions by the mass of the selected pollutants. This figure shows how the NH 3 emission is changing in relation to other pollutants during the WHTC cycle. In contrast to the NH 3 /THC emission ratio, which is constant for most of the test, the NH 3 /CO and NH 3 /NO ratios have their visible maxima near the beginning of the highway driving, which is caused by the rapid increase of engine loading after idling.
The conditions for the formation of NH 3 emissions occur when NO and CO are present in the exhaust gas. The concentrations shown in Figure 8 were measured at the TWC outlet of single cylinder row and are not real concentrations of CO and NO in raw exhaust at the inlet to the catalytic converter. In TWC, a significant part of the CO is oxidized and only a part of the CO is involved in the conversion of NO according to the following reactions: If no NH 3 was formed in the TWC, the emissions of NO, NO x , and CO (Table 2) could be greater. It is possible to trace ( Figure 8) the coincidence of CO and NO concentration peaks with NH 3 peaks. As already mentioned, in positive ignition engines, the NH 3 emission depends on the CO emission, which in turn depends on the air-fuel ratio. In a stoichiometric engine equipped with a lambda sensor, the air-fuel ratio oscillates around 1.00 (Figure 9). The frequency and amplitude of these oscillations depend on the characteristics of the fuel control unit. If the changes of the air-fuel ratio take place in a narrower range than shown in Figure 9, lower NH 3 emissions can be expected. different pollutants in the rural fragment of the WHTC cycle, it can be observed that the relative mass rate of NH3 increases faster than the NO and CO emissions. Reactions (1)- (3) in which CO and NO are involved directly or indirectly may be responsible for this phenomenon.
In the final phase of the WHTC cycle (highway driving), the exhaust gas temperature increases to its maximum value. It reduces the NH3 concentration as an effect of more stable driving. This results in the decrease of the NH3 emission rate in the final part of the WHTC cycle, and the equalization of the rate of NH3, CH4, and PN of emissions changes as a consequence. Table 2 shows the masses of emitted pollutants and their specific emissions. Based on the stoichiometry of reactions (1) and (2), it can be determined that 0.56 kg of NH3 was produced from 1 kg of NO, which means that 1.41 g of NO was used to produce 0.79 g of NH3 stated in Table 2. Similarly, it can be determined that an additional 3.9 g of CO was consumed in reaction (1) to make 0.79 g of NH3. The above stoichiometric calculations show that if the engine did not generate NH3, NO and CO emissions would have to increase by 360 and 19 %, respectively.  7 show cumulative masses of emissions and their ratios in the WHTC cycle determined by dividing the mass of NH3 emissions by the mass of the selected pollutants. This figure shows how the NH3 emission is changing in relation to other pollutants during the WHTC cycle. In contrast to the NH3/THC emission ratio, which is constant for most of the test, the NH3/CO and NH3/NO ratios have their visible maxima near the beginning of the highway driving, which is caused by the rapid increase of engine loading after idling.  The conditions for the formation of NH3 emissions occur when NO and CO are present in the exhaust gas. The concentrations shown in Figure 8 were measured at the TWC outlet of single cylinder row and are not real concentrations of CO and NO in raw exhaust at the inlet to the catalytic converter. In TWC, a significant part of the CO is oxidized and only a part of the CO is involved in the conversion of NO according to the following reactions: 2 NO + CO → N2O + CO2 (6) 2 NO + H2 → N2O + H2O (7) If no NH3 was formed in the TWC, the emissions of NO, NOx, and CO (Table 2) could be greater. It is possible to trace ( Figure 8) the coincidence of CO and NO concentration peaks with NH3 peaks. As already mentioned, in positive ignition engines, the NH3 emission depends on the CO emission, which in turn depends on the air-fuel ratio. In a stoichiometric engine equipped with a lambda sensor, the air-fuel ratio oscillates around 1.00 ( Figure 9). The frequency and amplitude of these oscillations depend on the characteristics of the fuel control unit. If the changes of the air-fuel ratio take place in a narrower range than shown in Figure 9, lower NH3 emissions can be expected.  The conditions for the formation of NH3 emissions occur when NO and CO are present in the exhaust gas. The concentrations shown in Figure 8 were measured at the TWC outlet of single cylinder row and are not real concentrations of CO and NO in raw exhaust at the inlet to the catalytic converter. In TWC, a significant part of the CO is oxidized and only a part of the CO is involved in the conversion of NO according to the following reactions: 2 NO + CO → N2O + CO2 (6) 2 NO + H2 → N2O + H2O (7) If no NH3 was formed in the TWC, the emissions of NO, NOx, and CO (Table 2) could be greater. It is possible to trace ( Figure 8) the coincidence of CO and NO concentration peaks with NH3 peaks. As already mentioned, in positive ignition engines, the NH3 emission depends on the CO emission, which in turn depends on the air-fuel ratio. In a stoichiometric engine equipped with a lambda sensor, the air-fuel ratio oscillates around 1.00 ( Figure 9). The frequency and amplitude of these oscillations depend on the characteristics of the fuel control unit. If the changes of the air-fuel ratio take place in a narrower range than shown in Figure 9, lower NH3 emissions can be expected.  Figure 10 shows that, as a result of various changes in the regulation of the engine control system and the exhaust after-treatment system, a significant change in the concentration of NH3 in the exhaust gas was achieved. As predicted, during greater CO emissions, more NH3 is formed due to the increased concentration of free H2 in the exhaust gas. When NH3 concentration is lower, NO emissions increase due to ceasing NO conversion to NH3.  Figure 9. Histogram of relative air-fuel ratio distribution in classes with a width of 0.005. Hot WHTC cycle. Figure 10 shows that, as a result of various changes in the regulation of the engine control system and the exhaust after-treatment system, a significant change in the concentration of NH 3 in the exhaust gas was achieved. As predicted, during greater CO emissions, more NH 3 is formed due to the increased concentration of free H 2 in the exhaust gas. When NH 3 concentration is lower, NO emissions increase due to ceasing NO conversion to NH 3 .
Emissions of particulate mass (PM) and particulate number (PN) decrease with the increase in NH 3 concentration, which may be partly due to the effect of converting ammonia to solid particles by the reaction of NH 3 with nitric and sulfuric acids. Figure 10 shows that the desire to reduce NH 3 emissions may result in an increase in particulate emissions. In the tested engine, it was necessary to use a particulate filter (DPF) as a part of the exhaust after-treatment system. λ Figure 9. Histogram of relative air-fuel ratio distribution in classes with a width of 0.005. Ho WHTC cycle. Figure 10 shows that, as a result of various changes in the regulation of the e control system and the exhaust after-treatment system, a significant change in the co tration of NH3 in the exhaust gas was achieved. As predicted, during greater CO sions, more NH3 is formed due to the increased concentration of free H2 in the ex gas. When NH3 concentration is lower, NO emissions increase due to ceasing NO co sion to NH3. CO Figure 10. Emissions in the WHTC cycle versus mean NH3 concentration in the test.
Emissions of particulate mass (PM) and particulate number (PN) decrease wi increase in NH3 concentration, which may be partly due to the effect of converting a nia to solid particles by the reaction of NH3 with nitric and sulfuric acids. Figure 10 s that the desire to reduce NH3 emissions may result in an increase in particulate emis

Discussion
The conducted tests showed the ability to emit NH 3 by SI engines. The combination of selected factors like TWC temperature, air-fuel ratio, or the type of precious metal coating in the TWC may cause an increase in NH 3 emissions. The tested engine met the Euro VI requirements for all limited exhaust gas components (CO, NOx, NMHC, PM, and PN) in each tested variant, but its NH 3 emission was clearly dependent on the air-fuel regulation and the TWC used. If an engine taken from the tested engine family was used in a light-duty vehicle, it could still be characterized by excessive NH 3 emissions despite passing the type-approval process. The influence of ammonia formation in the TWC on the emissions of particulate matter resulting from its reaction with nitric and sulfuric acids is not sufficiently investigated. It would be interesting to find an answer to the question of how much the fight against NH 3 emissions would reduce the PM emissions at the same time.
NH 3 is a toxic pollutant, the concentration of which in the atmosphere has been systematically increasing for many years [26]. This is mainly the effect of processes taking place in agriculture, but emissions from vehicles also participate in the increasing pollution. In vehicles' engines, the source of NH 3 emissions can be ammonia slip in the diesel's SCR and the processes taking place in the three-way catalyst in SI engines. The article shows that the main way to reduce NH 3 emissions in the SI engines is to reduce CO and NO emissions in the exhaust gas. In particular, this may be a significant problem in gaseous engines. Due to the greater value of hydrogen/carbon ratio in fuels such as CNG or LPG as compared to gasoline, exhaust gases from engines powered by gaseous fuels are characterized by a lower CO concentration and a higher content of H 2 O. Based on reactions (2)-(4), a greater share of H 2 in fuel and H 2 O in the exhaust gas can be considered stimulating factors in NH 3 emissions. As it is commonly known, the TWC technology uses CO for NO conversion. To ensure CO required mass for NO conversion, the range of acceptable air-fuel ratio variability in gaseous engines should be lower than for gasoline engine. For this reason, the gaseous engine will be more liable to increase NO emissions during dynamic driving and rapid changes of vehicle speed and more liable to greater NH 3 emissions. The cases presented in Figure 10 prove that in the engine fueled with propane, high NH 3 emissions occurred only in the case of a simultaneous increase of CO and NO emissions (the invisible increase of NO emissions in Figure 10 was almost completely converted to NH 3 ).

Conclusions
Widespread use of TWC in exhaust after-treatment systems of spark-ignition engines used in light-duty vehicles can result in ammonia formation in the exhaust gas. So far, this pollutant is not limited by international regulation regarding the light-duty vehicle, although emitted NH 3 may increase the content of this component in the ambient air. Currently, a regulation exists of ammonia slip from ships and heavy-duty vehicles, and it is necessary to propose a method to eliminate this problem in other vehicles. Now is the right time to take this issue into account in the future Euro 7 regulation that is being developed.