Recently, more NOx emissions from diesel engines made by a major commercial car company have been a great issue worldwide during all normal driving than in the laboratory or testing station. This violates not only each country’s stringent emission standard, but also may cause a serious health concern and environmental problems (Palash et al., 2013, U.S. EPA, 2018). Although diesel engines have various advantages like fuel economy, high engine efficiency (~52%) and lower emissions of CO/CO2 species, the efforts to decrease diesel NOx emissions have been little successful since the current state of the art three-way catalysts (noble metals) for gasoline engines hardly work particularly under oxygen-rich exhaust gas conditions (Nakatsuji et al., 2008; Toops, et al., 2010). Thus, many investigations into diesel NOx reduction have been conducted first of all using the Selective Catalytic Reduction (SCR) process by NH3/urea or by hydrocarbon agents in search of so called “lean NOx catalysts.” Especially, the fact that NOx could be reduced by hydrocarbons over Cu loaded/metal oxide catalysts has brought much attention over the possibility of diesel NOx reduction, but the technology is still far away from commercial production (Heimrich and Deviney, 1993; Anderson, et al., 1997; Yang, et al., 2011).
In the present work, the alternative Selective Non-Catalytic Reduction (SNCR) process was implemented for diesel NOx reduction under oxidizing exhaust conditions, which has been well practiced as a proven NOx control technology mostly in stationary power plants, incinerators and industrial boilers (Lyon, 1987, Jodal, et al., 1990, Capener, 2008). One major drawback of the SNCR process is the narrow temperature window (1100~1300 K), close to isothermal temperatures preferred, in which NO selectively reacts with reducing agents (NH3 and urea) in the presence of oxygen among combustion products. Another challenge is that diesel engine environments for NOx reduction are so severe, which could include high gas quenching rates (dT/dt= -8000 to -5000 K/sec) from the cylinder to the exhaust pipe, relatively high CO/unreacted hydrocarbons (<0.3% CO, <400 ppm UHC) and a wide range of oxygen levels (3~14%), short residence times and so on. These are all severely different from those in stationary sources like power plants (-200 to -300 K/sec) and waste incinerators (including reactor geometries), which may influence NOx reduction characteristics. Despite practical limitations, several SNCR applications to mobile or stationary diesel engines have been previously reported for NOx reduction. Miyamoto et al.(1995) however conducted to achieve diesel NOx reduction by directly injecting NH3/urea into the actual 4-stroke diesel engine cylinder (796 cm3). But unfortunately too much chemicals (NH3/NOx=16,6) were injected for about 50% NOx reduction at 90° CA ATDC (Crank Angle After Top Dead Center), which dropped quickly to less than 10% beyond 100° CA ATDC (less than 1000 K). Nam and Gibbs(2002) have systematically investigated the application of SNCR for DeNOx of diesel engines through direct injections of NH3/urea into a simulated cylinder (965 cm3 capacity) and exhaust pipe section. Recently, another SNCR experiments were performed under a stationary single cylinder diesel engine (bore~stroke of 87.5~110 mm2) by Vedharaj et al.(2014). Practically, they injected reducing agents into the low temperature exhaust manifold, where only 13.4% NOx reduction was obtained because of lower exhaust temperatures. It is quite evident that NOx reduction potential is absolutely dependent on injection temperatures and their profiles, although many other reaction parameters are simultaneously involved. To overcome such trends of lower NOx reduction at lower exhaust temperatures is practically required, considering severe diesel engine environments. Numerous investigations (Duo, et al., 1990; Weijuan et al., 2009; Yao, et al., 2017) have shown that the addition of various additives such as H2, H2O2, CO, various alkanes and alkenes (CH4, C2H4, C2H6, C3H8), alcohols (CH3OH, C2H5OH) and aldehydes with the primary reducing agents tends to shift the optimum temperature to lower temperatures by 100~200 K. But unfortunately, most SNCR experiments injecting hydrocarbon additives have been conducted under stationary power plant and laboratory conditions, which provide more amenable DeNOx environments than those of diesel engines. Until now, few experiments have been conducted regarding the direct injection of hydrocarbon additives into exhaust pipe conditions for diesel NOx reduction. This paper interestingly presents hydrocarbon-enhanced, DeNOx experimental results by describing the effects of hydrocarbon additives on diesel SNCR process, following previous NOx reduction characteristics obtained in the simulated cylinder and head section (Nam and Gibbs, 2012). The exhaust pipe having a longer residence time (~50 msec) than the cylinder and head section was used as the flow reactor for diesel NOx reduction into which additives were separately or simultaneously injected together with the reducing agent. The effects of hydrocarbon additives on temperature shifts and NOx reduction characteristics are discussed with the possibility of the exhaust pipe and practical implications, based on N/H/O species reactions. Other emission increases and trends of exhaust byproducts caused by hydrocarbon additives are further discussed throughout.
To investigate the effects of hydrocarbon additives on the diesel SNCR process under oxidizing exhaust gas conditions, a diesel-fueled, combustion-driven flow reactor was set up to simulate as closely as possible gas compositions, temperature profiles (gas quenching rates), residence times and aerodynamics of a diesel engine cylinder and exhaust pipe. The detailed experimental system was previously described by Nam and Gibbs(2012), based on a Perkins 4.236 single cylinder diesel engine (965 cm3 capacity), and a schematic diagram is illustrated in Fig. 1. The flow reactor consisted of a combustor, firing diesel fuel and the simulated cylinder (76 mm ID and 213 mm long, stainless steel tube) which was bolted onto an engine cylinder head/exhaust manifold (cut from a real engine) and finally an exhaust pipe of 2 m long tube (70 mm ID). The exhaust pipe insulated with kaowool fibre included 4 sampling ports, which allowed gas samplings at different residence times along the pipe. And the primary reducing agent (NH3) was first directly injected into the inlet of the exhaust pipe (Injector 2), used as the SNCR flow reactor. Then, C2H6 or diesel fuel additive was further introduced simultaneously or separately together with NH3 to investigate the effects of hydrocarbon additives on NOx reduction characteristics both in diesel combustion products and in low temperature exhaust conditions. Three different airflow rates were used such as 1000, 800 and 600 L/min, which corresponded to engine speeds of 3000 rpm, 2000 rpm and 1200 rpm, respectively, based on the above single cylinder (Perkins 4.236). And fuel flow rates normally varied in the range of 0.0832~0.0368 L/min (69~30 g/min) that are equivalent to air/fuel ratios (AFR) of 20~45 (ϕ =0.72~0.32), close to actual diesel operating conditions. At a typical air/fuel ratio of 27.5, the residence time of the exhaust pipe section reached about 50 msec, while the simulated cylinder only provided ~20 msec, varying with air flow rate and temperature. The gas concentration measurements (NOx, N2O, CO, CO2, O2 and UHC) were achieved using the on-line gas analyzing systems, directly connected into the exhaust pipe section. Meanwhile, NH3 slip measurements were separately carried out by a wet chemical method, using a diluted acid solution. NOx concentrations were continuously measured by a chemiluminescent analyser (Rotork Model 440), while other instruments for emission concentrations were described in previous study (Nam and Gibbs, 2012). Typical initial levels of combustion products at the inlet of the exhaust pipe for all 8 flows are presented in Table 1. The initial NOx concentrations for all flows were set to ~530 ppm. For instance, CO/UHC levels (flow #1) were maintained around <482/180 ppm, but these gases continuously oxidized to the end of exhaust pipe, eventually remaining very low, less than 30/4 ppm respectively, while those of flow #6 reached high about 1250/365 ppm due to the lower temperatures. As O2 level increases from 3.7 to 12.8%, CO2 concentration inversely decreases from 11.31 to 6.37% according to air/fuel ratios. Typical temperature profiles along the exhaust pipe section are shown in Fig. 2, where the initial gas temperatures for all flows covering a wide range of engine operating conditions were maintained in the range of 1000~1350 K for possible DeNOx experiments. Particularly, the gas quenching rates (temperature gradients) of 1430 to -3260 K/sec were developed along the exhaust pipe for all flows due to natural cooling. Gas temperatures were measured by shielded thermocouples (type K) and corrected for radiation losses (20~40 K).
3. Results and Discussion
3.1. DeNOx characteristics without additive
The selective noncatalytic reduction (SNCR) of NOx was investigated under diesel combustion conditions, by directly injecting NH3 into the exhaust pipe, used as the SNCR flow reactor. The initial NOx level was set to ~530 ppm, changing the molar ratio (β=NH3/NOx) from 1.0 to 2.0, when air flow rate was normally maintained at 1000 L/min. The effect of temperature on NOx reduction without additives is shown in Fig. 3 in which O2 levels varied from 3.7 to 12.8% at each flow according to the air/fuel ratios. About 50~68% NOx reduction occurred as the molar ratio was changed from 1.0 to 2.0 at the optimum flow #1 (Tin=~1260 K) and O2 level of 6.6%. Temperature windows were formed in the range of 1200~1350 K, not much different from previous laboratory and power plant conditions (Stohr et al., 1997; Quang Dao et al., 2009). NOx reduction tends to decrease with decreasing injection temperatures or with increasing above the optimum flow #1. As molar ratio increases, the extent of NOx reduction at the higher temperatures turned out to be much better than those at the lower temperatures. When temperature further decreased, much closer to the typical exhaust temperature (~1000 K), only 8% NOx reduction was obtained at a molar ratio of 1.5. First of all, the SNCR process proved to be highly dependent on optimum injection temperature even in diesel conditions, particularly requiring uniform and isothermal conditions, which are not available, although many other parameters such as residence time, initial NOx levels and various combustion products were simultaneously involved. On the other hand, the above sufficient diesel combustion products (CO/UHC levels and a wide range of O2, as in Table 1) were not significantly affecting NOx reduction potential in three major reaction components (NH3/NO/O2), implying that the key reaction factor should be still the temperature.
From kinetic calculations (Miller and Bowman, 1989; Nam and Gibbs, 2002; An et al., 2015), NH3 is mostly initiated by active OH radical among many species (OH/O/H/HO2) to produce NH2 (amidogen) intermediate in oxidizing diesel exhaust conditions as follows;
Meanwhile, over the higher temperatures (>1260 K), NH2 radicals rather react with active oxidants (O, OH) than NO to produce NHi and HNO species, which further oxidize and decompose to produce NO, following so called NH3 oxidation paths. So, the competition and balance between NO reduction and NH3 oxidation is maintained throughout, depending on the injection temperature, as in experimental results.
Another big challenge is that the temperature gradients (-2240 K/sec at flow #1, Fig. 2) are also quite substantial both axially and radially along the exhaust pipe, which must be one of the main reasons for low NOx reductions. Ostberg and Dam-Johansen (1994) reported that when the axial temperature gradient was -1000 K/sec in SNCR experiments in a plug flow reactor, it had a significant impact on the temperature window shifting upwards and developed into a little lower NOx reduction. This implies that direct injection of NH3 or other reducing agents into the exhaust pipe would not probably be a suitable DeNOx strategy in mobile or stationary diesel engines, because of the lower temperatures and even severe temperature gradients along the exhaust pipe. Therefore, an alternative strategy by adopting hydrocarbon additives was implemented to investigate possible increase of NOx reduction and further emission trends, particularly in the exhaust temperatures.
3.2. C2H6 additive
C2H6 was injected simultaneously with the primary reducing agent (NH3) into the inlet of the exhaust pipe, using a molar (mass) ratio (ϒ=C2H6/NOx) of 0.2 to 0.5 and 1.0, when the value of β=NH3/NOx was set to a constant value of 1.5.
The influence of C2H6 additive on NOx reduction is shown in Fig. 4. The first interesting thing is that the addition of C2H6 with the range of ϒ=0.2~1.0 mostly shifted the temperature windows downwards. With the molar ratio of ϒ=0.2, the temperature window was shifted downwards by about 50~100 K with little change of shape where the optimum temperature was centered at around 1150~1200 K from 1260 K. However, further increases of ϒ from 0.2 to 0.5 and 1.0 caused further shifts of the temperature windows at higher temperatures (1150~1300 K), but had little effect at lower temperatures (1000~1100 K).
The second, prominent feature is the change of NOx reduction potential caused by additive. At a molar ratio of ϒ=0.2, the degree of NOx reduction is similarly maintained, showing about 60% NOx reduction, compared with those of NH3 alone (no additive). Any further increase of ϒ from 0.2 to0.5~1.0 caused the NOx reduction to diminish significantly up to 40% at the higher temperatures. In the lower exhaust temperatures, more NOx reduction efficiencies (reaching about 15~25%) were surprisingly achieved by the injection of C2H6, which is compared with about 8% NOx reduction potential (NH3 only). According to the previous results (Duo et al., 1990; Yao et al., 2017), C2H6 addition caused temperature windows to be shifted downwards by about 150~ 200 K, mostly under laboratory and oil fired boiler (1MJ/sec) conditions (negligible CO/UHC and <3% O2). The present results show that the shift of temperature windows and the extent of NOx reduction were less effective, due to the severe temperature loss along the exhaust pipe and sufficient diesel combustion products (CO/UHC and O2). Nevertheless, the most encouraging finding is that less amount of C2H6 (ϒ=0.2) gave better NOx reduction potential, reaching almost as good level as with no additive.
The hydrocarbons are ignited and broken into their fragments in the presence of excess O2 than NH3 because of their lower activation energies (40~55 kcal/mole), compared with 62 kcal/mole for the NH3+O2 reaction (Niu et al., 2010; Raj et al., 2011). The addition of hydrocarbons prompts more active radical production even at lower temperatures, hence shifting the temperature window downwards, based on the following general reactions (R4~R6).(R5)
In these reactions, alkyl radicals further decompose mainly via dissociation reactions or via reactions with molecular O2. In case of C2H6, alkyl radicals such as C2H5, C2H3, and HCO are dissociated into highly reactive H and HO2 radicals, which would again lead to further formation of sufficient OH/H/O radicals. The existence of these radicals not only shifts the temperature window downwards, but also accelerates NO reaction with NH3 in the presence of O2 to increase NOx reduction potential, particularly in the lower exhaust temperatures.
3.3. Diesel fuel additive
The injection of diesel fuel as an additive for diesel SNCR process will be much more practical than any other hydrocarbons (CH4, C2H6, C3H8 and CO) investigated, due to the ease of use and transport safety to actual diesel engines.
Therefore, diesel fuel was simultaneously added together with NH3, based on the mass ratio (ϒ=diesel fuel/NOx) from 1.6 to 3.2~5.4. Fig. 5 shows the effects of diesel fuel on NOx reduction as a function of temperature under the same conditions as in Fig. 4. Relatively, at the smaller mass ratio (ϒ=1.6), the temperature window was first shifted downwards by about 100 K, where the optimum temperature was centered at around 1150~1200 K. As the mass ratio increased up to 5.4, the shape of the temperature curves turned out to become more flat, compared with no additive DeNOx characteristics. The degree of NOx reduction at a mass ratio of 1.6 was slightly diminished, achieving about 50% at the optimum temperature. With further increases in the amount of diesel fuel to ϒ=3.2 and 5.4, NOx reduction potentials were more decreased at the higher temperatures. However, diesel fuel proved to be much more effective than C2H6 additive, particularly in the lower exhaust temperatures (1000~1100 K). About 20~35% NOx reductions were achieved, recording about 3 times higher than that of no additive (8%). Mostly, the exhaust section having a longer residence time (~50 msec) depending on engine speeds (air flows) is of little use to achieve extra NOx reduction because of too lower temperatures, compared with those of the cylinder and head section (~20 msec) and higher temperatures, In actual engine cylinders, proper temperatures (1100~1400 K) for NOx reduction can be obtained from the latter half of expansion stroke, close to 70~90° CA ATDC, depending on the power outputs (Miyamoto et al., 1995). Usually, increasing engine speeds not only causes the optimum injection timing advanced (i.e. from 90° to 70° and 50° CA ATDC), but also decreases the residence time, which might be explained in a way for the present lower NOx reduction potentials, compared with the required residence time for enough NOx reduction reaching around 200 msec (Glarborg et al., 1994; Srivastava et al., 2005). Through previous direct injection of NH3 (β=1.5) into the simulated cylinder (Nam and Gibbs, 2012), only 34% NOx reduction was achieved since severe diesel engine environments existed such as large temperature drops (i.e. -7000 K/sec) and the above insufficient residence time. Among them, especially less than 1/3 of NOx reduction potential (9%) was only obtained from the low temperature, exhaust pipe section (~1000 K). Therefore, the above finding suggests that further NOx reduction could be developed by injecting a small amount of diesel fuel as an additive into the exhaust pipe. Diesel fuel would be much more useful than any other hydrocarbons due to the practical reason, where the smallest mass rate of diesel fuel (ϒ=1.6) promised the best performance for the present diesel NOx reduction. This is encouraging for the potential application of the SNCR process to diesel engines, and further investigations may be needed, regarding the optimization of chemicals used and their emission trends.
3.4. Emission Levels of CO, N2O and NO2
Fig. 6(A, B) shows the emissions of CO and N2O before and after injecting C2H6 and diesel fuel as additives for the whole flows (#A to #6). Simultaneous measurements were carried out for these emissions. UHC emission levels were negligible, almost a factor of 10 smaller than CO levels for all flows except #6 and not depicted due to complexity.
First, without additive (NH3 only, β=1.5), maximum N2O emission (~23 ppm) peaks around 1260 K (flow #1), which is consistent with maximum NOx reduction flow. At higher temperatures (1300~1350 K), N2O emissions tend to decrease slightly by decomposition, where less NOx reduction was also achieved. At lower temperatures (1000~1100 K), N2O production sharply decreased to ~8 ppm even at flow #2 (1210 K), despite substantial NOx reduction occurred at the same flow.
When C2H6 was added together with NH3 (Fig. 6A), N2O emissions were almost doubled (45~50 ppm) with increasing C2H6 amounts (ϒ=0.2~1.0) at and above the optimum temperature, despite the adverse effect on the NOx reduction (Fig. 4). However, at the optimum temperature of 1210 K, N2O emissions fall and continue to decrease as the temperature goes down further.
The temperature characteristics of N2O emissions with diesel fuel (Fig. 6B) are quite different from those of C2H6. Injection of diesel fuel triggered temperatures a little high by about 20~40℃, depending on the mass ratios (diesel/NOx), so that less N2O production occurred at the higher temperatures due to decomposition reaction. However, at lower temperatures, unlike C2H6, a substantial amount of N2O (20~65 ppm at 1100K) was still produced, depending on the mass ratios of diesel fuel.
Secondly, the CO emissions with no additives at the higher temperatures (1200~1350 K) usually remained quite low at the end of exhaust pipe because of very efficient combustion (Fig. 6A). As temperatures go down, the CO emission levels become significant, and increased up to ~500 ppm at flow #4 (1100 K), and again rose up to about 1600 ppm due to the low combustion temperatures (1000 K). With C2H6 injection at the molar ratios (ϒ =C2H6/NOx) of 0.2 and 0.5, CO emissions were slightly increased by about 100~300 ppm at the lower temperature flows, but at the higher temperatures, the extent of increase was negligible. As diesel fuel was injected, two different aspects emerged according to the temperature (Fig. 6B). At higher temperatures, CO emissions were further decreased by additional combustion. However, at the lower temperatures (1000~1100 K of flow #6~#4), CO emission levels were substantially increased by almost two or three times those of no additive because of incomplete combustion and higher carbon content of diesel fuel, compared to C2H6.
Table 2 shows a comparison of “carbon content” between C2H6 and diesel fuel for the injected quantities. The carbon content of diesel fuel was 6~9 times higher than C2H6 according to the mass ratios. Therefore, more CO and N2O emissions were observed with diesel fuel, although more NOx reductions were achieved at the same conditions.
Fig. 7 shows NO2 increase, resulting from the NO/NH3/O2 reactions with/without the injection of C2H6 and diesel fuel, as observed in Fig. 6. At higher temperatures (1150~1200 K), actually lower parts of temperature windows for NOx reduction, the NO2 emissions were less than 2% (<10 ppm), and not affected by either existing combustion products (CO/UHC) or by further injecting C2H6/diesel fuel additives. However, as temperature was lowered to 1040~1000 K (flows #5 and #6), the NO2 emissions sharply increased and reached ~50 ppm at 1000 K without additives. Injecting C2H6 or diesel fuel caused the NO2 emissions to increase substantially from ~50 ppm to about 80~95 ppm at Tin=1000 K where for diesel fuel (ϒ=3.2), more NO2 was produced due to its higher carbon content.
The injection of NH3 for NO reduction retards the main CO oxidation reaction rate (R7), resulting in the increase of CO concentration because many chain reactions compete for OH radicals, and particularly the reaction (R1) is very rapid, compared with reaction (R7). The major sensitive reactions for N2O formation are (R8) and (R9), while the conversion of NO to NO2 is mainly dominated by the reaction R10 at the lower temperatures. Increasing C2H6 and diesel fuel additives would increase the concentrations of not only active radicals (OH/O/HO2/H), leading to further NOx reduction particularly at lower temperatures, but also of nitro-carbonaceous species (NCO, NH, HCN and CN) and hydrocarbon fragments (CxHy), which then eventually contribute to the increase of N2O and NO2 byproducts and CO emissions as well through complex chain reactions.
The effects of hydrocarbon additives for diesel NOx reduction were investigated using the SNCR process under oxidizing diesel exhaust gas conditions. A diesel-fueled combustion system was set up to simulate the actual cylinder and head, exhaust pipe and combustion products, where the reducing agent NH3 and C2H6/diesel fuel additives were separately or simultaneously injected into the exhaust pipe, used as the SNCR flow reactor. With NH3 alone, temperature windows were formed in the range of 1200~1350 K, where about 50~68% NOx reduction was possible with the molar ratio (β) of 1.0~2.0 at the optimum flow #1 (Tin=1260 K). As temperature goes down to the typical exhaust temperatures (~1000 K), only 8% NOx reduction was obtained because of too lower temperatures. More encouragingly, together with additives of 0.2 (C2H6/ NOx) and 1.6(diesel/NOx), temperature windows were shifted downwards by about 50~100K from the optimum temperatures with little change in NOx reduction potential. Diesel fuel turned out to be much more effective than C2H6 especially at lower temperatures (Fig. 5) in which about 20~35% NOx reductions were maintained, recording about 3 times higher than that of no additive (8%). However, the addition of hydrocarbons increased emission levels of byproducts such as CO, UHC, N2O and NO2. At and above the optimum temperature (1260~1300 K of flow #1~#B), the increase in N2O emissions was substantial from ~23 ppm (no additives) to ~45 ppm for both additives, but other species emissions (CO, UHC and NO2) remained surprisingly very low because there were relatively enough temperatures for oxidation. However, at the lower temperatures (1000~1100K of flow #6~#4), diesel fuel produced higher CO, UHC, N2O and NO2 emissions than C2H6 because of the higher carbon content of diesel fuel (6~9 times).
After all, the injection of C2H6 and diesel fuel into the exhaust pipe is believed to have promoted the production of OH/O/HO2/H radicals in N/H/O/ hydrocarbon reactions which certainly contributed to further NOx reduction at lower temperatures. In addition, the existence of higher CO/UHC levels would accelerate the rate of NOx reduction in short residence times (~50 msec), but decrease the selectivity for the reaction of NH3 with NOx, resulting in overall slightly lower NOx reduction, and contribute to higher CO/N2O/NO2 emission levels.
Nevertheless, Injecting a small amount of diesel fuel together with a primary reducing agent (NH3) seems to be more encouraging for practical reason and could be suggested as an alternative SNCR DeNOx strategy under diesel exhaust systems, following further strict optimization of chemicals used for lower emission levels of byproducts.