CHAPTER 2 Genesis of Engine Emissions 2.1 COMBUSTION GENERATED AND OTHER ENGINE EMISSIONS The carbon monoxide (CO), unburned hydrocarbons (HC), nitrogen oxides (NO,), particulate matter (PM) and aldehydes emitted from the vehicles puwered by intemal combustion engines are the main source of urban air pollution. Except a part of unburned hydrocarbons, all these pollutants are generated during combustion in the engine and are emitted through the engine exhaust gas. The CO, HC and aldehydes result from incomplete combustion of fuel while the nitrogen oxides and sulphur di- and t oxides when fuel contains some sulphur, are the products of complete combustion. Amount of sulphur in the current engine fuels; gasoline and diesel, is quite small (mostly < 50 ppm by mass). Therefore, emissions of sulphur oxides from vehicle engines are not significant, The spark-ignition (SI), gasoline engines are the main contributors to CO, HC and NO, emissions while from compression ignition (CI), diesel engines the emissions of NO,, soot and particulate matter are of major concern, The amount of pollutants emitted depends on several factors such as engine design, operating conditions, ambient conditions, fuel type and the exhaust after-treatment employed. Depending upon the operating conditions, a gasoline fueled, four-stroke, SI engine without exhaust treatment typically emits 0.2 to 5.0 % CO, 300 to 6000 (C;) ppm HC and up to 2000 -3000 ppm NO,. These concentrations are for Taw, undiluted exhaust gas on volume basis. Under engine idle, CO and HC emissions are high due to engine operation at rich-mixtures. During engine warm-up and under engine transients like acceleration, HC emissions are again in the higher range. In addition to exhaust emissions, unbumed fuel or hydrocarbons from evaporation in the fuel tank, fuel system and carburetor in older vehicle designs also escape to the environment from the conventional gasoline fueled engines. Further, HC are also emitted through crankcase blow by gases in the homogeneous, four-stroke, spark-ignition (SI) engines. The crankcase blowby and evaporative emissions from uncontrolled gasoline fucled vehicles contribute about 40% of total HC emissions, each source accounting for nearly 20%. In the diesel engines, carbon Monoxide emissions are quite low (0.02 to 0.1 % by volume). The exhaust hydrocarbon emissions from diesel engine are lower by a factor of 5 to 10 compared to typical SI engine levels and, the evaporative and crankcase HC emissions are practically absent, The nitrogen oxide concentrations in the exhaust gas however, are comparable to those from the typical SI engines. The mechanism of formation of engine exhaust emissions is governed by the comb chemistry of combustion. The sources of exhaust emissions in conventional spark ignition engine are shown schematically in Fig. 2.1. After the spark ignites the compressed fuel-air mixture, flame propagates Across the combustion chamber and burns the charge in the engine cylinder, Nitric oxide and carbon Monoxide are formed in the bulk of burned gases by chemical reactions during combustion and post combustion period. Nitric oxide (NO) is formed in the high temperature burned gases in the engine cylinder primarily through oxidation of nitrogen present in the intake air, Nitrogen oxides emissions20 Engine Emissions yen . Carbon monoxide (CO) is Primary, amounts of nitrogen dioxide (NO). C eo thet int ctaeney a onygen tn the foel-ae Jnmete, formation of iol ener freezin a ae deficiency of esa the eee he conic Products and fees at ‘ ae fuel ratio eo 0 a ; Soar tut ti cel a aa casey nate enaleed Ik NO tad the chemical equilibriur engine warm-up and traniens f toch higher eles than srs being vch during cod ‘Significantly to CO emiesions ae faring Wie ons CouUeIStaes ea saa ; tion du s acceleration, engine opera Figure 2.1 ine cyliné ‘unburned hy Pressure inside the engine decreases during expansion stroke, the unbumed mixture from the crevices expand in he 6 a jn the? fee it B18 entrained into the hor bulk gases, Sim, ly, the hydrocarbons adsor desorbed ® film and combustion chamber i mpression strokes are also sentra Partly entrained by the bulk gases, A large fracti umned hydrocarbons that are re- beret EAS inside combust luring expansion before the exhaust pase leave the engine. Fi Products along with CO, NO, crevices leave the ‘dif xii Stroke is partially or completely © when the exh and part of the entrain, engine and are bulk of pases and j mbt ‘aust valve(s) opens, the CO! ted unburned hydro usted. From the above, | involve combustion from quench 1%" ; Seen that NO and CO are form and chemical ki isi? netics. On the other hand, HC etGenesis of Engine Emissions 21 result from flame quenching at the combustion chamber walls and in the crevices, and from adsorption and desorption in lubricating oil film and combustion chamber deposits. HC emission mechanism involves boundary layer acrodynamics, mixing with bulk gases and post-flame oxidation. Air-fuel ratio is one of the most important parameters that influence the engine exhaust emissions. Typical variations of exhaust emissions with air-fuel ratio in a SI engines are shown in Fig. 2.2. Nitric oxide emissions are maximum at slightly (5-10 %) leaner than stoichiometric mixture due to combination of high combustion temperatures and availability of excess oxygen at this point. Carbon monoxide and HC emissions reduce as the air-fuel ratio increases due to availability of more oxygen for combustion. Use of a high air —fuel ratio tends to reduce all the three pollutants. The SI engines are normally operated close to stoichiometric mixture condition to provide smooth engine operation. Leaning of mixture beyond a critical air- fuel ratio that depends to some extent on engine design and operating conditions, results in poor quality of combustion and eventually in engine misfiring causing an erratic engine operation and sharp increase in HC emissions. During engine cold starting and warm up duc to poor fuel vaporization rich fuel-air mixtures are required for smooth engine operation resulting in high CO and HC emissions. Normally, one would like to operate engine on lean fuel-air mixtures that would give low CO and HC with moderate NO, emissions, and also low fuel consumption. But, presently most SI engines are operated close to stoichiometric conditions for an efficient catalytic control of NOx emissions, as discussed in Chapter 4. . 6000 6 e 5 € & 5000 co 3 Nox §& 4000 4 a & — 3000 3 § S 2000 2 5 2 1000 He / eS O07 (08: 09 1 1 12 Fuel-Air Equivalence Ratio Figure 2.2 Variation of HC, CO, and NO concentration in the exhaust of a conventional spark ignition engine with fuel/air equivalence ratio, [(actual fueVair ratio)/(stoichiometric fuel-air ratio)]. In ion ignition (diesel) engine, fuel is injected into the hot, compressed air inside the oie aE Er ey crake A nonuniform furl cstibson ta vanes wth time and Space exists in the engine cylinder. A non-homogeneous fuel-air mixture prevails throughout the entire Combustion period. Mixing of fuel and air is controlled by several parameters related to injection, air motion, turbulence and fuel evaporation. To control engine power output, the quantity of injected fuel is7 on in the eng, ng Parameters, varies with tr + ‘premixed? illustrated (Fig, 2.3) schematically ho rmation of NO, HC and soot Particulate 22 Engine Emissions i istributi i i i eyele ii ly unchanged. The state of fuel dist ied while the air quantity per eycle is relatively ; i Sia 7 ; ee aa in engine cycle varies with engine es a and oon 7 ion of i yi d by the local fuel-air ratio, Formation of pollutants is strongly influences b 0 during combustion. The diesel combustion process is composed of two main phases: “mixing controlled’ combustion. Heywood [1] ‘Sh engine loads from the fuel over-rich regions, The mechanism of formation of the mer main engine pollutants, CO, NO,, HC and Particulates is Giscussed in more details in the sections that follow. HC from lean flame blow-out Rich mixture region in fuel jet NO forms in bumed gas, Premixed Combustion Soot Oxidation in white/yellow flame Soot formation in tich zone in fuel jot Fuel vapor M Nozzle Sac Volume , = mpression-ignition engine durist ei| Genesis of Engine Emissions 23 2.2 CARBON MONOXIDE Carbon monoxide emissions result due to deficiency of oxygen during combustion of fuel rich mixtures. A bwostep process may approximate complete combustion of hydrocarbon fuel to form finally the carbon dioxide, Step 1: Conversion of hydrocarbons to CO. During this step, several oxidation reactions occur involving formation of intermediate species like smaller hydrocarbon molecules, aldehydes and ketones. Oxidation of a hydrocarbon, RH where R is the hydrocarbon radical may be represented by the following chain reaction RH > R R+0;>RCHO > RCO >CO @1) Step 2: Conversion of CO to CO; provided sufficient oxygen is available. One of the principal reactions for conversion of CO to CO; is, CO+OH + CO,+H (2.2) During combustion and expansion in the engine cycle until the temperatures decrease to about 1800 K, the C-O-H system is more or less under thermodynamic chemical equilibrium. Once the temperatures during expansion drop below 1800 K, the reactions in C-O-H system become kinetically controlled. The mate controlling reactions for oxidation of CO to CO, especially at temperatures lower than! 800 K include the following three body recombination reactions (3), H+H+MeH.+M (2.3) H+OH+MeH,0+M (2.4) H+0;+MHO;4+M @5) The water-gas reaction (2.2) is quite fast and at high temperatures is continuously under equilibrium. Only Iate into expansion stroke conversion of CO to CO; is kinetically controlled and takes place at a slower rate. Newhall [2] studied kinetics of CO formation and established that the CO emitted is higher than the equilibrium concentrations corresponding to the temperature and pressure conditions at the end of expansion. The non-equilibrium rate controlled concentrations of H and OH from the reactions (2.3) to (2.5) were used for prediction of equilibrium CO concentrations by the reaction (2.2). The calculated CO termed as partial-equilibrium values are compared with the measured and equilibrium CO on Figure 2.4, ‘The equilibrium concentrations were calculated corresponding to peak cycle temperatures and pressures, and corresponding to exhaust conditions. The measured CO data from several sources at different air-fuel tatios, different fuel composition and fuel volatility were used. Measured data are correlated by a single curve. The CO increases sharply as the air-fuel ratio decreases below stoichiometric condition. For lean mixtures, CO concentration is very small, less than 0.1%. The partial equilibrium calculations gave good Correlation with the experimental values fo For the rich mixtures, experimental values are quite close to the equilibrium concentrations at peak pressure conditions. For the Stoichiometric and lean mixtures however, partial equilibrium CO predictions have a better agreement ‘ith the equilibrium concentrations corresponding to exhaust conditions. Measured carbon monoxide for leaner than stoichiometric mixtures are higher than the partial equilibrium calculation results. One factor that would have significant contribution to CO at lean mixture conditions is partial oxidation of hy \ydrocarbons entrained into burned gases from the crevices during ‘expansion.24 Engine Emissions Calculated equilibrium ~ at peak pressure Calculated partial ‘equilibrium and measured data Carbon Monoxide, % vol. @ 6 44 Calculated 24 Equilibrium at Exhaust eS 15 ao ae 1 09 08 Fuel-Air Equivalence Ratio i bution may not be entirely homogeneo, between ¢ 5 Secondly, fuel. Foal aton with the intake wn ©, POT fuel vapor; the aircfuey i 8 from th © internal combustion-engines ignition e wou ld ‘ngines, whil tial ines, ubstant 8 8S Woul gines, while si "cussed later, NO is formedGenesis of Engine Emissions 25 Thermal NO is the dominant source of nitrogen oxides in IC engines. NO is formed in the high temperature bumed gases behind the flame front. The rate of formation of NO increases exponentially with the bumed gas temperature, although it is slower compared to the overall rate of combustion. Mechanism of formation of thermal NO is discussed in detail later in this section. Fuel NO is formed by combustion of fueis with chemically bound nitrogen. The route of fuel nitrogen to form NO is proposed to be via the reactions of fuel nitrogen to first produce intermediate nitrogen containing compounds and reactive radicals like HCN, NHs, CN, NH etc. These species are subsequently oxidized to NO by the oxygen or oxygen containing species. Although petroleum crude may contain nitrogen of up to 0.6 % by mass but most of it is concentrated in the heavy, high boiling fractions. Gasoline has negligible nitrogen. Diesel fuels have higher nitrogen content, but this too is generally less than 0.1% by mass. Fuel nitrogen for gasoline, diesel and natural gas is not considered significant for contribution to NO formation. Prompt NO: Fennimore [3] found that NO is rapidly formed in the flame front and its amount is ignificant. He termed it as prompt NO. The prompt NO is formed in the flame by reaction of intermediate chemical species of CN group with O and OH radicals. The hydrocarbon radicals CH, CH, C, C; ete. formed in the flame front react with molecular nitrogen to give CN group intermediates like HCN and CN by the reactions (2.6) to (2.8). Bowman [4] observed large concentrations of HCN near the reaction zone of fuel rich flames. A rapid formation of NO was associated with rapid decay of HCN. CH+N; GHCN+N 2.6) CH, + Nz «+ HCN + NH Q7) C#HN, «+ CNEN (2.8) The contribution of prompt NO in the stoichiometric laminar flames has been estimated to be about 5 to 10 percent only [5]. In the engines, as the combustion occurs at high pressures the thickness of flame front is very small (~ 0.1 mm) and the residence time of chemical species in this zone is very short. Moreover, the burned gases produced by the charge elements that burn early during the combustion Process are compressed to a much higher temperature than the temperatures attained immediately after combustion. The formation of thermal NO in the bumed gases behind the flame front therefore, is much higher compared to any NO formation in the flame front. However, contribution of prompt NO becomes significant under the conditions of engine operation with high dilution such as lean burn operation and use of exhaust gas recirculation [6]. In diesel engines, contribution of this mechanism could be significant in regions of rich mixtures close to fuel spray. i 23.1 Kinetics and Modeling of Thermal NO Formation The following three principal reactions govern the formation of thermal NO, and constitute what is commonly referred to as the extended Ze/dovich mechanism. % (2.9) +N, t NO+N @9) ky (2.10) N+0, 8 NO+O 2.10) N+oH 8 NO+H (2.11) 328 Lagive tmiainns 2 e original Zeldovich mechanisn consist avse al [Ff bho ae tn wget UD othe orginal Zeklowieh ettanlac tint es a S) arnt LOK The fromant part of the be a Pane wwvotowuler mitragen Dy atonnde aaygonr Detinnd the Mame, This rea NW . \ 0 formation eteRQY of about TSO Row mrad and ts a rate: determining reac yah in ‘ peace selected by Bosman annt Raglan [N] thou literature ity amnits of emtimates s el See Nemathn of NO using the three Rowing equations ing af, NPOsition ¢g high activa, The rate co ature, T in Kay waretions (2.9) t© GUD) can be expressed by th {NO} = & LON = OTN *RINTOL] = ks [NO][O] os *RUNTON] =k o [NOTH] 2.12) Nese Ie arty are rate oonstants fhe the forward reactions 2.9), 2.10) and (211), Respectively: subscript (.) dcnvees the rate constants fOr the mene reactions and_[NO}, [0], [OH] ete, denote tk Soncentration of species within [ ] int moles: em’, Table 21 Reaction rate constants for NO formation mechanism, emimoks, Tink fo) Reachon Foowant Reverse, O-N: NON Ke LSS TOFS ap Ry RQ=3y NUE exp CGD) ess 1O"T Ss exp (& 4e807T) Aye 721 19! KS= 38x TS exp QaoT) LAUT 10" X ex “exp 4807) CSOT) n and radicals ON, H and OH are tobe estimate S Very small and the con Nentration of Ny remains nearly constant, Two Approximation of [N}: Concent Pration oP atomic N is rather smal of the ox ce only compared to other acting chemical species, Rate of firmatioe is small Telative to its Soncentration, and FO, OH, 0; and O are quilibrated, {NO}Genesis of Engine Emissions 27 d Gy NOV = 2 the CO} — k 4 NOJIN]) (2.14) from Equation 2.13, k, mm. -*& LOIN 2 14-k.y [NOI] +k = (NO}[H) (2.15) 8s ky INO] +k, [0,]+k, [OH] Using reactions 2.10 and 2.11, from the assumption of equilibration of O, OH, H and Oz, O.H)k \ 2101 ka ky (2.16) loyfon] kk, Eliminating [N) and (H) using equations 2.15 and 2.16, the equation 2.14 gives, @ oe 2 tno} = 2k, toy ,j NOP M10, 1085) in at 1+k_[NOV(k,[0,]+ LOH) where K'= (ky/k 1)(ka/k .) is equilibrium constant for the reaction Nj +O,++2NO, The NO formation rates may be calculated by Eq, 2.17 using concentrations of ©, Os, OH and Ns. REE oF NO formation is much slower than the combustion rate. Most of NO formation takes place in the post flame bumed gases after combustion is completed locally. The NO formation procese ig decoupled from combustion process and rate of formation of NO can be calculated assuming equilibrium values of temperature and concentrations of O, O2, OH and Np Bs 'y introducing equilibrium assumption in the calculations, the Eq. 2.17 is further simplified by using the following notations; Ry= ky [O]- [Nae = k 4 [NO]-[N]e Where R, is the reaction rate using equilibrium concentrations for the reaction (2.9). Similarly, Ro = kp [N]e[Oz]e = k-2[NO}-[O]., and Rok INK LOM -& SNOW, ‘Using the above notations the Eq. 3.17 is simplified to give rate of formation of NO as below, 2R, (1 (NOVINO 6)" ei “(+ wINOVINO} ¢)28 Engine Emissions 2.3.2 Dependence of NO Formation on Temperature and [02] Let us consider initial rate of NO formation when [NOJ/[NO].<< 1. From Eq, 2.18; et us cons No} 2 [OLINsl dt Qy The soncentation of atomic oxy at equilibrium, using reaction %4 Oa ++ O is given by, K oylO,}” nore (le ny? 2% and, equilibrium constant Kyo, is given by K y@) = 3.610" exp] ) (225 Using value of k.y and the Eq. 2.19, 2.20 and 2.2 |, the initial rate of NO formation is determined by, {NO} _ Spe 222) 113; (0, J. “(N2]. , mol/em?-s QZ) Temperature being in exponential term in Eq.2.22, it st i 222 it follows that the NO formation is maximized una sochereeneentrations, These conditions oceur at fuelan equivalence ratios 5-10% leaner thm Stoichiometric mixture, ted below by an example: reach equilibrium concentration of NO may 223) +N2 + 2NO, as va ‘INO}, -(« NO*Os1 IN, } + Mole fraction Using equilibrium constant fy) = 203xexp(-21,650/7) ang Fert 3.38 x 107% yp « 338210 Tep(sh35/1) 25)Genesis of Engine Famissions 29 time fyo is in seconds, pressure, P is in atmospheres and temperature, Tin in K. For the charge that burs carly in the cycle the peak burned suas temperatures of 2700 K or higher could be obtained. Using cylinder pressure of 35 atm and = 2700 K, kinetically formed NO would reach equilibrium concentrations quite rapidly in about 0,38 ms, At an engine speed of 3000 rpm, this period is equal to 6.8 degrees CA, which is significantly shorter than the typical total combustion duration. On the other hand for the charge buming Inte, maximum temperature may reach 2300 K or lower and also would be low. Even if this charge after combustion is held at 2300 K and 25 atm to reach equilibrium NO concentration it would take 18.3 ms, This period equals to 330 degrees CA at 3000 rpm, which is too long a period compared to the combustion event and the period during, which high bumed gas temperatures prevail in the engine cycle. The NO concentrations reached kinetically in the charge that ‘burs late would hence, be much lower than the equilibrium concentrations. 23.3 NO Formation in SI Engines Unmixed Combustion Model of Thermal NO Formation For estimation of NO formation in SI engines using Eq. 2.18, data on thermodynamic state and equilibrium composition of the burned gases must be known, Several methods are used to calculate temperature of bumed gases for a given engine geometry , which include (i) using measured cylinder pressure - time history and thermodynamic model (ji) use of empirical burn rates, and (iii) fundamental combustion models that are based on one-dimensional flame propagation models or multi-dimensional combustion models. In one-zone combustion model, after combustion the post flame bumed gases are assumed to mix instantaneously with the gases burned earlier so that all the burned gas at a given instant is uniform in composition and temperature. This is commonly referred to as the fully mixed model. ‘Another combustion model at the extreme is an unmixed multi-zone model where no mixing occurs between the mixture clements that bum at different instants in the cycle, In the unmixed model, each bumed gas element maintains its separate identity and undergoes isentropic compression and expansion 2s the cylinder pressure changes. The elements that burn early get compressed to peak pressure after combustion and reach higher temperatures than those burning later. The unmixed model predicts that a temperature gradient exists in the burned gases, the early-burn elements being at higher temperatures. Eniatence of temperature gradient in the engine combustion chamber have been observed, the difference in temperatures of an early-burn element (ncar spark plug) with a late- burn element at the far end of the Combustion chamber being as high as 400 K [9]. Actual situation would be somewhere between the filly nized and unmised models as some mixing among the different burned gas elements would occur. The tumized model however, is more realistic. It has been shown'that use of 5 to 10 different unmixed burned ga clements predict reasonably well the realistic temperature and composition distribution in the ‘combustion chamber (10). Using umixed model, Komiyama et al (11] have studied formation of NO in the different charge elements, which burn sequentially us the combustion progresses. Figure 2.5 (a) shows measured cylinder pressure and calculated mass fraction burned for an engine. The temperatures of unburned mixture and bumed gases for the charge elements burning at the beginning and at the end of combustion are shown on Fig. 2.506), Using unmixed model of combustion and kinetic model of NO formation (Eq, 2.18), NO conceotrations are calculated as a function of crank angle in these two elements and are shown in Fig 2.5(c). The equilibrium concentrations of NO for the two elements are also shown, In the early-burn. 2s the peak temperature is reached duc t0 compression to peak cylinder pressure; the kinetically30 Engine Emissions er, as the temperature s, to near equilibrium a ie eontrotied by the back . noled NO teaches 10 ee comp freezes due to expansion. In the yt formed i., the rate co NO starts decomposing 11 TG Sie decomposition reactions g fling due to enpansion NO stars Kinet jue to expansion even se Se tel Cee Ee NO ication Fen i no ve sgn ne ‘burning elements, tion reactions, Rate cont uilibrium concentratior of magnitude lower than jn thy to dominate epee well above the ea eons on pan eat 1400 KN formed in the bum mg Ce bin is ice in a ers ae Sens NO with highest concentrations e ‘pode aves bon obsevedexpesimentally 9) jel have als Equilibrium x= 0.0 EEE Menu SUTE and cate temperature of the unbume slated b ‘charge elements (c) NO concen rm ; MEE, Ty an rations in th one element that bums finer : 6) ned et°M, X a8 a function of erank angle Wo el lements. ed gas, Ty for an early first and late ey i and another by, ing last ip enstion of crank angle degreeGenesis of fingine Emissions 31 Mans fraction of NO in the exhaust can be caleulated by summing up of the frozen mass NO fraction yvor all the burned ysis elements ws below ! Fo” |M noe (2.26) where, (vay Ws the final frozen NO nase fraction in the element of charge which burned when mass frnction burned wax x. The (rua) {NOJMyr/y. snd Myy is molecular mass of NO, pis the density of bumed element, The average NO concentration in extust as mole fraction, Juois given by /M Ino = Buon (2.21) 2.34 NO Formation in Cl Engines In the compression ignition engines, rapid combustion in pre-mixed phase is followed by diffusion combustion process, The rate at which fuel and air are mixed controls the diffusion combustion process. ‘The reaction kinctics leading to formation of nitrogen oxides and other assumptions of equilibration of C- 1-0 nystem used in the spark ignition enginies also apply to the CI engines. As the fuel is injected in the hot compressed air, the fuel spray entrains air and non-uniform fuel distribution exists in the combustion chamber, Pquivalence ratio varies widely from very rich at the core of spray to very lean at the spray boundaries. A fuel spray injected radially outvard in swirling air is shown schematically in Fig. 2.6. As the sir is entrained into spray and spray slows dovmn it gets deflected more and more towards the direction of swirl. The shape of fucl-air equivalence ratio distribution in the jet is also shown on the spray. Spray core contains moxtly liquid fuel and very rich mixture exists in the vicinity of spray core. Large regions containing fuel vapour downstream of the spray core are obtained. Ignition occurs in the slightly leaner than eoichiometric mixture region downstream of the spray core where fuel has spent most time within the flanenable limits, After the ignition delay, during premixed phase of combustion, fuel air mixture within flammable region and around stoichiometric composition buns spontancously. In the mixing controlled phave, combustion is believed to occur in those regions of spray where equivalence ratio is dove to unity. Thus, NO is formed at varying, rates depending upon the local equivalence ratio and temperature, As the combustion progrestes, the already burned gases keep on mixing with colder air and fuel vapour changing. its composition and temperature, Temperature of the reacting gases also changes ve 10 compression and expansion. Conceptual Models In the classical spray combustion models, formation of NO starts in the bumed gases produced during premixed combustion phase when clove to stoichiometric and lean flammable mixtures burn. However, Dec {12} in his combustion model suggests that premixed combustion occurs in rich mixtures, with = 2 to 4, In such rich mixtures, neither thermal NO nor prompt NO is likely to form in significant concentrations, Adiabatic fame temperatures are ~ 1600 K and very little oxygen is available for thermal NO Kinetics to be significant. Alo, very little prompt NO is formed for mixtures with > > LA, Dee suggests that most NO is formed in mixing controlled diffusion combustion at spray boundaries 404 in the post combustion high temperature yaves, The diffusion combustion takes place at near stoichiometric conditions. However, other investigators {7} propose that in the naturally aspirated engines32. Engine Emissions ° ir, the cContributi ixing of fuel and air, . fi mixing 0! Sad ind . \d sufficient time available for ete d gases from premixed combustion, " with long ignition delays do eacastion is saksankial’ The bal d stion to a ‘xed combustion premixe a ressure and ence ation rate: h, n rates are high, |O formati het pressure and temperature and hence Nt compressed to a higher press , Nozzle hole ~ Re’ ‘Spray tail Trailing Edge Core 8 Air Swirl Spray leading edge Flame blowout region Figure 2.6 Schematic ofa fuel spray injected in Swirling air. Fuel-air equivalence ratio (0) distribution spray is shown, © formation is more Significant. Firstly, during ths Phase the mixtures that are lean or close to Stoichiometric get burned, Secondly, the delay period is ie and a larger fraction of fuel bums as premixed. In the turbo-charged, quiescent combustion clan Sreints, the ignition delay i short and the e) xi ignition is smatier. Thus, a relatively Halet fraction of fuel burns durin, Moreover, with Tetarded injection timing and high injection pressures in the turbo-charged engines sa initial small posiee the premixed gmhion ise, i rest being rich (8), Therefore, overall a Significantly smatier fraction of fuel burs turbocharged, bigest ehtasd cngines samp ared to the nat Y aspirated engines, In the modem is produced in the eyete ane pars pate “ngines with retarded injection timing, more than half of NO: roe es decrease due tg “xPansion after peak pressure. In freezing of NO kinetics he come neal due 10 mixing oa cooler ait and conse! engines on mixing with conte fF and fuel sir igen tONS, Thu, vid nin Of bumed gases in dies! resus in NO concentation ones Higher levels comm ettseS More i jes, whic! TPid freezing of Nj Kinetics, whic! Cylinder averag history for eavy-duty mPared to that in sy engines" NO . With MBT injection timing wa Studied experimen tly aspira tect injection engine operating evacasnTied out by a method op it °XP ¥ Vioculesey evacuated tank at a pre ntire rine i ‘srman [13]. The experime! sntite cylinder 7 aaa angle. A gj 8888 by dis them in termined oP brag at Sealed the engine ene from ue © sudden and rapid maka or egilinder Bases to be suddenly Sracuated tank was ruptured at tre Pred discharged into the evacuated tank. ThGenesis of Engine Emissions 33 quenched NO reactions. The results of this experiment are shown in Fig 2.7. The data on three engine speeds, 850, 1000 and 1200 rpm were observed to fall along the same curve. Ratio of cylinder average NO to exhaust NO concentration reaches maximum value shortly after top dead center before peak pressure occurs. The decomposition reactions are not prominent as the NO from peak reduces only slightly by just about 10 % and tends to be frozen rapidly at a higher level. It is further, observed that most of the NO forms within 20° CA from the start of combustion. These experiments support the hypotheses that freezing of NO kinetics and NO decomposition reactions results in NO levels to be frozen at higher levels in the diesel engines as the bumed gases are more likely to mix with cooler air and ‘mixture in the diesel engines than in the SI engines. 14 12 1 08 06 NOMNO}exh 04 02 o 30-20 -10 0 10 2 30 40 Crank Angle, ATDC | Figure 2.7. Ratio of cylinder-average and exhaust NO concentration histories for a DI diesel engine, __ overall equivalence ratio, § = 0.6, injection timing at 27°BTC (13). Stoichiometric adiabatic flame temperature, 7;g- has been correlated with NO, formation expressed as emission formation index, EINO, in diesel engines using the following Arrhenius type expression [14]: EINO,, =constant x exp(E/RT 921) (2.28) T,.9-1's in K, E is an overall activation energy. T,-r is evaluated at initial T and P corresponding to those at tde under motoring using polytropic compression process. The polytropic index, n = 1.33 has used. The use of diluents like nitrogen, exhaust gas, oxygen and water varies the intake air composition and hence the stoichiometric adiabatic flame temperature. The measured NO, emissions with varying amounts of diluents at different engine loads and speeds correlated well with Ti», according to the relationship given in Eq. 2.28. On a log plot (Fig. 2.8), the EINO, data normalized with respect to emissions obtained with standard air (without diluents) for several engines showed a linear correlation With the reciprocal of T4-1,. The activation energy for DI engine was determined to be equal to - 285.0 ‘W/gmol, and for IDI engines equal to - 304.9 kJ/gmol. A single value of E correlates the emission data Varying by a factor of nearly 40. The good correlation obtained between EINO, and T;,., demonstrates the dominant importance of NO formation in close-to-stoichiometric burned gas regions. This model is a Very simple model and showed good corelation for the engines of 1980s. The engine design and34 Engine Emissions ed engines with very differ ion process. For advancec M22 ity : os affect the NO, formation pr ition teeny Wa ne oe mixing processes, single over ao taney as fe eh ser Gran net hoc Fig. 2815), Bot, the NO formation and decomposition ri Ent jeant i Jecomposition rates ‘ n in en i ignificant impact and NO d : ; t reactions to kinetics of yy Engine lo bas be Ne a temperature, the contribution of a tons Ee i becienaas changes and hence the deviations are observed from 8 bats flame temperature model. 10 IDI Engine of BIR=-36,700 £ 10 a DI Engine g EIR =- 34,300 3 5 2 0.10 oor Figure 28 Correlation of No, ixture (7; conditions at tap desde for my ‘emperature of stoichiometti® a .. and determined for mixtut Role of N:O in NO Kinetics Kinetic models based on the extg widely used for calculatio i Pressures typic formation. Reacti h mechanism, of diesel combust NO emisgi disoy ‘sed Carlier fy ines, a ; the Zekger HONS from qh wee for the SI engines, | ions involving Nyo formation sou omg alone dene e*8ines, However, at a !ecomposition may ate nt adequately predict significant contribu!Genesis of Engine Emissions 35 1 NO formation in diffusion and premixed combustion at high pronsuren [15} tn addition to the reactions 2,910 2.11 the Following additional renctions are connidered to affect NO formation in DI diesel engines; ' N,Ow oO 2NO (2.29) n OWN, 1M 8 NOM (2.30) 5 N,0 concentration in neatly at stendy sate, Making steady ntate assumption for [NJ and [N,O}, use of forward reactions 2,9, 2.10, 2.11, 2. and 2.30 yields the rate of NO formation as: : ALNOV/ at AON Hk, ks ) (231) where ks is the forward reaction rite constant for the reaction (2.30). The reaction rate constants for the __ eactions 2.29 and 2.30 are given in Table 2.2, In terms of rates shown in Table 2.2, ksi) a) (2.32) Vk s¢m)/k sq) )IM] ky ‘Table 2.2 Rate constants for the reactions involved in NO mechanism of NO formation, cm’/gmol-s, ‘Temperature range 2000 ~ 2800 K [15] Reaction Forward Reverse NO+0-2NO Ke= 2.9 7 10% exp C1057) Ka = L1S*10 > exp 3156871) OF N+ Mae NO4M key AM] = 2.0410! + k .qo/[M]= 6.2 * 10! exp (- 10672/T) exp (- 28247/T) OF N+ Mee NO 4M kesyay 5.45% 10" rexp (-12430/1) K soa) ® 13 * 10! exp(- 30005/T) —————— ee eeeeseseseseseSFSse ‘The reaction (2.30) ix pressure sensitive, ‘The rate constant, ks is a function of temperature and the third body concentration {M. The concentration [M] in computed from equilibrium considerations using modules like CHPMKIN, The sate constants in Tuble 2.2 however, have already accounted for the ‘concentration, {4}, Logarithms of k, and ks, which yovern the NO formation kinetics, are plotted in Fig, 2.9 for different Srre® ws 8 function of reciprocal of stoichiometric adiabatic fume temperature ranging from 2200 to 2900 Ke which covern the temperatures of interest at high und low ends for diesel combustion, The rein rates crovs each other at a temperature of about 2500 K, The eross over temperature slightly idiahai with increae in pressure in the range typical of motored diesel conditions (40 to 60 atm), At iilabatic Woichiometric combustion temperatures higher than about 2500 K, the contribution of reads tae constant k; and hence the Zeldovieh ix the dominant mechanism of NO. formation, However, the36 Engine Emissions é i yeratures is reversed at stoichiometric temperatures trend is. abn much more signi mechanism 10° 107 kor ks em? /gmol. § 38 conditions, The 28. [15] Hi may be noted that atthe tem Zeldovich mechanism, the reaction the normal burning zone Kemperatures of arou engines being typically in the ‘Ange 50 t0 60 atm, contribution fo engine out No emissions for the {ate injection (HCCI engines) that re ‘Sull in lower ag peratures of Tate constany NO Formation in 1py Engines Indirect injection (IDy Hence, NO formatio are; swirl-chamber a burs in the pre-char ) diesel n in these ind pre-chamber engines ang mber where al] the fuel is ‘ngines haye Mts are already ind 2600 engin, cen al eiBiNes Would be qi C1 is the injecte, 4 the contribution of ch mechanism. 2400 K an lower than \e Zeldovi jominates that of th ky (Zeldovich) 60 atm) =k, (3-body, tat 20 2 — ke (-body, 10. 40°42 44 10,0077, 5 4 (11K) about 2250 K Where the NO mechanism dominates in order of magnitude lower than those st © peak motoring Pressures in DI diese hanism is expected tg make a significant sing high other diluents or very combust tures, the N20 mec ines iabatic t ut due to their poor fuel efficiency- ace Y° Main designs of this OPE TON take place into stages. A rich mixiore 1S Pattlly burned sic fuel air mixture from theGenesis of Engine Emissions 37 pre-chamber is transported to the main chamber where excess air is present, The jet of partially burned gases issuing out of pre-chamber generates high turbulence that causes rapid mixing. Thus, in the main chamber most fuel bums as lean mixture, At light londs, most NO may form in the pre-chamber. But, at higher loads additional NO formation would occur in the main chamber. Although temperatures are higher in the pre-chamber, but except at light londs mixture is overall rich and hence a lower NO formation in the indirect injection engines compared (o DI engines results as shown in Fig, 2.10. 1600 ——noxor NOxIDI ~coo! 4200] ----co0 1400 1000 800 600 Concentration, ppm 7 ee 02 03 04 05 06 O07 08 Overall Fuel-Air Equivalence Ratio Figure 2.10 Comparison of NOx and CO concentrations in DI and IDI diesel engines [16] Main parameters that affect NO formation in the diesel engines are: injection timing, injection rate. ‘charge temperature and exhaust gas recirculation. These parameters also simultaneously influence smoke (Soot) emissions and fuel consumption. Their effects would be discussed later in the chapter 23.5 Formation of NOz Nitrogen dioxide in the exhaust gases of the spark-ignition engines is quite small 10 70 pans per million (ppm) compared to several hundred to thousands of ppm cr ative cose 0} a Spark-ignition engines NO» is generally les than 2%, while in diesel engines NOs ean constitute 1 2:44 Percent ofthe total emissions of nitrogen oxides. Concentrations of NOs and NO in the SI and dec engines exhaust are compared in Fig. 2.11. As stated earlier, NO: concentrations in the digest on jase significantly higher than those for the Sl engine. Hesel engine are NOs is rapidly formed in the combustion zone by reaction (2.33) between NO and HOs radical 18] » radical [18). NO + HO, ~ NO, + OH (2.33)38 Engine Emissions 5 ck © and O; on reaction wit ’ uipsequentty, in the post fame region NO} is converted back to N th atomig oxygen. NO; +0 + NOF Oz (2.34) der air or air-fuel mixture ¢; emperature bummed Bast v aused by ‘actions responsible for conversion of NO3 back to NO. Such a situation would result in relatively high NO: Soncentrations. This mechanism 18 supported by the high where the measured NONO ratio concentrations of NO» emissions observed in the diesel engines, jncrease as the engine load decreases (Fig. 2.10). es rapidly mix with col However, if the high | thigh turbulence it may quench re x 0 ea | —~ NO? xz + 80 E 3000 | | & | os | ¢ | z ace 0 oO | Diesel a = 1000 | SS fe 07 05 09 10 11 12 13 i ae ee eeu oe Se 06 07 708 Equivalence Ratio Figure 211 Exhaust N < 10 and NO; Sons i fuchar equivalence ratio 6.[13, 17) soncentrations in exhaust of SI and diesel engi engines as a function of 2.4 UNBURNED HYDROCARBONS Unburned hydrocarbons unburned by are emitted as part of : ydrocarbons nea y composed of 100 to 200 differ ney at volatil fuel inducted into the ensi hydrocarbons thet are present se oeatbons ile organic compo vans nite escapes combustion. The {actions of aromatics nd olefin ee Of Which 10 1 20 are NOCe™, Petroleum fuels ae 1¢ exhaust hydrocarbons luce jor compounds. Most of the compounds emitted mS consequent} ns Whi i the fuel molecules dee ahaa p eed hotochemveally wae are rich in aromatics and olefins The different organic compounds 0 Process avd PXTOlysis, chemical ensee: However, mam organic in atmospheric photochemtstn reeset it the na constitute about cal synthesis and partial dation of panicularly of the SI engines: Mean, 889 eine exhaust have dines eo hydrocarbon emissions measured neglecting methane cmisctan oot PNB photocheetitean Tent levels of toxicity and role (NMHCNMOG). Hydrocarbon concemmr term ically, reamounts in the exhaust gase NMOG). H ee . cys which is basically a carbon tion on in the exhausy methane Iauve: HC emissions are al ethod is specified in part per million as Gr neues! hYdrocase BY ons or organic BS VPPMC, oF sis lame ionization detecjons 39 Genesis of Engine En with propane (CsH,), its reading of HC is to be multiplied by 3 to obtain HC concentration in the units of C. PPM several processes are responsible for the inducted fuel to“eséape combustion, The sources of hydrocarbon emissions are also specific to basic engine type whether it is a homogencous spark (or heterogeneous compression ignition engine. In the crankcase scavenged, small two stroke SI engine designs, fuel-air mixture bypasses combustion and is ‘short-circuited’ to exhaust port. The mixture short- circuiting is the main source of hydrocarbon emissions in the small 2-stroke, SI engines. 2.4.1 HC Emissions from SI Engines In the SI engines, several processes contribute to unbumed hydrocarbon emissions. Most of the HC emissions result from “flame quenching’ as the flame is unable to propagate in fucl-air mixture located in cold regions and narrow passages i.c., the “crevices’ located around the combustion chamber. The narrow passages such as the annular clearance between piston top land and cylinder walls, the spark plug threads ec. are called ‘crevices’. The fuel and fuel air mixture in contact with colder walls and crevices are at a significantly lower temperature than the bulk gases. Flame gets quenched in these regions leaving the fuel unbumed. Absorption of fuel in lubricating oil film on the cylinder walls and in combustion chamber Geposits also contributes to unbumed HC. Incomplete fuel vaporization and leakage of fuel through exhaust valve(s) are some other sources of HC emissions. Under very lean mixture operation, flame may extinguish during expansion leading to quenching of combustion reactions in the bulk of mixture. This ally’ or ‘fully’ misfired combustion results in a sharp rise in unbumed HC. The relative contribution different sources to HC emissions varies with engine design and operating conditions. Exhaust valve Spark plug Miscellaneous crevices 3% “—~ Wall quench <10% il storage 25-50% Deposit storage 10% [7 Ring crevices 25-50%40 Engine Emi ns, Main sources of hydrocarbon emissions in the four-stroke, homogencous charge Spark in engines are shown schematically in Fig.2.12. These are: @ Flame quenching on the cylinder walls Gi) Crevices flame quenching Gil) Absorption and desorption in oil film on cylinder walls (iv) Absorption and desorption in carbon deposits in the combustion chamber () _ Misfired combustion or bulk gas quenching (i) Liquid fuel in the cylinder (vii) Exhaust valve leakage, and (viii) Crankcase blow by gases Flame Quenching For a flame to propagate, the energy release during chemical reactions should keep the reaction zone high enough temperatures to sustain the rapid combustion reactions. As flame approaches walls in combustion chamber, the walls being at a lower temperature than the gases in the flame, more and mo heat is lost by conduction from the flame and hot gases to the walls. This results in drop in temperature the reaction zone, which slows down reactions thereby lowering heat release rates. Finally, this proce leads to lowering of gas temperature below ignition point and flame quenching occurs. The fla propagating normal to the single wall will quench at some distance away where heat release in flame heat transfer from the flame to the walls become equal. Flame may not propagate in a tube or betwe two parallel plates if the tube diameter or distance between the parallel plates is below a critical val called quench distance or quench layer thickness. Basically, wall- quenching of flame is primarily du: heat transfer and not slowing of diffusion of reacting species For quenching of flame propagating normal to a single wall, energy balance at the instant ° quenching gives, 035 where k = Thermal conductivity of the unburned mixture, AT, = Characteristic temperature difference for heat transfer, juench distance, 4 a = Unburned mixture density, 5, = Laminar flame speed, h = Heat release per unit mass of the mixture burned, Zy» = Average specific heat of burned gases, and y AT; = Temperature rise on combustion in the flame, Introducing thermal diffusivity, a, = k/p,€,, in the Eq. 2.35, we getGenesis of Engine Emissions 41 cot. Pe 236) a © AT, a 5S, "f From equation (2.36), quench distance is inversely proportional to the laminar flame speed. The SoS, asionless quantity is known as ‘Peclet number’. For a given fuel- air mixture composition, a u 3 [pressure and temperature, 5, is proportional to the Peclet number at the flame quenching conditions. ‘Quench distance or quench laycr thickness depends on several parameters such as the wall geometry, fuel ‘composition, stoichiometry, flame speed, temperature and pressure of reactants, thermal conductivity and ‘uxbulence. Quench distance for stoichiometric mixtures of different fuels for laminar flame combustion ‘ge given in Table 23. Table 2.3 ‘Quench distance (mm) for different fuel-air mixtures, 6=1, P= latm and T = 20°C [8] Fuel 4, mn Foal 3, mam ee ee ee Hydrozea 06 Prope HOr 25 Methane 19 Methanol 18 Propane-air 21 Tsooctane 2.0 *H, replaced nitrogen in air proportions In Sl engine like conditions, typical two wall quench distance was estimated in the range from 0.2 to mm. Peclet number for flame quenching between two parallel plates is about 5 times of that for the ‘wall quench [19]. Thus, the single wall quench layer thickness would be in the range 0.04 to 0.2 For a Si engine operating on normal gasoline, using Eq. (2.35) and an average cylinder pressure = 10 S.= 0.2 m’s and a= 10° m/s, single wall quench distance is estimated to be about 0.05 to 0.1 mm 20}. The value of AT. and AT; were assumed to be nearly equal for these estimates. A good conespondence on quench distance is observed among the results of different studies. HC Emissions from Wall Quenching {21} in a spark ignition engine conducted photographic studies of flame region close to the bustion chamber wall immediately after arrival of flame. He showed existence of a thin non-radiating ayer adhering to the wall. Thickness of the quench layer varied from 0.05 to 0.4 mm and the quench er thickness decreased with increase of engine load. Increase in wall temperature reduces heat loss to the wall and consequently a smaller quench layer thickness results. At top dead center, the surface to ‘Volume ratio of the combustion chamber is at its maximum. At this point, the wall quench layer ‘Consist of 0.1 to 0.2 percent of total charge inducted. However, it was shown by studies on combustion ef ‘pre-mixed fuel air mixtures in combustion bombs that when all the crevices in the bomb were removed bs filing with solid material, unbumed HC concentrations were only about 10 ppmC [22]. This reaaiey from the fact that the hydrocarbons in the quench layer on_single wall diffuse in the hot burned se quite carly after flame quenching and get oxidized. After flame quenching, typically witht sos42. Engine Emissions Rents igh temperature gases Prox di diffusion in the high would get oxidized on ingle wall are one of i milliseconds aly wvas thought that te elon attic ae nas above pe ‘on combustion [22]. Initia ines. However, the ' hee Ha OE ee ae i. walt ench layers fo total unburned HC emission is rather ama that the contributic Crevice HC ‘i i regions into which fuel-air mixture can flow but flame Gcorsopase Hgiauhe Glee eect es ee ea gh rates of heat transfer to qe wall abd hence the quenching of flame occurs, Largest of all the crevices in combustion chambe i between the cylinder wall and the piston top land, and second land. Other crevices present are along the fasket between cylinder head and block, around intake and exhaust valve seats, threads around spark pag ‘and space around the central schematically in Fig, 2.13, Geerance, volume behind rings and ring gap. Some of these volumes change si motion. Table 2.4 gives the i generally decreased due cylinder crevice remains the ‘ume, as it is easier to reduce crevice volume crevice, 4 f03-0.4mm Top land crevice ‘Wall tat ing side crey) Compression ‘string back erevica-> ie Piston Figure 2.13 typica, dimensions oj The crevices constitute @ major source (0 to sis 60 pere 3 crovie ga imaller than the two-plate quench dine 0 PEtCent) of hydro a crevice plate cent otbute to HC emissions. Under non Same wat t2e°bON emissions. Ifthe plate quench distance is of the order of 0.4 es charge is filled in these crevices. At peak pressure Piston top lang crevices, Ot Pro} i ice and the engin, Pagate into the crevic e tm (23}. During med Operating conditions, the ® Maximum o™PteSsion and combuetion wpa SMOUNE Of gas would be stored intGenesis of Engine Emissions 43 crevices. The gas composition depends on the location of spark plug. With central plu: i forced into crevice would be unburned charge. Very little ured gases would be aad tn te ns jinder crevices unless the flame has reached piston top in some location nearest to spark plug before the k pressure occurs. The other crevices close to spark plug would be filled with a larger fraction of the fumed gases. During expansion the stored gases in the crevices begin to expand back into the cylinder. Part of the unburned charge from crevices that expands back into the cylinder is oxidized on mixing with the hot burned gases. Table 2.4 Typical Volume Contained in Engine Crevices, cm’ (Engine Displacement Volume/ Cylinder = 632 em’) [2] a Percent Volume, em ‘Clearance volume per eylinder 89.0 100 Volume above first ring (top land) 0.93 1.05 Volume behind first ring 0.47 0.52 Volume between Ist and 2nd rings (Second land) 0.68, 0.77 ‘Volume behind second ring 0.47 0.52 Total ring crevice volume 2.55 29 Spark plug thread crevice 0.25 0.28 0.30 034 Head gasket crevice Total crevice volume 31 35 Several investigations have been carried out to determine contribution of different crevice volumes to (C emissions [23-26]. Wentworth [24] was among the first researchers to investigate the contribution of iston and ring crevices to exhaust HC emissions. He moved the top ring as close to the piston crown as sible and the ring was sealed at top and bottom in its groove with ’O” rings. Tests with this sealed ring a production engine gave 47 to 74 percent reduction in exhaust HC emissions depending upon the ting conditions. Estigation of contribution of crevice volume to HC emissions is briefly discussed as follows. The ice temperatures are nearly equal to the wall temperatures that are cooled. Hence, the density of the Moed in the crevices is higher than in the cylinder. The maximum fraction of the unburned E, stored in crevices occurs at peak pressure and it is given bys sey, Yer? masta Bes mg VoFoler ; Es pressure and temperature and subscripts cr and 0 refer to the of intake stroke in the cylinder. Pm is the peak pressure in the ction in the crevice region, let us take Pmay /Py= 40, T,= 300 K, P= 400 K. conditions typical of an operating SI engine. Taking piston top land crevice volume of about .9 cm’ and the engine cylinder volume of 300 ‘em’ for a compact car and compression ratio = 9.0, about ‘of the charge would be stored in the piston ring crevice. The crevice charge would consist of 10 to 15 residual gases and some burned gases forced into it when flame propagates across the crevice ing. em, V, T and P ase mass volume, nnditions in the crevices and at the end inder. For estimation of the charge fra44 Engine Emissions Reduction in crevice volume is imy portant for reduction in HC emissions. Radial clearance beta Piston and cylinder in the production engines is of the order of 0.4 mm and top ang height around s mm. Reduction in these design parameters is limited by the thermal expansion of the mating piston cylinder, and the thermal stresses. Increase in radial clearance beyond two-plate quench distance wou allow flame penetration in the crevice. It would result in reduction of HC due to combustion progressn im the crevice volume [28, 29]. However, increase in radial clearance would tend to increase blow gases and loss in engine power output. Addition of chamfers to the crevice opening also allows flan Penetration in the crevice and hence HC reduction. The extent of HC reduction possible by change design of the piston ring crevice region also depends upon the operating conditions Such as air-fuel Tati, high residual gas dilution that oceurs under idle or use of exhaust gas recirculation (EGR). Unde conditions of high residual gas dilution or use of very lean mixtures, the flame may quench much before), Teaches the crevice region. Thus, increase in crevice volume by increasing radial clearance or chamfering of piston crown can result in an increase in HC emissions when using lean mixtures or EGR. 60 Fiston-Ring Crevice Head Gasket Cervice 40 50 30 Change in HC Emissions, % 8 revice Volume on lower h ., [26] foun ion ene Of HC Piston, piston ring excice wots none in head gasket Piston, piston ring crevice region Sensitivity of te °8, SHOWN oF crevice it was about 1.2. A higher yet i time available for oxida particularly on the exhaust is relatively small, emissions de, : ?Pends also, on the location of erevice Yolume is more effective than 's 2.14, They observed that for t! Sensitivity to head pean WAS about Q:2 While for the head gasket Tala burned hyde vice Volume ig attributed to low resident? valve side. Overalt Mons flowin, fives however, the contributions ope pl ee, crevices ig HC from Lubricating Oil Film Fuel hydrocarbons can be absorbe: compression strokes, which later g As the pressure increases, absorbed in the oil film, lin the oft fitm ‘tlm present on : et desorbed the cyl; ae and the pal pressure of the mene ses dane, Walls during inate 2 ‘ the fuel in the mitts Combustion and exp set ‘Afler combustion, the paral pret of fate increases and the fuel i in the gases is nearlyGenesis of Engine Emissions 45 Therefore, the fuel gets desorbed carly from oil after combustion and diffuses back into the burned gases. ‘The desorbed fuel vapours from oil film may get oxidized depending upon the temperature, pressure and composition of the burned gases. ‘Absorption and desorption of fucl vapors in the oil may be modeled using Henry's Law for dilute solutions in equilibrium [30-31]. The maximum amount of fel that can be dissolved per unit volume of oil is given by: Mf =M0X pePlH p (2.38) and mole fraction of fuel vapour in oil, ” So Xp eX P/N (2.39) Lon fe n+" ty fee's vhere, ny is the number of moles of fuel absorbed in oil, n, is number of moles of oil per unit volume, Xe ‘mole fraction of fuel in the combustion chamber gases close to the oil film, P is instantaneous cylinder sure and His the Henry's constant. Henry’s constant is a measure of fugacity of the fuel components Golute) in liquid phase or inverse of its solubility. Henry’s constant increases with temperature and rreases exponentially with increase in molecular weight of the solute, in this case the fuel. At 400 K fenry's constant for n-hexane, iso-octane and ethyl-benzene is 200, 120 and 45 kPa, respectively. The iue of Hy for hexane at 350 K reduces to approximately 90 kPa [19]. A larger fraction of the heavier J components would be absorbed in the oil as they have a smaller value of Hy, Taking average cylinder under compression stroke as 0.5MPa and typical oil film temperature equal to 400 K, the mole .ction of fuel vapour absorbed in oil film at equilibrium for the stoichiometric mixture of isooctane and would be 0.07. Diffusion rate of fuel vapours in the oil has to be fast enough for estimation of the fuel fraction issolved in the oil film according to Equation 2.38 as it relates to the state of equilibrium reached. The iffusion time has to be short compared to the characteristic engine time, The diffusion time constant may be approximated as, (2.40) where 6 is the oil film thickness and D is the diffusion constant, The lubricant oil film thickness is a strong function of oil viscosity and hence the oil temperature. It also varies with engine speed, The oil film thickness on the cylinder wall varies between 1 and 10 jim. The diffusion constant, D is of the order of 10° m’/s at 400 K and 10°"° m’/s at 300 K. The diffusion time at temperature of 400 K for a 1 ym thick oil film would be I ms and for a 10 pum thick film it would equal to 100 ms, For an engine speed of 3000 rpm, intake and compression strokes together would take 20 ms. Thus, for oil films of 1 to 2 um thickness state of equilibrium in fuel vapour absorption would be achieved under engine conditions, __ The absorption and desorption of fuel in the oil film and its contribution to the HC emissions involves several processes, Some of the absorbed hydrocarbon vapours in oil are carried to the crankcase. The lubricating oil film with absorbed fuel mixes dynamically with oil in the oil control ring and part of the fuel absorbed is transported to the crankcase where it is desorbed,46 Engine Emissions ‘ i sited on the pisto, When lubricating oi! was added to fuel or to the engine syliniee Ta sea ae i! cm chaust hydrocarbons increased in proportion to the oil added [32,. wtally oxidized fuel sp 01 (ae o erating on isooctane were identified as unburned fuel and tae original 1S ay the unburned ol or al exidstion species, HC eessionn desesed back fo the oi imal ev ina minutes time as the engine oil added to the cylinder was oxtnuses faved fuel emissions, whieh engine fuel, addition of engine oil did not result in an Sean [34] used sealed ring ae W attributed to insolubitity of propane in the engine oil. Wentwo nd observed ino 8 Tee deg completely eliminating crevice between piston, ring and cylinder 7 TBS] used graeme EC emi as the engine oil consumption eae On the ae aaaee ree hs ne ee a am ine oi served 30 percent reduction in HC emi . Redu ; s pees te low solbtlity poly-glycol lubricants replaced the mineral oil in the engine. Therefore Potential HC contribution of the engine oil film depends on the solubility of fuel in the engine oil and the amount of engine oil present on the combustion chamber walls. Wt may be concluded that the fuel absorption in the en, missions, although magnitude of its contribu HC emissions may be estimated generally less ecies ang gine oil film is responsible for some He tion has been shown to vary widely. Its contribution to ty s than 10 percent. ‘Combustion Chamber Deposits of deposits. As the pore size is small re not Penetrate into depos, % Pore ut ies pack into bumed gases, Soe e uMbumMed mixture comes eylinder. By i ‘bne of these hydrocarbons will bum UP tre mina the temperatiives dig tinder” at exhaust port and manifold or recycled inthe residual gageg 3 4° i8 either oxidized in the cy atoly converter is then, used fo reduce the exhaust HC emissions ro rankease blow by gases: THE A? dl supplied. "S to less than O.1percent of theFuel Inducted Genesis of Epgine Emissions 53 91% burnt flame 9% ” <— Feel Mixture | Fuol only Quench layers: Oillayers Deposits Liquid fue! In-cylinder 213 oxidized 118 oxidized) J In cylinder oxidation T 2.7% Exhaust oxidation 1.8% 1 Engine-out HC 1.8% 1 Catalyst oxidation Bumed Gases Oz, H20, CO 0.05 -0.1 % Tailoioe Emissions Figure2.19 Flow chart for gasoline fuel entering the engine: its normal combustion and pathways to generation of HC emissions in SI engines. Estimated magnitudes of ‘contribution of different sources and oxidation inside cylinder, at exhaust port and in exhaust catalyst [51]. 24.2 HC Emissions from CI Engines Diesel fuel has a higher boiling range and contains hydrocarbons of higher boiling point and molecular ‘mass compared to gasoline. The unburned hydrocarbons in diesel exhaust consist of almost 400 organic compounds ranging from methane to heaviest fuel components. Diesel exhaust hydrocarbons are composed of original fuel molecules, pyrolysis products of fuel compounds and partially oxidized hydrocarbons. In diesel engines, several events such as liquid fuel injection, fuel evaporation, fuel-air mixing, combustion and, mixing of bumed and unbumed gases can occur simultaneously. _$4 Engine Emissions cn i eral processes © nb Diesel combustion being heterogeneous in nature, Senet fel and ar can result in ne sdrocarbon emissions, During ignition delay period, overmising of fil a a ae a mi ie ae t00 lean to bum. Over- penetration of fuel Fe onitolted combustion, over-ich fa i Combustion chamber walls with Tiguid fuel, During mixing : combustion cha i contribute to HC emissions. The five main sources mixtures that are too rich to ignite and combust cont hydrocarbons in diesel engines are: ility limits, Overmixing of ftel and air beyond lean cama ee ‘Undermixing to fucl-air ratios too rich for comp! ee Impingement af fuel sprays on walls ic., spray over- See ct oramin Bulk quenching of combustion reactions due to mixing one Poorly atontized fuel from the nozzle sac volume and nozzle hol Overmixing and under-mixing of fuel in air by far are the most important causes of diesel HC emissions. Overmixing of Fuel Schematic of fucl spray injected radially outward into the swirling air before combustion (Fig.2.6) wa: Giscussed earlier, Toward downstream of swirling flow, the leading edge of the spray would have large concentration of the smaller droplets, which at ire expected to vaporize faster than the larger droplets inte Spry core. The local fuel-air ratio distributi ies wit i shown in Fig. 2.2 where equivelence ratio is less than th: region (LFOR) in which flame is un, Seichiometric region outside of the spray core, which i ‘Overmixing of fuel in air near the spray boundary results in an overlean Jean bility limits. This overmixed and ove combustion. In this region i n only slow oxidati lean mixture will neith slow oxidation reacti Products, unbumed fire] reaction: ining fuel vapours exists inside the outward bounday ie lean limit (= 0.3). This tegion is termed as lean flame blow ot able to propagate. The ignition starts in the slightly leaner than tht § termed as lean flammable region (LFR). mixture that is beyond t® (source ot Hoy Figure 2.20. Schematic of diese fuel g - ition, Lean flame out region (LOR) op ela eGenesis of Engine Emissions 55 ‘The width of the LFOR region depends on several parameters, prominent being ignition delay, ressure and temperature in the chamber, fuel type, and air swirl, Increase in temperature and prewure Presid reduce width of LFOR as the lean limit of combustion is extended. The contribution of LFOR 'o wepumed HC would depend on the length of ignition delay, amount of fuel injected and rate of fuel -air vnixing during the delay period. A longer ignition delay permits more time for the fuel vapours f diffuse farther into air and a higher percentage of fuel would be contained in the LFOR, An increas in HC ‘eesions with increase in the ignition delay for a direct-injection engine is observed (Fi 2.21). Thus, cmummixing of fuel and air during delay period is a significant source of HC emissions and is more pronounced under light load conditions that result in longer delay periods. 251 © 2500 & & 2000 2 & 1500 = 1000 o 9 a A i 3 & a oO S12 16 20) 28 Ignition Delay, deg. CA Figure 2.21 Effect of ignition delay on HC emissions in DI diesel engine [52]. Undermixing of Fuel is undermixing of the fuel with air. This can happen for the fuel overfuelling of the engine. The fuel left in the injector sac volume fully or partially vaporized during combustion process. It enters the engine cylinder at low velocity during expansion stroke, Therefore, during the period when although the gas temperatures are still high, yet as the fuel coming cut of nozzle sac has little time to, mix with air this fuel may not burn fully. Nozzle sac volume has been observed to be the main contributor to HC emissions in DI engines through the process ‘of undermixing of fuel (52, 53]. Effect of nozzle sac volume on HC emissions at different engine speeds for two types of nozzles, one with minimized sac So a eisevrlve covered orifice (VCO) type, is shown on Fig 2.22 [53], Fuel contained in the yhen the curve is extrapolated to zero sac nozzle holes also contributes to the HC emissions as seen W volume, Part of the fuel may remain in the sac volume and part may get oxidized, the balance being exhausted ‘as unburned HC. In the modern engines, nozzle sac volume has been decreased significantly. Engines also employed valve covered orifice nozzles to drastically reduce HC contribution of nozzle sac. However, as the liquid fuel provides cooling of the injector tip, VCO nozzles suffer from poor durability. Another cause of high HC emissions injected later in the cycle or because o| and nozzle holes at the end of injection gets_ 5 jssions Minimized sac- 56 Engine Emiss t volume 30 Increpsing engine speed 8 se i 3 VCO-nozzle ae 2s | ce / s2 ae f SE 10 by wiinimized So | {nozzle -sac; as | volume} (= —veo-nozate 0 10 20 3.0 Nozzle orifice and sac-hole volume, mm? Figure2.22. Effect of nozzle sac volume and hole type on HC emissions from DI diesel engine [53]. ‘Other Sources For the DI engines, at full load a minimum of about 40 i i i i f percent excess air ($< 0.7) is supplied to lis! smoke emissions. Overfuelling may occur during acceleration especially in turbocharged engines, a response of turbocharger to increase airflow rate i i : high EGR or disubed fuel injection 1 rate is slower than the increase in fuel injection rate. Use? engine fond Gncseses fa overs foe ystem calibration may also cause overfuelling. With increas? increase. Due to increase in the cyindsy pe eene® tio), engine and cylinder gas temperatures in engine load until critical fuel-sir oa HC emissions generally decrease with incre 09) Spte HC emissions increase shasply. and excess air is reduced to around 10 percent (9 ry impingement on combusti , and can be significant ion chamber walls parti i : ‘warm up or a ie of HC emissions, Under | ‘ara in small size engines is quite com™ tnd high HC emissions Dag 428 ming, some engine eyeles ray ere aee operation during engi exhaust white colored sppearance, which ns UH fuel dro may misfire resulting in bulk quenchité called as ‘white segtes™’Y @PPear in the exhaust gine 2.8 SOOT AND PARTICULATE FORMATION black smoke emissions. TH? of fuel type on soot formatio! ‘ ion 4 1» Th types in the decreas eon For The he effect a low [9]; mes, the soot formatiGenesis of Engine Emissions 57 ies> Alcohols>Paraffins> Olefins> ene For the diffusion flames, fuel molecules undergo pyrolysis and the soot-forming tendency is in the order; Aromatic icetylene> Olefins> Paraffins> Alcohols As the spark ignition engines generally operate close to stoichiometric air-fuel rat from these engines are not significant. With the use of unleaded gasoline, lead particulates from the SI engines have been climinated. However, soot may form in the gasoline direct injection (GDI) or direct injection stratified- charge (DISC) engines under some operating conditions as the ignition starts in a rich mixture that exists near spark plug. In stoichiometric GDI engines as the fuel may not be uniformly mixed and some fuel-rich zones may i ns have been associated with respiratory problems and are thought to be carcinogenic in nature, The particle size distribution is important as the particles smaller than 2.5 yum are respirable and are of main concern. The particles smaller than 2.5 ym constitute more than 90 percent mass of the total particulate matter in the diesel exhaust. From the GDI cagines, most of particulate mass is made of sub-micron particles. 2.5.1 Characteristics and Composition of Particulates In the test procedure defined by the US Environmental Protection Agency, the particulate matter is measured as any substance other than water collected by filtration of the diluted exhaust gases at or below 325K [54]. The composition of diesel particulate matter collected on a filter is schematically shown on Fig. 223. It_has two main components; one is dry soot or solid carbon material and the other is soluble organic fraction (SOF) that can be extracted by a solvent like dichloromethane. In addition, sulphates originating from fuel sulphur, nitrogen dioxide and water are also absorbed on the particle core formed by soot. The soluble organic fraction adsorbed on the solid soot core consists of hydrocarbons originating from fuel and lubricating oil, partial oxidation products and poly aromatic hydrocarbons. Depending upon engine design and operating conditions, the soluble fraction may vary from 10 to 90 percent of particulate mass but mostly lies in‘the range from 20 to 45 percent. Dry soot is a carbonaceous matter, which results from several processes like pyrolysis, dehydrogenation and condensation of fuel molecules. Dry soot is mainly the carbonaceous fraction of the particulate and its typical chemical formulae are CiH, CsH and CygH [55]. About 5 to 10 % by mass oxygen and 0.5% nitrogen are also present. Typical empirical formula of dry soot would be CHo11On0«sNomis (1, $5]. The soluble organic fraction originates from the fuel and oil hydrocarbons, and therefore has H/C ratio = 2, although depending upon engine operating conditions it may vary from 1.25 to 2.0. The hydrocarbons Cy; to Cy are present in particulate SOF, the C23 ~ C2, being close to the mean. Typically SOF has an empirical formula CH,,6sOo.1Nocor Other inorganic compounds of iron, silicon (fuel contamination), phosphorous, calcium, zine (source is Cil) etc, are also present in traces in the particulates. The composition of particulate matter depends on the engine operating parameters, conditions in the exhaust system and the sampling and collection system, The magnitude of different constituents of diesel ae matter can be significantly different for different engine designs These are also affected by the aoe fae technology employed. The particulate composition is varying as the emission standards content ep Btened due to improved combustion systems and fuel quality particularly tow sulphur ae seduction in lubricating oil consumption of the engines, The average particulate enki equip rie 1990 and Euro 3 turbocharged aflercooled diesel engine and a US heavy duty diesel ipped with a particulate filter are compared on Fig 2.24. Considerable reduction in total PM58 Engine Emissions Solid Carbon spheres (0.01 | 0.08ym dia) form to make s: particle agglomerates (0.0! um dia) with adsorbed hydrocarbons Adsarbied Hydiucasinue Liguid candenend hydroeartion partioloe Sulphate with hydration Figure 2.23 Schematic representation of diesel particulate matter collected on Silver Pro-1980, Carbon ‘ios 1D o1-soF 5 Fuct-soF TD Sulphate | 82, 0.152 ED Remainder 0.028 0.083 PM=0.37 gkWwn 0.0003. US 2007 HD Engine 6.0008 0.0004. PM = 0.0018 aw Figure 2.24 Typical diesel particulate matter cor i J Mpasition far Pye [NW ys 3007 #4 heavy duty engines 1000, Fane 8 and a USGenesis of Engine Emissions 59 emissions, carbon (dry soot) content, fuel and oil derived SOF and sulphates per unit work are taking place. SOF content also depends to a large extent on the design of combustion system. Sulphates depend ee the sulphur content of diesel fuel. For low PM emissions, a low sulphur diese! fuel is an essential requirement. Currently, low sulphur fuels with sulphur content of less than 15 ppm are being widely used in Europe, the USA and many Asian countries. Electron micrographs of soot formed in diffusion flames and in the engine combustion chamber show these to be aggregates of primary spherical particles of size range 10 to 80 nm, most of them being in the 15 to 30 nm size range [S7, $8]. Typical electron micrographs of soot in acetylene-air diffusion flames are shown on Fig. 2.25 [58]. Primary particles of spherical shape of about 30 nm size and branched hainlike soot aggregates are clearly observed in micrograph at 80000 magnifications. A single spherical soot particle contains 10° to 10° carbon atoms. Initially, the combustion-generated soot particles have one hydrogen atom for every § or so carbon atoms with a density of 1.8 g/cm’. During expansion more ns may be absorbed on the particles. The primary soot particles form aggregates in the Combustion chamber of about 100-200 nm in size containing up to several hundred primary particles. ‘These aggregates may further agglomerate to particles as large as 1 um. The aggregates may’ resemble a cluster of spheres in round shape, branched or chain like structure. 80 k magnification 8k magnification Figure 2.25 Transmission electron microscope (TEM) images of soot in acetylene-air diffusion flames. Micrograph at 80K magnification shows branched chainlike soot aggregates [58] Soot is formed in diesel engine environment at temperatures in the range 1000 to 2800 K and at 50 ~ 100 atm pressure. The fuel molecules containing 12 to 22 carbon atoms and atomic H/C ratio of about 2 finally end in spherical soot particles having H/C ratio of about 0.1, which grow in size and agglomerate. In the combustion chamber, fragmentation of large molecules to smaller ones, dehydrogenation, polymerization, condensation reactions and agglomeration progress simultaneously forming soot Soot