Emissions of Regulated Pollutants from a Spark Ignition Engine

Jun 14, 2003 - Institut Franc¸ais du Pétrole, 1 et 4 avenue du Bois Préau,. F-92500 Rueil Malmaison ...... 1991 (in French). (24) Zervas, E.; Monta...
0 downloads 0 Views 217KB Size
Download PDF
Environ. Sci. Technol. 2003, 37, 3232-3238

Emissions of Regulated Pollutants from a Spark Ignition Engine. Influence of Fuel and Air/Fuel Equivalence Ratio E . Z E R V A S , * ,† X . M O N T A G N E , † A N D J. LAHAYE‡ Institut Franc¸ ais du Pe´trole, 1 et 4 avenue du Bois Pre´au, F-92500 Rueil Malmaison Cedex, France, and Institut de Chimie des Surfaces et Interfaces, 15 rue Jean Starcky, F-68057 Mulhouse Cedex, France

A spark ignition engine is used to determine the influence of fuel composition and air/fuel equivalence ratio on the exhaust emissions of regulated pollutants. Two specific fuel matrices are used: the first contains eight hydrocarbons and the second contains four oxygenated compounds. A specific experimental design is used for these tests. Fuel aromatics increase the exhaust CO, HC, and NOx at stoichiometry, lean and rich conditions. Lambda is more important than fuel composition in the case of CO and HC. At stoichiometry, the addition of oxygenated compounds can decrease exhaust CO, HC, and NOx up to 30%, 50%,and 60%, respectively. Under these conditions, the addition of 5% of 2-propanol is the most effective for the reduction of CO, the addition of 20% of ethanol for the reduction of HC, and this of 5% of methyl tributyl ester (MTBE) for the NOx. The addition of oxygenated compounds can decrease CO by 30% at lean conditions, while no decrease is observed at rich ones; HC and NOx can decrease up to 30% and 80%, respectively, under lean conditions and 50% under rich ones. At all lambda tested, exhaust NOx increases with the addition of 20% of 2-propanol.

of ethanol (6) or MTBE (2, 6). These differences can be explained by the different methods used for the precise determination of λ, especially in the older works (14-16). The addition of oxygenated compounds, as ethanol, produced from renewable biological sources can also decrease the emission of carbon dioxide, which is considered today as one of the factors for the global climate changes. Even if the addition of oxygenated compounds decreases generally the CO and HC emissions, it is not certain that the air quality improves. In one study, an increasing air quality, for CO, HC, and NOx, is observed after the introduction or increase of oxygenate content (17), in another one, atmospheric NOx do not change (18). The addition of oxygenated compounds increases the exhaust emission of other nonregulated pollutants. For example, the addition of methanol increases exhaust formaldehyde and those of ethanol increases exhaust acetaldehyde emissions (5, 10, 19), which are more reactive in the atmosphere than their parent alcohols (20). The addition of oxygenated compounds also increases the exhaust emission of organic acids (21). Spark ignition engines operate under stoichiometry and most of the authors studied these conditions. But modern engines also operate under lean or rich conditions (lean burn engines, cold start or very high loads), and the influence of fuel composition under these conditions has not been well studied. The addition of oxygenated compounds probably does not have the same effects under lean or rich conditions. Only one author studies the influence of fuel/air equivalence ratio (22). In our previous articles we presented the influence of fuel composition and air/fuel equivalence ratio (λ) on the emissions of organic acids (21), alcohols, and carbonyl compounds (19) of a SI engine. It was found that these two parameters have a significant influence on the emissions of these unregulated compounds. Continuing this research, we present here the influence of fuel composition and λ on the emissions of regulated pollutants (CO, HC, and NOx). Two specific fuel matrixes containing eight hydrocarbons and four oxygenated compounds are used to study the influence of fuel composition and air/fuel equivalence ratio on the emissions of CO, HC, and NOx. A comparison between the oxygenated fuels is also performed.

Introduction

Experimental Section

Fuel composition is one of the major parameters for the gas concentrations of exhaust pollutants of spark ignition engines. The correlations between fuel composition and exhaust concentration of the three regulated pollutants, CO, HC, and NOx, are presented in many articles (1-8 and many others). Gasoline is mainly composed by hydrocarbons, but all its components do not have the same participations to the formation of exhaust pollutants (2, 4, 5). The addition of oxygenated compounds into gasoline is also proposed to decrease exhaust emissions of regulated pollutants. The main oxygenated compounds studied are methanol (6, 9), ethanol (6, 9, 10), and MTBE (2, 6-8, 10-12, and many others). Many authors present a decrease of exhaust emissions by the addition of oxygenated compounds, but others present no change (for example the addition of MTBE does not change the emissions of CO (7, 13), HC (13), or NOx (11-13)); an increase is even noticed in the case of NOx after the addition

A Cooperative Fuel Research Committee (CFR) spark ignition engine was used for these tests. This engine is a small monocylinder engine (displacement ) 6.11 10-4 m3, bore ) 0.08255 m, stroke ) 0.1143 m) used for the octane number determination. The λ used varies from 0.83 to 1.25 (calculated from the exhaust gas analysis used five gases: CO2, CO, HC, NOx, and O2), while all other engine parameters were kept constant (speed ) 15 Hz, compression ratio ) 6:1, indicated mean effective pressure ) 4.5 × 105 Pa). Modern engines emit lower pollutant concentrations than the CFR engine, but this one allows the determination of the most important correlations between fuel composition and exhaust emissions of regulated pollutants. As the subject of this work was to find out the aforementioned correlations, no catalytic converter was used. Carbon monoxide was analyzed by nondispersed infrared, nitrogen oxides by chemiluminescence and total unburned hydrocarbons by a flame ionization detector. Two fuel matrixes were adopted in this study (Table 1). The first one (synthetic fuels matrix) contains eight hydrocarbons: n-hexane, 1-hexene, cyclohexane, n-octane, isooctane (2,2,4-trimethylpentane), toluene, o-xylene, and eth-

* Corresponding author phone: 331-69 27 84 77; fax: 331-69 27 82 92; e-mail: [email protected]. Present address: Renault, CTL L26 0 60, 1, Alle´e Cornuel, F-91510 Lardy, France. † Institut Franc ¸ ais du Pe´trole. ‡ Institut de Chimie des Surfaces et Interfaces. 3232

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 14, 2003

10.1021/es026321n CCC: $25.00

 2003 American Chemical Society Published on Web 06/14/2003

TABLE 1. Chemical Analysis and Octane Number of the Fuels Used (% Vol) hexane hexene cyclohexane octane isooctane toluene o-xylene ETB hexane 1-hexene cyclohexane n-octane isooctane toluene o-xylene ETB methanol ethanol 2-propanol MTBE oxygen (wt %) alkylate RON H/C O/C

R 7 7 7 7 7 7 7 7

M5

E5

P5

6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 5 0 0 0 3.08

6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 0 5 0 0 1.98

6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 0 0 5 0 1.45

6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 0 0 0 5 0.90

5.6 5.6 5.6 5.6 5.6 5.6 5.6 5.6 20 0 0 0 11.89

41.8 87.5 2.01 0.013

41.8 86.8 2.01 0.008

35.2 35.2 35.2 98.6 97.4 96.6 2.23 2.14 2.10 0.12 0.074 0.053

42 2 2 2 2 2 2 2

2 42 2 2 2 2 2 2

2 2 42 2 2 2 2 2

2 2 2 42 2 2 2 2

2 2 2 2 42 2 2 2

2 2 2 2 2 42 4 2

2 2 2 2 2 2 42 2

2 2 2 2 2 2 2 42

44 63.7 2.19

44 83.7 2.06

44 88.7 2.06

44 43.6 2.16

44 93.8 2.16

44 101.3 1.66

44 96.6 1.71

44 44 41.8 41.8 100 85.2 89.4 90.4 1.72 1.95 2.04 2.02 0.028 0.018

MTBE5 M20

E20 5.6 5.6 5.6 5.6 5.6 5.6 5.6 5.6 0 20 0 0 7.75

P20 MTBE20 5.6 5.6 5.6 5.6 5.6 5.6 5.6 5.6 0 0 20 0 5.72

5.6 5.6 5.6 5.6 5.6 5.6 5.6 5.6 0 0 0 20 3.60 35.2 93.1 2.06 0.033

ylbenzene (ETB), while the second one (oxygenated fuels matrix) also contains four oxygenated compounds: methanol, ethanol, 2-propanol, and MTBE. An experimental design, specially adapted for mixtures, is used to determine each component quantity in the blend of synthetic fuels (23). To avoid a high dispersion in physical properties of the fuels used, an alkylate, containing basically isooctane but also 1.5% of benzene, was used as the base fuel for these blends. The reference fuel R contains an equal content of each of the eight compounds, while the other fuels contain 42% of a major component. The oxygenated matrix was obtained by the addition of 5% or 20% of one of the four oxygenated compounds to the fuel R. Two simple fuels were also used: iC8, which is pure isooctane and iC8T, which is a mixture of 80% of isooctane and 20% of toluene. These fuels allow the study of the addition of an aromatic component to an alkylate basis. The name of each fuel was chosen to recall its major component. The chemical composition and physical properties of these fuels are quite different than the commercial ones, but these matrixes allow the study of the influence of the chemical composition of the fuels on the emission of regulated pollutants. More details about these fuels are presented elsewhere (24). All tests were doubled, and average values were used. The CO, HC, and NOx relative standard deviation is 3.5, 8.5, and 12.5%, respectively, estimated from five points of the reference fuel R.

Results and Discussion Emissions of Synthetic Fuels at Stoichiometry, Lean and Rich Conditions. CO. At stoichiometry, aromatics (1, 3, 4) and octane enhance exhaust CO, while isooctane, 1-hexene, n-hexane, and cyclohexane decrease it (Figure 1). The addition of 20% of toluene to pure isooctane increases the concentration of exhaust CO. Other authors present a smaller difference by the increase of aromatic content on the CO emissions on the European driving cycle (13) or even an increase of exhaust CO by the reduction of fuel aromatics (2). This latter author reports a small decrease of CO by the decrease of fuel olefins from 20 to 5%. CO emissions are enhanced at rich conditions due to lack of oxygen and decreased at lean ones due to oxygen excess. All fuels emit from 0.06 to 0.1% of CO at lean conditions, compared to 0.3-0.5% at λ ) 1.0 (Figure 1). The differences between λ ) 1.25 and λ ) 1.11 are negligible. Under these conditions, hexane, isooctane, and toluene enhance the emission of CO, while, 1-hexene, cyclohexane, and octane decrease it. At rich conditions, all fuels emit 2.8-3.1% and 5.7-6.8% of CO at λ ) 0.91 and λ ) 0.83, respectively.

FIGURE 1. Emission of CO for the fuels of the , synthetic matrix ., at stoichiometry (middle bars), lean (lower bars), and rich (upper bar) conditions. Mean: mean exhaust concentration, min and max: min and max exhaust concentration for a 95% confidence level. Aromatics still produce the higher concentrations, but these differences are lower than at λ ) 1.0. The addition of toluene VOL. 37, NO. 14, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3233

TABLE 2. Influence of Each Fuel Component on the CO, HC, and NOx Exhaust Emissionsa CO

hexane 1-hexene cyclohexane n-octane isooctane toluene o-xylene ETB M5 E5 IP5 MTBE5 M20 E20 IP20 MTBE20

HC

NOx

1.25

1.11

1

0.91

0.83

1.25

1.11

1

0.91

0.83

1.25

1.11

1

0.91

0.83

D D D D D D D D D D D D D

D D D D D D D D D D D D D

D D D D E E E D D D D D

D D -

D D D E D D

D D D D E D D D D D D D D

D D D E E E D D D D D D D

D D D E E D D D D D D D D D

D D D E E E D D D D D D D D

D D D D E E D D D D D D D D D

E D D D E E E D D D D D D E D

D D D D E E E D D D D D D E D

D D D D D D D D D D E D

D D D E E E D D D D D E D

D D D E E D D D D D D E E

a E ) enhance more than 5% relatively to fuel R, D ) decrease more than 5% relatively to fuel R, - ) no influence (enhance or decrease less than 5%).

into pure isooctane under lean conditions does not change CO emissions, while it increases them under rich ones. The differences between CO emissions of the fuels used are not very important at rich conditions. These effects are summarized in Table 2. The value of λ is more important than fuel composition for the emission of this pollutant. The differences between CO emitted from all synthetic fuels are less than 40%, 25%, and 10% under lean, stoichiometric, and rich conditions, respectively. Concerning the λ influence, exhaust CO is 6-10 and 12-20 times more important at rich conditions than at stoichiometry (for λ ) 0.91 or λ ) 0.83, respectively), while it is 3-6.8 times lower than at λ ) 1.0 at lean ones. HC. At stoichiometry, the emission of exhaust hydrocarbons is enhanced by fuel aromatics (1-4) except ETB and decreased by all other fuel components (Figure 2). The addition of toluene to pure isooctane enhances slightly the exhaust concentration of HC. Another author presents a smaller influence on exhaust HC by the increase of aromatic fuel content on the European driving cycle (13), and no change of HC by the decrease of fuel olefins from 20 to 5% (2). As in the case of CO emissions, HC emissions are enhanced under rich conditions due to lack of oxygen and decreased under lean ones. Under these latter conditions, the HC exhaust concentrations vary only from 155 to 326 ppmv (Figure 2), compared to 606-1131 ppmv at λ ) 1.0. Toluene and o-xylene still produce the higher concentrations, while 1-hexene and cyclohexane produce the lowest ones. Under rich conditions, the HC concentrations vary from 1034 to 2131 ppmv. Toluene and o-xylene still enhance these emissions, while 1-hexene and cyclohexane produce the less. The addition of toluene into pure isooctane increases slightly the HC emissions compared to pure isooctane, for both lean and rich conditions. These effects are summarized in Table 2. As for CO, lambda is generally more important than fuel composition for the HC emissions. The differences between HC emitted from all synthetic fuels are less than 30% under all λ tested. Concerning the influence of λ, exhaust HC is 1.3-1.7 and 1.9-2.2 times more important at rich conditions than at stoichiometry (for λ ) 0.91 and λ ) 0.83, respectively), while they are 2.9-3.8 and 3.8-5 times lower than at λ ) 1.0 at lean ones (λ ) 1.11 and 1.25, respectively). All these variations are less important than those of exhaust CO. NOx. It has been reported that exhaust NOx is enhanced by fuel aromatics at λ ) 1.0 (1-4). The exhaust temperature 3234

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 14, 2003

FIGURE 2. Emission of HC for the fuels of the , synthetic matrix ., at stoichiometry (middle bars), lean (lower bars), and rich (upper bar) conditions. Mean: mean exhaust concentration, min and max: min and max exhaust concentration for a 95% confidence level.

FIGURE 3. Emission of NOx for the fuels of the , synthetic matrix ., at stoichiometry (middle bars), lean (lower bars), and rich (upper bar) conditions. Mean: mean exhaust concentration, min and max: min and max exhaust concentration for a 95% confidence level. of aromatic fuels is higher than that of the others fuels, so NOx emission is higher (25). Our results show that hexane, toluene, o-xylene, and ethylbenzene enhance very slightly exhaust NOx (Figure 3 (13)), while the lowest concentrations come from 1-hexene. The addition of 20% of toluene to pure isooctane also increases the exhaust concentration of NOx. Another author presents a smaller influence on exhaust NOx by the increase of aromatic fuel content on the European driving cycle (13) or presents no change by the decrease of fuel olefins from 20 to 5% (2). NOx emissions present a maximum around λ ) 1.1, because at this point, the combination between flame temperature and oxygen concentration is more favorable, even if flame temperature presents a maximum at stoichiometry. The highest NOx emissions are emitted at λ ) 1.11 (Figure 3, from 529 to 1560 ppmv, 1.1-1.5 times more than at λ ) 1.0). Toluene, o-xylene, ETB, and hexane still enhance these emissions, while 1-hexene emits the lowest ones. These differences are more important than at λ ) 1.0. The same results are observed at λ ) 1.25, where NOx emissions vary from 264 to 836 ppmv. At rich conditions, NOx emissions are much lower than at stoichiometry: from 183 to 507 ppmv at λ ) 0.91 and from 67 to 157 ppmv and λ ) 0.83. The

FIGURE 4. Change of CO emissions from the addition of methanol, ethanol, 2-propanol, and MTBE for the five λ used. The curves of λ ) 1.25 and λ ) 1.11 are superposed. addition of toluene into pure isooctane increases NOx emissions compared to pure isooctane, for both lean and rich conditions. These effects are summarized in Table 2. Both λ and the type of fuel are important parameters for NOx emissions. The differences between NOx emitted from all synthetic fuels are less than 60% under all λ used. Concerning the influence of λ, NOx is 2-3 and 6-10 times lower at rich conditions than at stoichiometry (for λ ) 0.91 and λ ) 0.83, respectively). At lean conditions, they are 1.11.5 times more at λ ) 1.11 than at stoichiometry, while they are 1-2 times less at λ ) 1.25. For the three pollutants, aromatics increase exhaust emissions, while 1-hexene and cyclohexane produce the less ones. Fuel composition is much less important than lambda for the CO and HC emissions, especially under rich conditions. Emissions from Oxygenated Fuels at Stoichiometry, Lean and Rich Conditions. CO. Figure 4 presents the average change of CO emissions due to addition of oxygenated compounds into fuel R. The addition of methanol decreases 2.5-5% the exhaust CO at stoichiometry, with no significant differences between M5 and M20 fuels. Literature presents also a decrease at λ ) 1.0 from methanol-containing fuels (6, 9). This addition decreases exhaust CO also at lean conditions. This decrease is more important than at λ ) 1.0 and can VOL. 37, NO. 14, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3235

reach 30% in the case of M5 fuel but less (20%) in the case of the M20 one. The changes are the same in the case of λ ) 1.25 and λ ) 1.11. At rich conditions, exhaust CO remains practically unchanged (differences between -4% to 2%). The addition of 5 or 20% of ethanol decreases exhaust CO by 20% at stoichiometry, in accordance with literature (5, 6, 9, 26, 27). At lean conditions, this decrease is more important in the case of E5 fuel reaching 30%, while the E20 fuel decreases it by 20%. At rich conditions, the addition of ethanol does not influence significantly exhaust CO emissions (differences between -7% and 1%). Our results are not in accordance with those reported previously by other investigators, i.e., the addition of ethanol has no effect on CO emissions under lean conditions, while this decrease can reach 20% under rich ones for an E20 fuel (22). No obvious interpretation can be given to this difference; the only probable one is the presence of 7.8% of water in the alcohol used in this latter study, that can affect the miscibility of alcohol/gasoline. In the case of P5 fuel at stoichiometry, the addition of 2-propanol decreases exhaust CO by 28%, against 8% in the case of the P20 one. Literature presents also a decrease (about 50% at λ ) 1.0 and idle conditions from a P10 fuel (6)). At lean conditions, P5 fuel decreases exhaust CO by the same percentage, while the P20 one reaches 20%. At rich conditions, exhaust CO remains unchanged by the addition of this oxygenated compound (differences form -3 to -0.3%). The addition of MTBE decreases slightly the emission of CO at λ ) 1.0 (less than 8%). Literature presents a decrease of CO emissions at stoichiometry by the addition of MTBE (1-3, 6, 11, 12, 26, 28, 29) or no change (7, 13). Another author presents that a significant decrease of exhaust CO from the addition of MTBE occurs only under high engine loads (8). Figure 4 shows that this decrease is more important in the case of lean conditions reaching 30% for both MTBE5 and MTBE20 fuels. At rich conditions, exhaust CO remains practically unchanged by the addition of MTBE (differences from -6% to 1%). Comparing these four oxygenated compounds in case of stoichiometry, the most effective fuel is the P5 with a decrease of 28%, followed by the E5 (-20%), while the addition of methanol or MTBE decreases exhaust CO very slightly. At lean conditions, there is practically no difference between λ ) 1.25 and λ ) 1.11, and it must be noticed that low oxygenate content fuels lead to a greater decrease of CO emissions than the high oxygenated content ones (30% against about 20%). This decrease seems to be independent of the oxygenated compound. At rich conditions, the addition of oxygenated compounds does not influence significantly the CO emissions. All four oxygenated compounds emit less CO than the ,best. synthetic fuel for the CO emissions (the isooctane one) under lean conditions (20-30%), but they emit more under stoichiometry and rich conditions (3-30%). These effects are summarized in Table 2. At stoichiometry and lean conditions, the decrease of exhaust CO due to the addition of oxygenated compounds is more important than the percentage of this compound in the fuel. This statement indicates that the decrease of exhaust CO comes not only because of a dilution of the fuel but also that the addition of oxygenated compounds enhances the combustion of CO in the cylinder or during the postcombustion processes. This must also be the reason of the unchanged emissions of CO between reference fuel and the oxygenated ones at rich conditions. HC. Figure 5 presents the average change of HC emissions due to the addition of oxygenated compounds in the fuel R. At stoichiometry, exhaust HC decreases up to 18% by the addition of 5% of methanol and 29% using the M20 fuel. Literature presents almost similar results at stoichiometry (6). For the five λ used, the addition of 5% of methanol 3236

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 14, 2003

FIGURE 5. Change of HC emissions from the addition of methanol, ethanol, 2-propanol, and MTBE for the five λ used. decreases exhaust HC almost by the same percentage (1824%). The changes from, the M20, fuel are slightly more important (21-29%) but seem again independent of the λ. At stoichiometry, the addition of ethanol decreases exhaust HC by 12% and 48%, respectively, for the E5 and E20 fuels. Literature presents almost the same results at stoichiometry (5, 6, 26, 27). Figure 4 shows that the decrease due to E5 fuel is almost independent of λ and remains at 7-9% (except at λ ) 1.25 where it reaches 19%), while this of E20 presents two zones: lean conditions with a decrease of 9-28% and stoichiometry and rich conditions with a more important decrease, 46-48%. Our results are not in accordance with those already reported. Literature presents a very slight decrease (3-5%) from the addition of ethanol up to 20% in the same range of λ (22). No obvious interpretation can be given to this difference except this given previously in the case of CO. The addition of 2-propanol influences little exhaust HC: a decrease of only 6% at stoichiometry. Literature presents also a decrease from the use of a P10 fuel (6). For all five λ used, this decrease is from 4 to 8% in the case of P5 fuel and 1 to 13% in the case of P20. As for the addition of ethanol, the decrease due to addition of 20% of 2-propanol is higher under rich than under lean conditions (5-13% against 1-2%). The addition of MTBE decreases, at λ ) 1.0, exhaust HC by 20% and 12%, respectively, for the MTBE5 and MTBE20

fuels. Literature presents a decrease of exhaust HC by 5-20% after the addition of MTBE (1, 2, 3, 5, 6, 7, 11, 12, 26, 28, 29) or no significant difference (13). Another author shows that a significant decrease of exhaust HC from the addition of MTBE occurs only under high engine loads (8). Figure 5 presents that this decrease is quite independent of λ; it remains from 18 to 25% in the case of MTBE5 fuel and from 7 to 18% for the MTBE20 one, for the five λ used. No differences can be observed between rich and lean conditions. At stoichiometry, the most effective fuel for the decrease of exhaust HC is the E20, which decreases them by almost 50%, following by the M20 with a decrease of 30%. At lean conditions, the differences between λ ) 1.25 and λ ) 1.11 are generally small. The best oxygenate is methanol (decrease 20-29%) followed by ethanol (10-30%), MTBE (10-18%), and 2-propanol (1-9%). At rich conditions, the best fuel is the E20 one (46-47%), followed by methanol (18-28%), MTBE (7-25%), and 2-propanol (5-13%), while the E5 fuel decreases exhaust HC only by 7-9%. All four oxygenated compounds emit 5-38% more HC than the ,best. synthetic fuel for the HC emissions (the hexene one), except two fuels under stoichiometric and rich conditions: the M20 one that emits comparable concentrations and the E20 that emit 2229% less that the hexene one. These effects are summarized in Table 2. The decrease of exhaust HC is more than 5% in the case of low content oxygenated fuels, indicating that the addition of an oxygenated compound enhances the combustion of HC or their postoxidation. This is not always the case for the addition of 20% of an oxygenated compound. The addition of methanol or ethanol generally decreases exhaust HC by more than 20%, while this of 2-propanol or MTBE decreases them less, indicating that only the first two oxygenated compounds enhance the combustion of HC. NOx. The average change of NOx emissions due to addition of oxygenated compounds is presented in Figure 6. At stoichiometry, the addition of 5 or 20% of methanol decreases exhaust NOx by 18% and 7%, respectively. Literature presents also an increase at stoichiometry from a M10 fuel (6). For the five λ used, the addition of 5% of methanol decreases exhaust NOx by 11-27%, with the more important decrease under lean conditions (24-26% against 11-17% under rich ones). The addition of 20% of methanol changes very little exhaust NOx (decrease of 2-8% for all λ used). The addition of 5% or 20% of ethanol decreases exhaust NOx by 22% and 19%, respectively, at λ ) 1.0. At stoichiometry, literature presents an increase from the addition of ethanol (5, 6) or a decrease of 5-25% on the European cycle using an E5 one (26). For the five λ used, the addition of 5% or 20% of ethanol in the reference fuel decreases NOx by 15-30%. Generally, these changes are more important at lean conditions. Literature presents that the use of an E10 fuel decreases slightly exhaust NOx (3-7%), while this of an E20 one increases them; an exception is observed at very lean conditions, while even the E20 fuel decreases exhaust NOx (22). The addition of 2-propanol decreases exhaust NOx by 31% in the case of P5 fuel, while it increases them by 17% in the case of the P20 one (at λ ) 1.0). An increase of 5-10% is already presented in the case of a P10 fuel (6). For the five λ used, fuel P5 decreases NOx by 27-38%, while the P20 one increases them by 15-34%. In the last case, the increase observed is more important at lean conditions. The addition of 5% or 20% of MTBE at stoichiometry decreases exhaust NOx by 60% and 18%, respectively. Literature presents a decrease of 7-12% using MTBE15 (7, 26); other authors found that the addition of MTBE does not decrease significantly NOx emissions (1, 5, 11-13, 28, 29); others present even an increase of 5-15% by the addition of 10-12% of MTBE (2, 6). These differences can be explained

FIGURE 6. Change of NOx emissions from the addition of methanol, ethanol, 2-propanol, and MTBE for the five λ used.

by the different techniques used for the precise determination of λ. Figure 6 presents that, for the five λ used, the addition of 5% of MTBE decreases exhaust NOx by 51-82%. This decrease is more important at lean than at rich conditions (63-82% against 50-51%). In the case of MTBE20 fuel, the decrease is less important, from -6 (a small increase in the case of λ ) 0.83) to 54%. There is no clear correlation between the NOx emissions of the oxygenated fuels and the ,best. synthetic one (the hexene). All four oxygenated compounds emit 2-82% less NOx at λ ) 1.25 (except the P20 one). The MTBE5 emit 20-82% less NOx at all five λ used. The other fuels emit generally more NOx than the hexene one (from 1 to 114%). Generally, the addition of 5% of one of the four oxygenated compounds tested decreases exhaust NOx more than the addition of 20% (except for the addition of ethanol, where the decrease is about the same). At stoichiometry, among the four low content oxygenated compounds used, the most effective one for the decrease of NOx is the MTBE5 (59%), followed by the P5 (31%), E5 (22%), and M5 (around 18%). At lean conditions, MTBE5 is still the most effective one (6583%), followed by P5 (33-35%), E5 (24-29%), and M5 (2426%). This classification remains unchanged at rich conditions (MTBE5: 49-51%, P5: 27-37%, E5: 15-30%, M5: 1117%). These effects are summarized in Table 2. VOL. 37, NO. 14, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3237

the obtained results were also poor. The models proposed by Jeffrey (2) were also tested but without success. No links are found between the CO, HC, and NOx emissions and the number and type of oxygen-carbon or hydrogen-carbon bonds of the fuels.

Literature Cited

FIGURE 7. Influence of fuel H/C and O/C ratios on the emission of CO and HC. All fuels used for the five λ. The above results show that the use of an oxygenated fuel must take into account the global decrease of the three pollutants; the most effective fuel for the CO reduction is P5, while the E20 one is for HC and MTBE5 for NOx. Relations between Exhaust Emissions and H/C and O/C Fuel Ratio. Literature presents a decrease of exhaust CO and HC with increasing fuel H/C fuel ratio at stoichiometry (4, 6). Figure 7 shows that CO emissions decrease very slightly with the increasing H/C and O/C fuel ratio, while the HC emissions decrease with both these ratios. No correlation is found between NOx emissions and fuel H/C or O/C ratio, even if literature presents that NOx emissions decrease with increasing H/C ratio (7). The r2 of the best fitted lines of Figure 7 decreases with increasing λ for both CO and HC in the case of fuel H/C and O/C ratio, indicating that the linearity of these correlations is better at rich conditions. The slope a of the line y ) ax + b, with y ) exhaust concentration of CO or HC and x ) fuel H/C or O/C ratio, is lower at rich conditions (a is negative), indicating that the decrease of exhaust CO or HC due to increase of fuel H/C or O/C ratio is more important at rich than at lean conditions. The change of exhaust HC is much more important than that of CO. Models. In two previous articles (19, 21), we presented several linear models linking the exhaust concentration of a number of pollutants (organic acids, aldehydes, and alcohols) with the fuel composition. The same models were constructed in the case of CO, HC, and NOx emissions, but the results were not satisfactory. The r2 of these models are very low (less than 0.7), indicating than the correlations between the exhaust concentrations of regulated pollutants and fuel composition are not linear. The links between exhaust concentration and physical properties of the fuels (octane number, distillation curves, ...) were also tested, but

3238

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 14, 2003

(1) Hochhauser, A. M.; Benson, J. D.; Burns, V.; Gorse, R. A.; Koehl, W. J.; Painter, L. J.; Rippon, B. H.; Reuter, R. M.; Rutherford, J. A. SAE Tech. Pap. Ser. 912322 1991. (2) Jeffrey, J. G.; Elliot, N. G. SAE Tech. Pap. Ser. 932680 1993. (3) Morgan, T. D. B.; den Otter, G. J.; Lange, W. W.; Doyon, J.; Barnes, J. R.; Yamashita, T. SAE Tech. Pap. Ser. 932678 1993. (4) Petit, A.; Montagne, X. SAE Tech. Pap. Ser. 932681 1993. (5) Reuter, R. M.; Benson, J. D.; Burns, V. R.; Gorse, R. A.; Hochhauser, A. M.; Koehl, W. J.; Painter, L. J.; Rippon, B. H.; Rutherford, J. A. SAE Tech. Pap. Ser. 920326 1992. (6) Neimark, A.; Kholmer, V.; Sher, E. SAE Tech. Pap. Ser. 940311 1994. (7) Lange, W. W.; Muller, A.; McArragher, J. S.; Schafer, V. SAE Tech. Pap. Ser. 841867 1994. (8) Poulopoulos, S.; Philippopoulos, C. Atmos. Environ. 2000, 34, 4781-4786. (9) Li, H.; Prabhu, S. K.; Miller, D. L.; Cernansky, N. P. SAE Tech. Pap. Ser. 950682 1995. (10) Shifter, I.; Vera, M.; Diaz, L.; Guzman, E.; Ramos, F.; LopezSalinas, E. Environ. Sci. Technol. 2001, 10, 1893-1901. (11) Noorman, M. T. SAE Tech. Pap. Ser. 932668 1993. (12) Chou, D. C.; Long, J. SAE Tech. Pap. Ser. 961221 1996. (13) DePetris, C.; Giglio, V.; Police, G.; Prati, M. V. SAE Tech. Pap. Ser. 932679 1993. (14) Bresenham, D.; Reiser, J.; Neusen, K. SAE Tech. Pap. Ser. 982054 1998. (15) Chiang, M.; Manzie, C.; Watson, H.; Palaniswami, M. SAE Tech. Pap. Ser. 2002-01-2738 2002. (16) Chan, S. H.; Zhu, J. SAE Tech. Pap. Ser. 961020 1996. (17) Kirchstetter, T. W.; Singer, B. C.; Harley, R. A.; Kendall, G. R.; Traverse, M. Environ. Sci. Technol. 1999, 33, 318-328. (18) Kirchstetter, T. W.; Singer, B. C.; Harley, R. A.; Kendall, G. R.; Chan, W. Environ. Sci. Technol. 1996, 30, 661-670. (19) Zervas, E.; Montagne, X.; Lahaye, J. Environ. Sci. Technol. 2002, 36, 2414-2421. (20) Akutsu, Y.; Toyoda, F.; Tomina, K. I.; Yoshizawa, F.; Tamura, M.; Yoshida, T. Atmos. Environ. 1991, 25A, 1383-1389. (21) Zervas, E.; Montagne, X.; Lahaye, J. Environ. Sci. Technol. 2001, 35, 2746-2751. (22) Al-Farayedhi, A. A.; Al-Dawood, A. M.; Gandhidasan, P. SAE Tech. Pap. Ser. 2000-01-2857 2000. (23) Sado, G.; Sado, M. C. Experimental Design; Ed. ANFOR, Paris, 1991 (in French). (24) Zervas, E.; Montagne, X.; Lahaye, J. J. Air Waste Manag. Assoc. 1999, 49, 1304-1314. (25) Quader, A. A. SAE Tech. Pap. Ser. 890623 1989. (26) Kisenyi, J. M.; Savage, C. A.; Simmonds, A. C. SAE Tech. Pap. Ser. 940929 1994. (27) Poulopoulos, S. G.; Samaras, D. P.; Philippopoulos, C. J. Atmos. Environ. 2001, 35, 4399-4406. (28) McDonald, C. R.; Shore, P. R.; Lee, G. R.; den Otten, J.; Humphries, D. T. SAE Tech. Pap. Ser. 941868 1994. (29) Kivi, J.; Niemi, A.; Nylund, N. O.; Kyto¨, M.; Orre, K. SAE Tech. Pap. Ser. 922379 1992.

Received for review November 12, 2002. Revised manuscript received March 21, 2003. Accepted May 12, 2003. ES026321N