Environ. Sci. Technol. 2002, 36, 2414-2421
Emission of Alcohols and Carbonyl Compounds 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 was used to study the impact of fuel composition and of the air/fuel equivalence (λ) ratio on exhaust emissions of alcohols and aldehydes/ketones. Fuel blends contained eight hydrocarbons (n-hexane, 1-hexene, cyclohexane, n-octane, isooctane, toluene, o-xylene, and ethylbenzene (ETB)) and four oxygenated compounds (methanol, ethanol, 2-propanol, and methyl tert butyl ether (MTBE)). Exhaust methanol is principally produced from fuel methanol and MTBE but also from ethanol, 2-propanol, isooctane, and hexane. Exhaust ethanol and 2-propanol are produced only from the respective fuel compounds. Exhaust formaldehyde is mainly produced from fuel methanol, acetaldehyde from fuel ethanol, and propionaldehyde from straight-chain hydrocarbons. Exhaust acroleine comes from fuel 1-hexene, acetone from 2-propanol, n-hexane, n-octane, isooctane, and MTBE. Exhaust crotonaldehyde comes from fuel 1-hexene, cyclohexane, n-hexane, and n-octane, methacroleine from fuel isooctane, and benzaldehyde from fuel aromatics. Light pollutants (C1-C2) are most likely formed from intermediate species which are quite independent of the fuel composition. An increase in λ increases the exhaust concentration of acroleine, crotonaldehyde, methacroleine, and decreases these of the three alcohols for the alcoholblended fuels. The concentration of methanol, formaldehyde, propionaldehyde, and benzaldehyde is a maximum at stoichiometry. The exhaust concentration of acetaldehyde and acetone presents a complex behavior: it increases in some cases, decreases in others, or presents a maximum at stoichiometry. The concentration of four aldehydes (formaldehyde, acetaldehyde, propionaldehyde, and benzaldehyde) is also linked with the exhaust temperature and fuel H/C ratio.
Introduction Correlations between fuel composition and exhaust emissions from spark ignition (SI) engines operating at stoichiometric conditions have been well-established in the case of regulated pollutants (refs 1-4 and many others). However, * Corresponding author phone: 331-69 27 84 77; fax: 331-69 27 81 40; e-mail:
[email protected]. Present address: Renault-CTLL26146, 1, Alle´e Cornuel, Fr-91510 Lardy, France. † Institut Franc ¸ ais du Pe´trole. ‡ Institut de Chimie des Surfaces et Interfaces. 2414
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 11, 2002
exhaust gas contains other specific pollutants such as alcohols and carbonyl compounds that have, thus far, not been thoroughly investigated. Moreover, modern SI engines are not working only under stoichiometry but also under lean and rich conditions. The influence of air/fuel equivalence ratio (λ) on the emission of specific pollutants is not very well-established. Carbonyl compounds are emitted from internal combustion engines as products of incomplete combustion of hydrocarbons or oxygenated compounds. They are finally found in the atmosphere, and many authors studied their distribution and reactions in atmosphere of urban and rural areas (refs 5-9 and many others). These pollutants have multiple sources, but motor exhaust gas is considered as one of the most important (9). The fuel composition can influence the emissions of these pollutants (2, 4, 10). Gasoline contains basically hydrocarbons, but alcohols or methyl tert butyl ether (MTBE) are also added in fuels to decrease exhaust hydrocarbons and carbon monoxide. The addition of these compounds increases the emission of some oxygenated pollutants, as methanol from fuel containing methanol or MTBE (11, 12), formaldehyde from fuel containing methanol (11), or acetone and acroleine from this containing MTBE (4, 10). No study reports the influence of λ on the emission of alcohols and carbonyl compounds. In a previous work (13), we presented the influence of fuel and of λ on the emission of organic acids of an SI engine. The main conclusions were that the first four aliphatic acids are enhanced from fuel aromatic and oxygenated compounds and are strongly depended on λ. Continuing this work, this paper presents the influence of fuel and λ on the emission of alcohols and carbonyl compounds from an SI engine. The influence of the fuel was studied by the use of specific fuel blends containing hydrocarbons and oxygenated compounds. Relations between exhaust alcohols and aldehydes/ ketones and fuel parameters or other exhaust compounds were also researched. On the basis of the obtained results, some likely formation paths of these compounds are proposed.
Experimental Section Engine and Operating Conditions. A Cooperative Fuel Research Committee (CFR) SI engine was used for these tests. This engine is a small monocylinder engine used for octane number determination. The λ used was from 0.83 to 1.25 (calculated from the exhaust gas analysis), while all other engine parameters were kept constant (13, 14). Modern engines emit less pollutants than the CFR engine, but this engine allows for the determination of the most important correlations between fuel composition and pollutants emitted. As the subject of this work was to find out the aforementioned correlations, no catalytic converter was used. Our experience shows that tailpipe emissions of the pollutants presented in this paper (except probably formaldehyde and acetaldehyde) are lower than the engine out ones because of partial oxidation on the catalytic converter. Fuels Used. Two fuel matrixes were adopted in this study. 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 ethylbenzene (ETB)), while the second one (oxygenated fuels matrix) also contains four oxygenated compounds (methanol, ethanol, 2-propanol, and MTBE). An experimental design was used to determine the quantity of each component in the blend of synthetic fuels. To avoid a high dispersion in physical properties of the fuels used, an alkylate (containing 10.1021/es010265t CCC: $22.00
2002 American Chemical Society Published on Web 04/23/2002
TABLE 1. Chemical Analysis and Octane Number of the Synthetic Fuels Useda synthetic fuels cycloisohexane hexene hexane octane octane toluene xylene ETB hexane 1-hexene cyclohexane n-octane isooctane toluene o-xylene ETB methanol ethanol 2-propanol MTBE oxygen (% wt) alkylate RON a
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
44 63.7
44 83.7
44 88.7
44 43.6
44 93.8
2 2 2 2 2 42 2 2
44 101.3
2 2 2 2 2 2 42 2
44 96.6
2 2 2 2 2 2 2 42
oxygenated fuels R
M5
E5
P5
7 7 7 7 7 7 7 7
MTBE5
6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 6.65 5 0 0 0 0 5 0 0 0 0 5 0 0 0 0 5 3.08 1.98 1.45 0.90 44 44 41.8 41.8 41.8 41.8 100 85.2 89.4 90.4 87.5 86.8
M20
E20
P20
MTBE20
5.6 5.6 5.6 5.6 5.6 5.6 5.6 5.6 5.6 5.6 5.6 5.6 5.6 5.6 5.6 5.6 5.6 5.6 5.6 5.6 5.6 5.6 5.6 5.6 20 0 0 0 20 0 0 0 20 0 0 0 11.89 7.75 5.72 35.2 35.2 35.2 98.6 97.4 96.6
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
% vol contents.
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 other fuels were also used. The first fuel is pure isooctane (iC8); the second is a mixture of 80% 2,2,4-trimethylpentane and 20% toluene (iC8T). These fuels allow for the study of the addition of an aromatic component to an alkylate basis. Finally, a commercial gasoline was also tested. More details about these fuels (chemical composition and octane number) are presented in Table 1. The distillation properties of these fuels are presented elsewhere (14). The chemical composition and physical properties of these fuels are quite different from the commercial one, but these matrixes allow for the study of the influence of the chemical composition of the fuel on the pollutants emission. Analysis of Exhaust Gas. Alcohols were collected by passing a sample of raw exhaust gas brought through a heated line through two impingers in series containing 20 mL each of deionized water. The gas flow rate was regulated at 1 L/min. The final solution was analyzed by gas chromatography using flame ionization detector (GC/FID) following the procedure described by Williams et al. (11) and Siegel et al. (15). A standard solution containing six alcohols (methanol, ethanol, n-propanol, 2-propanol, n-butanol, and 2-butanol) was used for chromatograph calibration and for the identification of each alcohol. The detection limits of this method are under 0.5 ppm in the solution (14). Only methanol, ethanol, and 2-propanol were found in the exhaust gas and were wellseparated under these analytical conditions. Carbonyl compounds were collected in 2 × 30 mL of an acidified 2,4-dinitrophenylhydrazine (DNPH) solution in acetonitrile. The gas flow rate was regulated at 1 L/min. The final solution was analyzed by high performance liquid chromatography using ultraviolet detection (HPLC/UV) following the procedure described by Lipari and Swarin (16) and Swarin et al. (17). A standard solution containing 13 aldehydes (formaldehyde, acetaldehyde, acroleine, acetone, propionaldehyde, crotonaldehyde, methacroleine, methyl ethyl ketone (MEK), butyraldehyde, benzaldehyde, and o-, m-, and p-toluenealdehyde) was used for the chromatograph calibration and for the identification of each aldehyde. The detection limits of this method were under 5 ppb in the solution. All of these carbonyl compounds were detected in the exhaust gas and had no interference with the other exhaust gas components, but the column used cannot
separate MEK and n-butyraldehyde. The three toluenaldehydes were detected in very low concentrations, so they are not presented here. Organic acids and alcohols do not react with the DNPH solution (18). More details are presented elsewhere (14). Repeatability Tests. At each λ, five identical points of the fuel R were used to evaluate the repeatability of the engine and the analytical methods. The relative standard deviation of the concentration of most carbonyl compounds and methanol (fuel R does not emit other alcohols) was found to be less than 15%. All tests were repeated, and average values were used.
Results and Discussion Percentage of Alcohols and Carbonyl Compounds in the Other Exhaust Pollutants. The quantity of each aldehyde/ ketone, as a percentage of the total carbonyl compounds emitted, is presented in Table 2 for the commercial fuel at stoichiometry (calculated on a C1 basis). The major carbonyl compound emitted is benzaldehyde (due to fuel aromatics) followed by formaldehyde and acetaldehyde. Concerning the other fuels at λ ) 1.0, these percentages vary from 25% to 47% for the formaldehyde, 6-30% for the acetaldehyde, and 1-50% for the benzaldehyde. Concerning alcohols, the E5 and E20 fuels emit 71% and 91% of ethanol and the P5 and P20 fuels 72% and 92% of 2-propanol, respectively; the rest is methanol. All other fuels emit only methanol. The emission range of total and individual alcohol, carbonyl compounds, and total organic acids and HC are also presented in Table 2. In the case of the commercial fuel at stoichiometry, methanol corresponds to 0.02% of the total HC and carbonyl compounds to 3.5%. Formaldehyde corresponds to 1.1% of the total HC, acetaldehyde to 0.3%, and benzaldehyde to 1.2%. The higher percentage of methanol is 0.6% in the case of the M20 fuel. Carbonyl compounds vary from 2.3% (xylene fuel) to 6% (E20 fuel) of total HC emitted. The λ influences these ratios (Table 2). In the case of the commercial fuel, total HC decreases with λ, while methanol, total aldehydes, and organic acids present a maximum value at stoichiometry. These values indicate that a greater important percentage of exhaust pollutants are found in oxygenated form under lean conditions (7% at λ ) 0.83, 27% at λ ) 1.0, and 55% at λ ) 1.25). Methanol. Figure 1 presents that, at stoichiometry, exhaust methanol is principally emitted from the methanolblended fuels but also from MTBE (13). Fuel ethanol and VOL. 36, NO. 11, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2415
TABLE 2. Percentage of Each Carbonyl Compound over Total Carbonyl Compounds (as C1), Range of the Emissions of Alcohols and Carbonyl Compounds Detected (for All Fuel and λ Used, in ppmv), Emission of Total Alcohols, Carbonyl Compounds, Organic Acids and HC for the Commercial Fuel for Three λ (1.25, 1.0, and 0.8, as C1), and Percentage of These Totals over the Total HC Emitted for the Commercial Fuel at Stoichiometry % carbonyl/total carbonyls (fuel ) COM, λ ) 1.0, in C1 basis)
range, all fuels, λ ) 1.0 (ppmv in a C1 basis)
% over total HC, (COM, λ ) 1.0, in a C1 basis)
COM, λ ) 1.25, (ppmv)
COM, λ ) 1.0, (ppmv)
COM, λ ) 0.83, (ppmv)
alcohols total carbonyls total organic acids total HC
0.15-3.6 21-42 35-291 457-1131
0.02 3.5 24
0.07 41 244 2100
0.21 42 276 880
0.06 28 150 230
methanol ethanol 2-propanol formaldehyde acetaldehyde propionaldehyde acetone acroleine crotonaldehyde methacroleine MEC + n-butyraldehyde benzaldehyde
0.15-3.6 0-0.30 0-3.0 8.6-16.4 2.0-11.4 0.45-0.90 1.44-11.4 1.98-4.14 0.44-0.92 0.58-2.68 0.64-1.12 0.35-18.2
0.02
31.5 9.4 1.7 6.1 8.9 1.9 2.1 2.4 36
1.1 0.3 0.07 0.23 0.34 0.07 0.08 0.09 1.2
that toluene does not participate in the formation of methanol; the second fuel emits 80% of the first one. Commercial fuel produces about the same quantity of methanol as fuel R. Exhaust methanol can come from the unburned fuel methanol or from a recombination of a CH3 and OH radical or a CH3O and H (free or abstraction of an H from an hydrocarbon). Following the first reaction path, isooctane enhances the formation of methanol probably because of the higher amount of CH3 radicals formed, confirmed by the higher amounts of methane emitted from pure isooctane than from isooctane/toluene (18). However, other fuels (such as octane or xylene) also emit higher methane concentrations than the R fuel without enhancing the formation of methanol (18). Hexane enhances the methanol formation but emits lower methane concentration than the fuel R (18). Our results cannot prove the second possible path (CH3O + H), as no relation between methanol emissions and formaldehyde emissions is found. The third reaction path (CH3O + RH w CH3OH + R) probably takes also place. The exhaust concentration of methanol is maximum at stoichiometry for all fuels used except for M5 and M20, which increase exhaust methanol at rich conditions due to unburned fuel (Figure 1). In the case of all other fuels, two hypotheses can be proposed as in the case of organic acids emissions (13): the decrease of exhaust methanol concentration at rich conditions must be due to the preferable formation of CO, and at lean conditions, the precursors of methanol must rapidly oxidized to other products. A quantitative model relating the exhaust concentration of methanol with the percentages (in volume) of the fuel components is the following (all models presented here take also into account the composition of alkylate) FIGURE 1. (Bottom graph) Fuel effect on formaldehyde emissions at stoichiometry: (mean) mean exhaust concentration; (min, max) minimum and maximum concentration for a 95% confidence level. (Middle curves) λ effect on methanol emissions for some representative fuels. (Upper curves) λ effect on ethanol and 2-propanol emissions for the E5, E20, P5, and P20 fuels. 2-propanol also contribute to the emission of exhaust methanol as the E5, E20, P5, and P20 fuels emit more methanol comparing to fuel R. Figure 1 shows that methanol is also produced from two hydrocarbons: hexane and isooctane. A comparison between iC8 and iC8T fuels shows 2416
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 11, 2002
exhaust methanol (ppmv) ) a × fuel methanol (%) + b × fuel MTBE (%) + c × fuel isooctane (%) + d × fuel hexane (%) + e × fuel ethanol (%) + f × fuel 2-propanol (%) + g. The r 2 of the line ,predicted values. ) a1 × ,experimental values. + b1 is 0.981, with a1 ) 1.0012 and b1 ) 0.0697, indicating a very good accordance between predicted and experimental values. According to this model, at stoichiometry and for equal fuel components content, the majority (51.9%) of the exhaust methanol comes from the
TABLE 3. Participation of the Different Fuel Components in the Emissions of Exhaust Pollutantsa
a
methanol
hexane
isooctane
methanol
ethanol
2-propanol
MTBE
constant
λ ) 1.25 λ ) 1.11 λ ) 1.0 λ ) 0.91 λ ) 0.83
0.58 0.32 0.11 0.06 0.04
4.06 2.92 1.18 0.34 0.25
25.91 29.15 51.85 78.35 82.72
0.41 0.92 1.56 2.70 2.97
0.36 0.72 1.15 1.87 2.04
25.91 29.15 51.85 78.35 82.71
66.46 63.38 40.93 12.02 8.03
acetaldehyde
hexane
1-hexene
n-octane
ethanol
2-propanol
λ ) 1.25 λ ) 1.11 λ ) 1.0 λ ) 0.91 λ ) 0.83
0.42 0.88 1.10 0.97 0.99
0.56 0.74 1.13 0.75 0.48
0.50 0.99 1.55 1.69 1.39
6.17 7.23 13.72 21.23 29.41
1.65 1.65 3.21 4.27 3.47
acetone
1-hexene
benzene
toluene
2-propanol
MTBE
constant
λ ) 1.25 λ ) 1.11 λ ) 1.0 λ ) 0.91 λ ) 0.83
0 0 1.50 0.49 0.24
0 0 0.86 0.51 0.22
3.21 2.22 1.21 0.93 0.96
10.93 13.94 15.19 21.82 24.95
0.24 0.51 3.67 2.32 1.91
85.61 83.30 77.54 73.90 71.69
benzaldehyde
benzene
toluene
ETB
xylene
λ ) 1.25 λ ) 1.11 λ ) 1.0 λ ) 0.91 λ ) 0.83
47.54 26.35 6.10 42.85 46.74
28.04 36.99 42.6 24.85 23.85
19.60 28.73 32.37 21.95 20.60
4.82 7.93 18.93 10.36 8.81
Percentages of this participation for the five different λ (pollutants in ppmv, fuel components in %, constant ) all other fuel components).
fuel methanol, 5.9% from MTBE, less than 2% each from ethanol, isooctane, 2-propanol, and hexane, and the rest (40.5%) from all other fuel components (Table 3). This last percentage is quite important and indicates that, as methanol is a light compound, its exhaust concentration can come from all fuel components. The percentage of the four oxygenated compounds increases at rich conditions due to unburned fuel and to their cracking to methanol, while this of the other fuel components decreases (Table 3). Ethanol and 2-Propanol. Ethanol and 2-propanol are products only of the unburned fuel (they are detected only in the case of E5, E20, P5, and P20 fuels). The absence of these alcohols in the exhaust gas of other fuels indicates that the CH3CH2 and (CH3)2CH or CH3CH2O and (CH3)2CHO radicals formed during the combustion process cannot react with OH or H to give ethanol and 2-propanol, respectively, but give smaller compounds or aldehydes. As exhaust ethanol and 2-propanol come only from the unburned fuel, their concentration increases in rich conditions (Figure 1). The models exhaust ethanol ) a × fuel ethanol and exhaust 2-propanol ) a × fuel 2-propanol can be used. The r2 of the lines ,predicted values. ) a1 × ,experimental values. + b1 is 0.99 for both compounds, with a1 ) 1.004 and 0.999 and b1 ) -0.096 and 0.009, respectively, indicating a very good accordance between predicted and experimental values. Formaldehyde. Literature reports that exhaust formaldehyde is produced from fuel methanol, ethanol, and MTBE (12, 19, 20) and also by the decrease of fuel aromatics (4) because these compounds do not participate in its formation. Our results are in accordance with those reported; Figure 2 shows that fuels containing 42% of aromatics produce less HCHO than fuel R and that iC8T fuel produces 80% of HCHO comparing to iC8. Formaldehyde is mainly produced from octane, isooctane, methanol, and MTBE. At stoichiometry, the addition of methanol to fuel R increases exhaust HCHO by 40-80%. The other three oxygenated compounds enhance the formation of formaldehyde to a lesser extent (10-40%), with little difference between low and high content of
FIGURE 2. (Bottom graph) Fuel effect on formaldehyde emissions at stoichiometry: (mean) mean exhaust concentration; (min, max) minimum and maximum concentration for a 95% confidence level. (Middle) Fuel effect on acetaldehyde emissions at stoichiometry. (Upper curves) λ effect for some representative fuels. VOL. 36, NO. 11, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2417
oxygenated compounds. Commercial fuel produces about the same quantity of formaldehyde as does the R fuel. The formation mechanism of HCHO from methanol and MTBE is already known (21-23). The formation of HCHO from isooctane must follow the same initial path as methanol (excess of CH3 radicals). As formaldehyde is the lighter aldehyde, it must have multiple sources not always correlated with initial fuel composition. The exhaust concentration of formaldehyde is maximum at stoichiometry for all fuels tested (Figure 2). The same conclusions as for methanol may be presented here: the decrease of exhaust formaldehyde at rich conditions must be due to the preferable formation of CO, while, at lean conditions, the precursors of formaldehyde must rapidly oxidized to other products. As exhaust formaldehyde has multiple sources, not always linked with initial fuel composition, no model correlating its exhaust concentration with the fuel composition has been found. Acetaldehyde. Literature presents that fuel ethanol increases acetaldehyde, that MTBE has no effect (14, 19, 24), and that a decrease of fuel aromatics increases its emission (4). Our results are in accordance with those presented: exhaust acetaldehyde is principally produced from the fuel ethanol (Figure 2). Toluene and xylene decrease exhaust acetaldehyde as compared to fuel R, and the iC8T fuel produces less (85-90%) acetaldehyde than the iC8 fuel. Exhaust acetaldehyde is also produced from the straightchain hydrocarbons n-hexane, 1-hexene, and n-octane. Methanol and MTBE do not significantly influence its emission, but 2-propanol enhances it, especially at high content (P20). Commercial fuel produces less exhaust acetaldehyde than fuel R due to higher aromatic content. The addition of ethanol to fuel R increases exhaust acetaldehyde by 80-280%, of 2-propanol by 20-70%, while the changes from the addition of methanol and MTBE are within -15 to 15%. The formation mechanism of acetaldehyde is already known (21, 22, 25). Straight-chain hydrocarbons enhance the formation of acetaldehyde by the C2 radicals produced from β scissions. Acetaldehyde concentration increases at rich conditions in the case of ethanol-blended fuels (due to oxidation of ethanol to acetaldehyde). It presents a maximum at stoichiometry in the case of fuels containing the compounds that enhance its formation (n-hexane, 1-hexene, n-octane, and 2-propanol). The reason is that, in lean conditions, acetaldehyde precursors are rapidly oxidized, and in rich conditions, they are preferably transformed to CO than to acetaldehyde due to lack of oxygen. The exhaust concentration of acetaldehyde increases at lean conditions for all other fuels. The influence of λ on the emission of acetic acid is quite similar (13). A model can correlate the exhaust concentration of acetaldehyde with the content of fuel ethanol, n-hexane, 1-hexene, octane, and 2-propanol
exhaust acetaldehyde ) a × fuel ethanol + b × fuel n-hexane + c × fuel 1-hexene + d × fuel n-octane + e × fuel 2-propanol + e The r2 of the line ,predicted falues. ) a1 × ,experimental values. + b1 is 0.98, with a1 ) 0.99 and b1 ) 0.013, indicating a very good accordance between predicted and experimental values. According to this model, at stoichiometry and for equal content of all fuel components, the 13.7% of the exhaust acetaldehyde comes from fuel ethanol, 3% from fuel 2-propanol, and 1-1.7% each from hexane, 1-hexene, and octane (Table 3). The other 79% comes from all the other fuel components (from the C2 radicals which 2418
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 11, 2002
FIGURE 3. Exhaust concentrations of acetaldehyde versus ethanol and acetone versus 2-propanol. All experimental points used (with detectable concentration of ethanol and 2-propanol). are further oxidized to acetaldehyde). The specific weight of ethanol and 2-propanol increases at rich conditions while this of the three hydrocarbons presents a maximum value at stoichiometry and this of the other sources decreases (Table 3). In the case of the fuel R, the addition of ethanol increases linearly the exhaust acetaldehyde: exhaust acetaldehyde ) a × fuel ethanol + b (with a ) 0.208 and b ) 1.6 at λ ) 1.0). The exhaust concentrations of these two compounds are also linked (Figure 3). The line exhaust concentration of acetaldehyde ) 2.51 exhaust concentration of ethanol + 1.47 indicates that the majority of fuel ethanol is oxidized to acetaldehyde and that only 1/2.5 is emitted as ethanol. Propionaldehyde. Exhaust propionaldehyde is produced from the straight-chain hydrocarbons (hexane, 1-hexene, and octane, Figure 4), and fuel aromatics decrease its exhaust concentration. The addition of oxygenated compounds does not influence exhaust propionaldehyde. However, as no decrease of this concentration occurs (because of the decrease the content of the straight-chain hydrocarbons), it is probably that oxygenated compounds facilitate the oxidation of the propionaldehyde precursors. Fuel iC8 produces a slightly higher concentration of propionaldehyde than fuel iC8T, indicating that isooctane can generate slightly more easily than toluene a C3 radical that forms propionaldehyde. Commercial fuel produces less propionaldehyde than fuel R due to higher content of aromatics. The formation reaction of this aldehyde needs a propyl radical (21). Straight-chain hydrocarbons can give this radical by the following reactions: hexane w CH3(CH2)3CHCH3 w CH3CH2CH2 + CH2dCHCH3 (26). 1-Hexene preferably produces the CH3(CH2)2CHCHdCH2 (26), but this radical is further broken to CH2dCH2 and CH2CH2CHdCH2. The formation of a CH3CH2CH2 radical needs the following direct dissociation reaction: 1-hexene w CH3CH2CH2+CH2dCHd CH2, which can also occur but is minor as compared to the previous reaction. This can explain the lower concentration of propionaldehyde emitted from the cylcohexane fuel as compared to the hexene one. Octane can give an hexyl radical by β scission and then follows the previous reactions by internal arrangements (27). Propionaldehyde presents a maximum concentration at stoichiometry for all fuels tested, obviously for the same reasons as methanol. No model correlating its exhaust
FIGURE 4. (Bottom graph) Fuel effect on propionaldehyde emissions at stoichiometry: (mean) mean exhaust concentration; (min, max) minimum and maximum concentration for a 95% confidence level. (Middle) Fuel effect on acroleine emissions at stoichiometry. (Upper curves) λ effect for some representative fuels.
FIGURE 5. (Bottom graph) Fuel effect on acetone emissions at stoichiometry: (mean) mean exhaust concentration; (min, max) minimum and maximum concentration for a 95% confidence level. (Middle) Fuel effect on crotonaldehyde emissions at stoichiometry. (Upper curves) λ effect for some representative fuels.
concentration with fuel composition is found, indicating that this pollutant has multiple sources, more than these presented here. Acroleine. Exhaust acroleine is principally produced from fuel 1-hexene (Figure 4); octane probably contributes to its formation. The addition of oxygenated compounds in the fuel R does not influence the emissions of this pollutant (a very slight decrease can be observed for the four high oxygenated content fuels), indicating that oxygenated compounds must slightly enhance the oxidation of acroleine precursors. The addition of toluene to isooctane does not influence the quantity of acroleine formed, and the commercial fuel produces about the same quantity as fuel R. As almost all fuels give about the same exhaust concentration, this pollutant must have multiple sources. The formation mechanism of acroleine must be the following. 1-Hexene first loses an H and then is broken in two parts (26), which are further oxidized to acroleine. Octane can form, after a β scission, a hexyl radical that can continue the aforesaid reactions. However, every other compound that gives a C3 radical can participate in the formation of acroleine. The exhaust concentration of acroleine increases at lean conditions. As in the case of propionaldehyde, no model correlating the exhaust concentration with the fuel composition is found. Acetone. Exhaust acetone is principally produced from fuel 2-propanol, isooctane, hexane, octane, and MTBE (4, 11, 22; Figure 5). The addition of methanol and ethanol does not decrease the exhaust acetone, indicating that they enhance the oxidation of acetone precursors. Aromatics do
not participate to its formation, as they decrease its exhaust concentration. Commercial fuel produces about the same concentration of acetone than fuel R. The formation mechanism of acetone is the following. 2-Propanol gives acetone after an extraction of an R hydrogen (22). Hexane, octane, and isooctane must first produce a secondary C3 radical which is further oxidized to CH3COHCH3. The secondary C3 radical can come after the reactions. Hexane loses a hydrogen and is then broken in two parts: CH3CH2CH2 and CH2dCHCH3 (26). The first one can then give CH3CHCH3, which is further oxidized to acetone. Octane can give, by β scission, a hexyl radical which follows the previous reactions (26). Isooctane can give, directly, a secondary C3 radical after the reaction CH3C(CH3)2CH2CH(CH3)2 w CH3C(CH3)2CH2+CH3CHCH3 (28). MTBE must give a tertio-butyl radical, which loses a methyl to give CH3CHCH3. λ influences the exhaust concentration of this pollutant in a complex way. Acetone increases with λ in the case of 2-propanol-blended fuels (due to oxidation of 2-propanol to acetone) and presents a maximum at stoichiometry in the case of the fuels hexane, octane, isooctane, MTBE5, and MTBE20, which are the fuels containing the compounds that enhance its formation. The reasons must be that, in lean conditions, the precursors of acetone are rapidly oxidized and, in rich conditions, they are preferably transformed to CO. An increase of exhaust concentration of this pollutant at lean conditions is observed in the case of all other fuels. The model exhaust acetone ) a × fuel 2-propanol + b × fuel hexane + c × fuel octane + d × fuel isooctane + e × fuel VOL. 36, NO. 11, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2419
MTBE + f is valid but only at stoichiometry and rich conditions. The r2 of the line ,predicted values. ) a1 × ,experimental values. + b1 is 0.83, with a1 ) 1.02 and b1 ) -0.028, indicating a quite good accordance between predicted and experimental values. According to this model, at stoichiometry and for equal contents of fuels components, the 15% of exhaust acetone comes from fuel 2-propanol, the 3.7% from fuel MTBE, and less than 2% each from fuel hexane, isooctane, and octane (Table 3). The rest (77%) comes from all the other fuel components. The percentage of 2-propanol increases with λ, while the contribution of all other fuel compounds decreases (Table 3). No such model is valid at lean conditions. As in the case of acetaldehyde, the addition of 2-propanol in the fuel R increases exhaust acetone linearly: exhaust acetone ) a1 × fuel 2-propanol + b1 (with a1 ) 0.158 and b1 ) 0.71 at λ ) 1.0). The exhaust concentrations of these two compounds are also linked (Figure 3). The line exhaust concentration of acetone ) 2.38 × exhaust concentration of 2-propanol + 0.85 indicates that the majority of fuel 2-propanol is oxidized to acetone and that only the 1/2.7 is emitted as alcohol. Crotonaldehyde. Crotonaldehyde is principally produced from fuel 1-hexene and cyclohexane (Figure 5). Hexane and octane contribute slightly to its formation. No significant differences are observed between iC8 and iC8T fuels, indicating that isooctane and toluene participate equivalently in its formation. The addition of oxygenated compounds in the fuel R does not influence its emission, indicating that they enhance the oxidation of its precursors. Commercial fuel produces about the same quantity as fuel R. Crotonaldehyde needs a straight chain of four carbons; hexane, 1-hexene, and octane can give a primary C4 radical by β scissions (26): hexane w CH3CH(CH2)3CH3 w CH2d CH2 + CH2(CH2)2CH3, 1-hexene w CH2dC(CH2)3CH3 w CHt CH + CH2(CH2)2CH3, octane w CH3CH(CH2)5CH3 w CH2d CH2 + CH2(CH2)4CH3 w CH2dCH2 + CH2(CH2)2CH3. The primary C4 radical must then give a CH3CHdCHCH2, which is oxidized to crotonaldehyde. In the case of cyclohexane, a ring-opening is first necessary. As cyclohexane enhances the formation of 1-hexene (14), the mechanism passes by the formation of a hexyl radical and then a C4 straight chain. The exhaust concentration of crotonaldehyde increases at lean conditions. As many fuel compounds can give C4 radicals, no model correlating the fuel composition with the exhaust concentration of this pollutant is found. Methacroleine. Exhaust methacroleine is clearly produced from fuel isooctane (Figure 6) by the formation of isobutene as an intermediate product (4, 10). The comparison between iC8 and iC8T fuels show that the latter produces about the 80% of methacroleine of the first one. Even if MTBE is suspected to enhance the formation of methacroleine (because both can produce a (CH3)2CdCH2 radical), our results show no clear tendencies between fuel MTBE and exhaust methacroleine. The addition of oxygenated compounds does not influence the exhaust concentration of this pollutant, indicating that they facilitate the oxidation of isooctane or the intermediate radicals. Commercial fuel produces a lower concentration of this pollutant as compared to the fuel R, due to lower content of isooctane. The exhaust concentration of methacroleine increases at lean conditions. The model exhaust methacroleine ) a × fuel isooctane is valid. The r2 of the line ,predicted values. ) a1 × ,experimental values. + b1 is 0.94, with a1 ) 1.047 and b1 ) -0.051, indicating good accordance between predicted and experimental values. The addition of MTBE to this model gives a poorer correlation: the r2 is only 0.69, indicating that MTBE does not participate in the formation of this pollutant. 2420
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 11, 2002
FIGURE 6. (Bottom graph) Fuel effect on methacroleine emissions at stoichiometry: (mean) mean exhaust concentration; (min, max) minimum and maximum concentration for a 95% confidence level. (Middle) Fuel effect on benzaldehyde emissions at stoichiometry. (Upper curves) λ effect for some representative fuels. n-Butyraldehyde and MEK. n-Butyraldehyde and MEK cannot be separated under the analytical conditions used. Our results show that hexane, octane, and isooctane enhance their formation, obviously by the intermediate of C4 radicals. The sum of these two pollutants shows a maximum value at stoichiometry. Benzaldehyde. Exhaust benzaldehyde is produced from fuel aromatic hydrocarbons, as already presented (4, 14). The major source is fuel toluene, followed by ETB (Figure 6). A comparison between iC8 and iC8T fuels shows that isooctane produces small quantities of this pollutant but that the majority comes from toluene. Literature presents that benzaldehyde exhaust emission decreases with the addition of oxygenates (21), but our results do not show this trend. The low oxygenates content fuels enhance slightly the formation of benzaldehyde, probably because of the enhancement of aromatics oxidation. This pollutant shows a maximum at stoichiometry for all fuels used. The model exhaust benzaldehyde ) a × fuel benzene + b × fuel toluene + c × fuel ETB + d × fuel o-xylene is valid. The r2 of the line ,predicted values. ) a1 × ,experimental values. + b1 is 0.97, with a1 ) 0.996 and b1 ) -0.007, indicating a very good accordance between predicted and experimental values. According to this model, at stoichiometry and for equal fuel compound contents, the 42.6% of the exhaust benzaldehyde comes from fuel toluene, the 33.4% from ETB, 18.5% from o-xylene, and the rest (6%) from the fuel benzene (Table 3). This indicates that, at stoichiometry, the oxidation order is toluene > ETB > o-xylene > benzene. These percentages vary significantly with λ; the first three
ratio while benzaldehyde decreases. The other aldehydes do not present such a correlation. The reason is the same as for the exhaust temperature influence: benzaldehyde is produced from fuel aromatics which have low H/C ratio; the opposite is valid in the case of the other three aldehydes. No correlation between the exhaust concentration of the pollutants detected and fuel physical properties (octane number, distillation curve, etc.) are found (using all the experimental points or even only at lean or rich conditions).
Literature Cited
FIGURE 7. Correlation between exhaust concentration of formaldehyde, acetaldehyde, propionaldehyde, and benzaldehyde, and exhaust temperature (bottom curves) and fuel H/C ratio (upper curves). All fuels used at stoichiometric conditions. compounds present a maximum at stoichiometry (Table 3), while that of benzene is at a minimum at the same point. The reason is that, as the first three aromatics oxidized easier than benzene, they give final reaction products instead of benzaldehyde at lean conditions. At rich conditions, due to lack of oxygen, the first three aromatics give benzene rather than benzaldehyde. Influence of Exhaust Temperature, H/C Fuel Ratio, and Fuel Properties. The exhaust concentration of four aldehydes is linked with exhaust temperature. The lower curves of Figure 7 show that, at stoichiometry, formaldehyde, acetaldehyde, and propionaldehyde decrease with exhaust temperature, while benzaldehyde increases. No such correlation exists for the other aldehydes. The trend of benzaldehyde is easily explained by the increased exhaust temperature of fuels containing high aromatic content. The opposite happens in the case of the three light aldehydes. As aromatics do not participate in their formation, these pollutants are produced at lower exhaust temperatures. A second possible reason is that, in higher temperatures, light aldehydes are further oxidized to CO, while benzaldehyde is more stable. A third probable reason is that the three light aldehydes are probably formed during the cooling phase, where temperature is lower. Exhaust formic acid is also linked with exhaust temperature (its concentration decreases at high temperatures; 13, 29), but no correlation is found between the exhaust concentration of formic acid and those of the three aldehydes. As exhaust methanol presents a maximum value at stoichiometry, it shows a linear correlation with exhaust temperature for every fuel used, but no such correlation exists if all fuels are used. The four previous aldehydes are also linked with the H/C ratio of the fuel used. Upper curves of Figure 7 present that, at stoichiometry, the first three aldehydes increase with this
(1) Jeffrey, J. G.; Elliot, N. G. SAE Tech. Pap. Ser. 1993, 932680. (2) Kaiser, E. W.; Siegel, W. O.; Henig, Y. I.; Anderson, R. W.; Trinker, F. H. Environ. Sci. Technol. 1991, 25, 2005-2012. (3) Neimark, A.; Kholmer, V.; Sher, E. SAE Tech. Pap. Ser. 1994, 940311. (4) Petit, A.; Montagne, X. SAE Techn. Pap. Ser. 1993, 932681. (5) Chebbi, A.; Carlier, P. Atmos. Environ. 1996, 30, 4233-4249. (6) Granby, K.; Christensen, C. S.; Lohse, C. Atmos. Environ. 1997, 31, 1403-1415. (7) Souza, S.; Vasconcellos, P. C.; Carvalho, R. F. Atmos. Environ. 1999, 33, 2563-2574. (8) Anderson, L. G.; Lanning, J. A.; Barrel, R.; Miyagishima, J.; Jones, R. H.; Wolfe, P. Atmos. Environ. 1996, 30, 2112-2123. (9) Possanzini, M.; Di Palo, V.; Petricca, M.; Fratarcangeli, R.; Brocco, D. Atmos. Environ. 1996, 30, 3757-3764. (10) Bower, S. L.; Litzinger, T. A.; Frottier, V. SAE Tech. Pap. Ser. 1993, 932707. (11) Williams, R. L.; Lipari, F.; Potter, R. A. J. Air Waste Manage. Assoc. 1990, 40, 747-756. (12) Stump, F. D.; Knapp, K. T.; Ray, W. D.; Siudak, P. D.; Snow, R. F. J. Air Waste Manage. Assoc. 1994, 44, 781-786. (13) Zervas, E.; Montagne, X.; Lahaye, J. Environ. Sci. Technol. 2001, 35, 2746-2751. (14) Zervas, E.; Montagne, X.; Lahaye, J. J. Air Waste Manage. Assoc. 1999, 49, 1304-1314. (15) Siegel, W. O.; Richert, J. F. O.; Jensen, T. E.; Schuetzle, D.; Swarin, S. J.; Loo, J. F.; Prostak, A.; Nagy, D.; Schlenker, A. M. SAE Tech. Pap. Ser. 1993, 930142. (16) Lipari, F.; Swarin, S. J. J. Chromatogr. 1982, 247, 297-306. (17) Swarin, S. J.; Loo, J. F.; Chladek, E.; Drouillard, M. S.; Tejada, S. B. SAE Tech. Pap. Ser. 1992, 920321. (18) Zervas, E., Ph.D. Dissertation, University of Mulhouse, France, 1996. (19) Reuter, R. M.; Benson, J. D.; Burns, V.; Gorse, R.; Hauchhauser, A.; Koehl, W.; Painter, L.; Rippon, B.; Rutherford, J. SAE Tech. Pap. Ser. 1992, 920326. (20) Kirchstetter, T. W.; Singer, B. C.; Harley, R. A.; Kendall, G. R.; Chan, W. Environ. Sci. Technol. 1996, 30, 661-670. (21) Warnatz, J. 20th Symposium on Combustion; 1984; pp 845-856. (22) Norton, T. S.; Dryer, F. L. 23rd Symposium on Combustion; 1990; pp 179-185. (23) Egolfopoulos, F. N.; Du, D. X.; Law, C. K. Combust. Sci. Technol. 1992, 83, 33-75. (24) Poulopoulos, S. G..; Samaras, D. P.; Philippopoulos, C. J. Atmos. Environ. 2001, 35, 4399-4406. (25) Egolfopoulos, F. N.; Du, D. X.; Law, C. K. 24th Symposium on Combustion; 1992; pp 883-841. (26) Chakir, A.; Bellimam, M.; Boettner, J. C.; Cathonnet, M. Int. J. Chem. Kinet. 1992, 24, 385-410. (27) Chakir, A.; Cathonnet, M.; Boettner, J. C.; Gaillard, F. Combust. Sci. Technol. 1989, 65, 207-230. (28) Dryer, F. L.; Brezinsky, K. Combust. Sci. Technol. 1986, 45, 199212. (29) Zervas, E.; Montagne, X.; Lahaye, J. Atmos. Environ. 2001, 35, 1301-1306.
Received for review October 17, 2001. Revised manuscript received March 4, 2002. Accepted March 6, 2002. ES010265T
VOL. 36, NO. 11, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2421
