Impact of Biodiesel Source Material and Chemical Structure on

the impact of biodiesel chemical structure, specifically fatty acid chain length and number of double bonds, on emissions of NOx and particulate matte...
2 downloads 0 Views 100KB Size
Download PDF
Environ. Sci. Technol. 2001, 35, 1742-1747

Impact of Biodiesel Source Material and Chemical Structure on Emissions of Criteria Pollutants from a Heavy-Duty Engine

Downloaded via CALIFORNIA INST OF TECHNOLOGY on July 14, 2018 at 13:23:05 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

ROBERT L. MCCORMICK,* MICHAEL S. GRABOSKI, TERESA L. ALLEMAN, AND ANDREW M. HERRING Colorado Institute for Fuels and Engine Research and Department of Chemical Engineering and Petroleum Refining, Colorado School of Mines, Golden, Colorado 80401-1887 K. SHAINE TYSON National Renewable Energy Laboratory, Golden, Colorado 80401

Biodiesel is an oxygenated diesel fuel made from vegetable oils and animal fats by conversion of the triglyceride fats to esters via transesterification. In this study we examined biodiesels produced from a variety of realworld feedstocks as well as pure (technical grade) fatty acid methyl and ethyl esters for emissions performance in a heavy-duty truck engine. The objective was to understand the impact of biodiesel chemical structure, specifically fatty acid chain length and number of double bonds, on emissions of NOx and particulate matter (PM). A group of seven biodiesels produced from real-world feedstocks and 14 produced from pure fatty acids were tested in a heavyduty truck engine using the U.S. heavy-duty federal test procedure (transient test). It was found that the molecular structure of biodiesel can have a substantial impact on emissions. The properties of density, cetane number, and iodine number were found to be highly correlated with one another. For neat biodiesels, PM emissions were essentially constant at about 0.07 g/bhp-h for all biodiesels as long as density was less than 0.89 g/cm3 or cetane number was greater than about 45. NOx emissions increased with increasing fuel density or decreasing fuel cetane number. Increasing the number of double bonds, quantified as iodine number, correlated with increasing emissions of NOx. Thus the increased NOx observed for some fuels cannot be explained by the NOx/PM tradeoff and is therefore not driven by thermal NO formation. For fully saturated fatty acid chains the NOx emission increased with decreasing chain length for tests using 18, 16, and 12 carbon chain molecules. Additionally, there was no significant difference in NOx or PM emissions for the methyl and ethyl esters of identical fatty acids.

Introduction Oxygenated fuels are well-known to reduce exhaust emissions from motor vehicles. In particular, methyl-tertiary-butyl ether * To whom correspondence should be addressed. E-mail: [email protected]; phone: (303) 273-3967; fax: (303) 273-3730. 1742

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 9, 2001

(MTBE) and ethanol reduce emissions from gasoline engines. In high polluting automobiles, oxygenated gasoline can decrease carbon monoxide (CO) up to 50%. Oxygenates are now mandated under the Clean Air Act for use in reformulated and CO control gasoline. This success of oxygenated gasoline has led to interest in the use of oxygenated compounds as emissions reducing additives in diesel fuel. Numerous oxygenated compounds have been investigated as either diesel fuel additives or replacements and have shown emissions reducing properties. Oxygenated fuels are also of special interest since they are a potential source of renewable energy. Biodiesel is an oxygenated diesel fuel made from vegetable oils and animal fats by conversion of the triglyceride fats to esters via transesterification. In the study described here we examined biodiesels produced from a variety of real-world feedstocks, as well as pure (technical grade) fatty acid methyl and ethyl esters, for emissions performance in a heavy-duty truck engine. The objective of the study was to understand the impact of biodiesel chemical structure, specifically fatty acid chain length and number of double bonds, on emissions of oxides of nitrogen (NOx) and particulate matter (PM). Many studies have examined emissions from biodiesel and studies through 1997 have been reviewed (1). Studies employing heavy-duty transient testing generally show that emissions of PM and CO are reduced and NOx emissions are increased relative to petroleum diesel. PM can be viewed as being made of carbon or soot, a soluble organic fraction (SOF) and sulfate. Biodiesel causes a reduction in the soot fraction of PM (1). For example, Graboski and co-workers (2) tested methyl soyester and diesel-soy ester blends in a 1991 Detroit Diesel Series 60 (four-stroke) engine. For 35% biodiesel, the composite NOx emission increased by nearly 1%, while the composite particulate emission decreased by 26% relative to the reference diesel. The NOx increase of 1% was found to be statistically significant at the 99% level. For 100% biodiesel, the composite NOx increased by 11% while PM was decreased by 66%. More recent studies have shown similar levels of PM reduction and NOx increase (3). The cause of the increased NOx emissions for biodiesel is unknown; however, a number of fuel properties have been shown to effect emissions of NOx. For example, Signer and co-workers (4) report a 3-4% increase in NOx for a 3.5% increase in fuel density using the European Economic Community 13-mode test cycle. Cetane number and fuel aromatic content are known to influence NOx and PM emissions from diesel engines (5). For biodiesel blends with diesel, the blend aromatic content is lower than that of the base diesel fuel (biodiesel contains no aromatics). This dilution of the aromatics should lower both PM and NOx emissions. Biodiesels also have increased cetane number relative to typical no. 2 diesel, which should reduce both emissions. Results suggesting that the high boiling point of biodiesel relative to the petroleum diesel boiling range is the cause of the NOx increase have also been presented (6). The studies cited above clearly indicate that substantial reduction in particulate emissions can be obtained through the addition of biodiesel to diesel fuel. There is also strong evidence that soy derived biodiesel can cause NOx emissions to increase. The objective of the current investigation is to determine the effect of biodiesel source material and ester molecular structure (number of double bonds and chain length) on PM and NOx emissions. A further objective is to examine emissions differences between methyl and ethyl esters of identical fatty acids. 10.1021/es001636t CCC: $20.00

 2001 American Chemical Society Published on Web 03/30/2001

TABLE 1. Properties of Certification Diesel Fuel Lot D-434 Used as Reference in This Study property

lot D-434

ASTM method

API gravity viscosity, cs 40 °C net BTU/lb cetane number carbon, wt % hydrogen, wt % oxygen, wt % sulfur, ppm IBP, F T50, F T90, F EP, F aromatics, vol % olefins, vol % saturates, vol %

36.28 2.5 18456 46.0 86.6 13.4 0 300 353.9 498.7 583.7 646.4 29.2 2.0 68.8

D-287 D-445 D-3338 D-613 D-5291 D-5291 D-5291 D-2622 D-86 D-86 D-86 D-86 D-1319 D-1319 D-1319

Methods The reference fuel for this test program was certification diesel obtained from Phillips Petroleum (lot D434). The properties of this fuel are shown in Table 1. Tests on the reference diesel were performed both before and after each biodiesel fuel to provide an indication of engine drift. The Institute of Gas Technology supplied seven biodiesels prepared from various feedstocks (7). Additionally, 14 fuels were prepared from pure, or nearly pure, fatty acids and from several feedstock fats. For some feedstocks both methyl and ethyl esters were prepared and the preparation of these fuels is described elsewhere (8). For pure fatty acids, the notation CXX:Y is used, where XX is the number of carbon atoms in the fatty acid chain and Y is the number of double bonds (stearic acid is C18:0, for example). The biodiesel fuels examined in this study are listed in Table 2 along with many important fuel properties. Fuel properties were measured using the specified methods either in-house or at commercial analytical laboratories. In most cases, the fuels tested met the proposed specification for biodiesel, ASTM PS121. However, in some cases, it was not possible to meet the specification. In particular, several fuels exceeded the acid number and glyceride specifications of 0.8 and 0.02/0.24 wt % (free/total glycerine),

respectively. Emissions testing of fuels that were identical except for having a high versus a low acid number, or high versus a low glyceride content, has shown that these properties have no effect on emissions of criteria pollutants (8). The system for emissions measurement for regulated pollutants (THC, CO, NOx, and PM) is identical to that described by McCormick and co-workers (6). All components and procedures meet the requirements for heavy-duty engine emissions certification testing as specified in Code of Federal Regulations Title 40, Part 86, Subpart N. The fuel system was flushed before changing from biodiesel to certification diesel. Separate fuel filters were used for biodiesel and certification diesel to minimize cross contamination. Note that the palmitate (C16:0), stearate (C18:0), and hydrogenated soy fuels have high melting points (are solid at room temperature) and were therefore difficult to work with. To test these fuels they were melted using electric heaters and then poured into an electrically heated fuel tank. Heated fuel lines were used and the engine test cell temperature was maintained at the highest obtainable temperature (37.8 °C or 100 °F). Total hydrocarbon was determined by a continuous flame ionization detector, NOx by chemiluminescence, and CO by nondispersive IR. Emission gases are 1% EPA Protocol Standards. Gas standards were not changed during this test program. All gas mass emissions are determined by background corrected flow compensated integration of the instantaneous mass rates. Tedlar bag samples of background and sample are also collected. The exhaust sample is proportionally sampled through a critical flow orifice. The bag compositions are compared with the bag equivalent flow compensated emissions to validate the test runs. Agreement is always within 5% for the individual regulated gaseous emissions. Particulate matter is collected on Pallflex T60A20 70 mm filters of a common lot. Particulate matter is sampled through a secondary tunnel that ensures a filtered gas temperature below 52 °C (126 °F). Since the PM mass collected for the biodiesel samples was small, even minor differences in filter weight due to water adsorption can impact the particulate mass emission. Particle filter handling and weighing is conducted in a yellow light, constant humidity weigh room held at 9 ( 2 °C (48 ( 4 °F) dew point, 50% nominal relative humidity and 22 ( 1 °C (72 ( 2 °F).

TABLE 2. List of Biodiesel Fuels Tested in This Study (LFFAG ) low free fatty acid grease, HFFAG ) high free fatty acid grease) fuels tested

cetane no.

methyl soy edible methyl tallow inedible methyl tallow methyl canola methyl lard methyl LFFAG methyl HFFAG (Bio3000) methyl laurate (C12:0) methyl palmitate (C16:0) methyl stearate (C18:0) ethyl stearate (C18:0) methyl oleate (C18:1) methyl linoleate (C18:2) ethyl linoleate (C18:2) methyl linolenate (C18:3) ethyl linseed methyl soy (soyagold) methyl hydrogenated soy ethyl soy ethyl hydrogenated soy 1:2 M-sterate:M-linseed a

From Kinast (7).

b

ASTM D613 47.2 62.9 61.7 55.0 63.6 57.8 52.9a 61.2 74.3b 86.9b 76.8b 56.0 41.7 44.4 45.9 43.4 52.3 47.3

density, g/cm3 ASTM D4052 0.8877a 0.8708a 0.8767a 0.8811a 0.8762a 0.8789a 0.8767a 0.8730 0.8674 0.8684 0.8636 0.8796 0.8943 0.8869 0.8941 0.8942 0.8836 0.8688 0.8817 0.8643

water/sediment, glycerol free/bound, vol % wt %

iodine no. ASTM D1510 133b 64b 64b 97b

0.3 0.5 0.5 1 90 151 140 165 157 121 6 122 6 66

ASTM D2709 0a 0.05a 0a 0a 0.6a 0a 0.21a 0 0 0.005 0 0 0 0 0 0 0 0 0 0 0

Christina Planc 0.001/0.797a 0/0.102a 0/0.159a 0.001/0.196a 0/0.160a 0/0.256a 0.010/0.064a 0/0.003 0/0.011 0/0.035 0/0.024 0/0.022 0.001/0.126 0/0.089 0/0.089 0/0.041 0.007/0.223 0.001/0.099 0.003/0.031 0/0.097 0/0.032

acid no. ASTM D664 0.32a 0.32a 0.44a 0.13a 0.76a 0.41a 0.36 0.06 0.16 1.9 0.01 0.13 0.41 0.81 0.23 2.9 0.15 4.66 3.02 3.94 1.62

oxygen, wt % ASTM D5291 11.16 11.74 11.08 11.04 11.82 11.10 11.28 14.68 11.98 11.84 10.84 11.44 11.76 11.05 11.25 11.19 11.44 11.10 11.55 6.52

From Graboski and McCormick (1).

VOL. 35, NO. 9, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1743

for PM (95% confidence intervals given). The median and mean for the cert fuel runs are very close, suggesting a normal distribution and that the majority of the variance is due to random fluctuations in the data and not a large time series effect. This is confirmed by a statistical analysis presented elsewhere (8) and indicates that no significant engine drift occurred over the course of the testing.

FIGURE 1. NOx and PM emissions results for certification fuel runs performed over the study. The engine is a 1991 calibration, Series 60 production model. The six cylinder, four stroke engine is nominally rated at 345 bhp (257 kW) at 1800 rpm and is electronically controlled (DDEC-II), direct injected, turbocharged, and intercooled. More detailed specifications for this engine were presented by McCormick and co-workers (6). Note that this is the engine model specified for certification of California diesel fuels in Title 13, California Code of Regulations, Section 2282.

Results Each biodiesel testing sequence (typically three replicate tests on a single fuel) was bracketed with at least two replicate tests on certification diesel. Figure 1 shows all of the individual hot run NOx and PM emissions from the certification diesel. The solid line represents the average emission of 4.59 ( 0.0125 g/bhp-h for NOx, and the dashed line 0.261 ( 0.014 g/bhp-h

Average emissions results for all biodiesel fuels are presented in a table available as Supporting Information, along with the calculated coefficient of variation for repeated tests. The emissions results are shown in Figure 2 for NOx and in Figure 3 for PM. The horizontal lines in these figures show the certification diesel average (in g/bhp-h) for NOx and PM. For the fuels produced from agricultural and waste feedstocks PM emissions are to a good approximation independent of feedstock. With the exception of methyl palmitate (C16:0), methyl laurate (C12:0), methyl and ethyl stearate (C18:0), and the methyl and ethyl esters of hydrogenated soybean oil, all biodiesel fuels produced higher NOx than certification diesel and all lowered PM relative to certification diesel. Thus, highly saturated fuels, those with no double bonds in the fatty acid chain, appear to have the lowest NOx emissions. Examination of the series laurate, palmitate, and stearate (C12, C16, and C18) suggests that longer chain esters have lower NOx emissions. The highest PM emissions were observed for methyl linoleate (C18:2), which was the only biodiesel with PM emissions exceeding those of certification diesel. Engine performance was very poor for this fuel. A statistical analysis of the data was performed to allow a determination as to whether observed differences in NOx or PM emissions relative to certification diesel are significant. This analysis employed a two sample t-test comparing certification fuel mean emission with biodiesel mean emission (hot start runs). The t-test tool in Microsoft Excel was used under the assumptions of equal variance, two tailed t-distribution, and hypothesized mean difference of zero. A

FIGURE 2. NOx emission results for testing of various biodiesel fuels. M ) methyl, E ) ethyl, I ) inedible, Ed ) edible, E ) ethyl, Hydro ) hydrogenated, LFFA ) low free fatty acid, HFFA ) high free fatty acid. 1744

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 9, 2001

FIGURE 3. PM emission results for testing of various biodiesel fuels. M ) methyl, E ) ethyl, I ) inedible, Ed ) edible, E ) ethyl, Hydro ) hydrogenated, LFFA ) low free fatty acid, HFFA ) high free fatty acid.

TABLE 3. Correlation Coefficients for Several Fuel Properties

cetane no. density, g/cm3 iodine no.

cetane no.

density, g/cm3

iodine no.

1 -0.897 474 5 -0.896 791 8

1 0.958 306 1

1

total of 81 certification fuel runs versus three biodiesel runs were used in each analysis for 82 degrees of freedom. For methyl ester of edible tallow the observed NOx increase is significant at the 91.5% confidence level. All other changes in NOx are highly significant (96% at a minimum with many well above 99.9%). All PM changes are highly significant. For details see Graboski and co-workers (8). The data also allow a direct comparison of methyl and ethyl esters of identical fatty acids for the following fuels: stearate, linoleate, soy, and hydrogenated soy. The Microsoft Excel paired two-sample for means t-test tool was used with a hypothesized mean difference of zero. No significant difference in NOx or PM emissions was observed between methyl and ethyl esters of the same fatty acids (p-values of 0.79 and 0.21, respectively).

Discussion The main goal of this project was to reveal the effect of fuel properties and molecular structure of biodiesel on emissions. Initially we focus on empirical correlation of fuel properties and emissions. Correlation coefficients between the various fuel properties were estimated for a group of 21 fuels for which measured or literature property values were available. These are shown in Table 3. Cetane number, density, and iodine number are all interrelated. Viscosity is weakly correlated with the other properties but is not well correlated with emissions (8). Density is the easiest and least expensive property to measure and thus is a good correlating parameter. It is important to note that the results presented here are

engine specific although other engines and calibrations will probably give similar results. This discussion is not dependent on whether pure esters or natural mixtures are considered. The correlation analysis treats mixed and pure esters simultaneously. This analysis does not include petroleum diesel, which behaves differently. Figure 4 shows a regression model for NOx emission with density. The regression is highly significant and the single parameter explains 88% of the variance. Figure 4 also shows how PM correlates poorly with density for all fuels. It is evident that there is a critical fuel density where PM dramatically increases. Below that point, the PM emission is essentially constant and seemingly independent of the biodiesel source. To investigate whether the PM emissions are constant below the critical fuel density, a regression analysis was conducted with a data set modified by deleting the three fuels with a density above 0.89. The regression analysis results show that the F-statistic is not significant. Further, the coefficients of the regression have significance levels of 0.2 and 0.3, suggesting that any variation in PM is likely due to chance. Assuming the model is significant, it predicts that PM changes 0.01 g/bhp-h for density changes from 0.86 to 0.89. This difference is less than experimental error. Thus, we conclude that an adequate model for PM emissions from biodiesel fuels assumes constant PM of 0.070 g/bhp-h, providing fuel density is less than 0.89. As Table 3 demonstrated, there is a near perfect correlation between density and cetane number for biodiesels. Cetane number can be used to discuss emissions effects and the relationship between cetane number and emissions is shown in Figure 5. The regression shown in Figure 5 indicates that a NOx neutral biodiesel, relative to certification diesel, would have a cetane number of 68. Furthermore, to get the full PM benefit, the cetane number of the biodiesel needs to exceed about 45. PM appears to be impacted only at cetane number values less than those of conventional diesel fuels today. Results of other studies, including studies of blends of biodiesel and petroleum diesel, indicate that the PM reducVOL. 35, NO. 9, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1745

FIGURE 4. Effect of density on NOx and PM emissions. tion is proportional to the fuel oxygen content (1, 8). Here we show that this is only true if cetane number is greater than about 45 or density is less than 0.89. The impact of molecular structure is implicit in either the density or cetane number. More saturated esters have higher cetane numbers and lower densities than less saturated esters. The dataset includes direct comparisons of fuels with differing numbers of double bonds in the fatty acid chain. These include both fuels prepared from pure or nearly pure fatty acids and fuels prepared from various more practical feedstocks. For fuels containing a mixture of molecules, the iodine number is a measure of the degree of unsaturation or number of double bonds. Iodine number has been measured for the pure ester fuels and many of the fuels prepared from more practical feedstocks. Iodine numbers are also available in the literature for several of the other fuels. Figure 6 shows the relationship between iodine number and emissions of NOx and PM, with emissions values for several specific fuels noted for reference. There is a highly linear relationship between iodine number and NOx, and the regression predicts that a biodiesel with an iodine number of 38 will be NOx neutral relative to certification diesel. This corresponds to an average of 1.5 double bonds/molecule. High stearate fuels with few double bonds produce significantly less NOx than certification diesel. Unfortunately, these materials have poor cold flow properties and some are even solid at room temperature. 1746

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 9, 2001

FIGURE 5. Effect of cetane number on NOx and PM emissions. Cetane number data from this study and reported in the literature (1). If all other factors are held constant (i.e., number of double bonds, etc), decreasing fatty acid chain length or molecular weight will lower boiling point and viscosity, and effect other fuel properties. One direct comparison of the impact of chain length on emissions was performed in this study. A series of saturated methyl esters based on lauric (C12:0), palmitic (C16:0), and stearic (C18:0) acids was prepared and tested. Ethyl stearate was also examined. This comparison can be seen upon inspection of Figure 2. On the basis of this evidence, we conclude that shorter chain esters produce higher NOx emissions, but note that the fully saturated methyl laurate still produced NOx at or below the certification fuel level. Therefore, shortening of the hydrocarbon chain may be a route to NOx neutral fuels with improved properties. Chain length has no significant impact on PM emissions for this small dataset. While these data clearly show the effect of biodiesel molecular structure on NOx emissions, the reason that this occurs remains open for speculation. One possibility is that the NOx increase is some manifestation of the well-known NOx/PM tradeoff. In the classic example of the tradeoff, as injection timing is advanced combustion temperature increases leading to increased NOx and reduced PM (9). However, the inspection of the plots presented in Figures 4, 5, and 6 indicates that this is not the case. While NOx emissions increase with increasing density or iodine number, or decreasing cetane number, PM emissions remain unchanged.

However, this argument cannot explain higher NOx emissions relative to petroleum diesel because of the lower heating value of biodiesel. Another possibility is that differences in the speed of sound and isentropic bulk modulus of biodiesel relative to petroleum diesel can advance the effective injection timing and thereby cause NOx to increase. Tat and Van Gerpen (11, 12) have measured these properties for the same methyl soyester and certification diesel used in this study. They find that the speed of sound and isentropic bulk modulus are 3 and 1% higher, respectively, in methyl soyester than in certification diesel. They also suggest that this could advance injection timing by as much as 1°. A timing change of this magnitude can significantly effect NOx; however, it is unlikely to be large enough to produce the increase observed for methyl soyester versus certification diesel. The important conclusion of this work is that fuel chemistry is at the root of all of these fuel properties and the increased NOx emissions observed for many biodiesel fuels. The most fundamental way to alter the emissions performance of a fuel is to alter molecular structure.

Acknowledgments This work was supported by the USDOE/National Renewable Energy Laboratory under Contract ACG-8-17106-02. The authors also wish to thank Dr. Terry Parker of the Colorado School of Mines for reading the manuscript and providing constructive comments.

Supporting Information Available Table of the average emissions data for various biodeiesels and certification fuel, hot start transient tests. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited

FIGURE 6. Effect of iodine number on emissions of NOx and PM, data for all fuels tested in this study for which iodine number was available. This clearly implies that the NOx increase, relative to petroleum diesel, that is observed for some biodiesels is not driven by the thermal or Zeldovich NOx formation mechanism. It should also be noted that iodine number is highly inverse correlated with cetane number (high iodine number correlates with low cetane number). Thus, excessive ignition delay and poor combustion performance may also be proposed as a cause of the high NOx; however, fuels such as methyl soyester exhibit increased NOx but have a substantially higher cetane number than petroleum diesel. Another possible cause of the NOx increase is that the double bonds participate in some combustion or precombustion chemistry to increase NOx. If these chemical structures result in formation of higher levels of certain hydrocarbon radicals in the fuel rich zone of the diesel spray, increased formation of so-called prompt NO might result (10). Density is also effected by number of double bonds and might also cause NOx to increase. Increasing density may increase NOx because the fuel injectors inject a constant volume, but larger mass, of the more dense fuels. Because a larger mass of fuel is burned more NOx is produced.

(1) Graboski, M. S.; McCormick, R. L. Prog. Energy Combust. Sci. 1998, 24, 125-164. (2) Graboski, M. S.; Ross, J. D.; McCormick, R. L. SAE Technol. Paper 1996, no. 961166. (3) Sharp, C. A.; Howell, S. A.; Jobe, J. SAE Technol. Paper 2000, no. 2000-01-1967. (4) Signer, M.; Heinze, P.; Mercogliano, R.; Stein, H. J. SAE Technol. Paper 1996, No. 961074. (5) Ullman, T. L.; Mason, R. L.; Montalvo, D. A. SAE Technol. Paper 1990, No. 902171. (6) McCormick, R. L.; Ross, J. D.; Graboski, M. S. Environ. Sci. Technol. 1997, 31, 1144-1150. (7) Kinast, J. A. Production and Properties of Biodiesels from Multiple Feedstocks; Final Report to USDOE/National Renewable Energy Laboratory, Contract ACG-7-15177-02, September 1999. (8) Graboski, M. S.; McCormick, R. L.; Alleman, T. L.; Herring, A. M. The Effect Of Biodiesel Composition On Engine Emissions From A DDC Series 60 Diesel Engine; Final Report to USDOE/ National Renewable Energy Laboratory, Contract ACG-8-1710602; June 8, 2000. (9) Heywood, J. B. Internal Combustion Engine Fundamentals, McGraw-Hill: New York, 1988; p 866. (10) Miller, J. A.; Bowman, C. T. Prog. Energy Combust. Sci. 1989, 15, 287-338. (11) Tat, M. E.; Van Gerpen, J. H. Measurement of Biodiesel Speed of Sound and Its Impact on Injection Timing; Final Report to USDOE/National Renewable Energy Laboratory, Contract ACG8-18066-01, May 10, 1999. (12) Tat, M. E.; Van Gerpen, J. H.; Soylu, S.; Canakci, M.; Monyen, A.; Wormley, S. J. Am. Oil Chem. Soc. 2000, 77, 285-289.

Received for review August 30, 2000. Revised manuscript received February 9, 2001. Accepted February 15, 2001. ES001636T

VOL. 35, NO. 9, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1747