Emissions of Acrolein and Other Aldehydes from Biodiesel-Fueled

Article pubs.acs.org/est

Emissions of Acrolein and Other Aldehydes from Biodiesel-Fueled Heavy-Duty Vehicles Thomas M. Cahill†,* and Robert A. Okamoto‡ †

Division of Mathematical and Natural Sciences, Arizona State University, West Campus, P.O. Box 37100, Phoenix, Arizona 85069, United States ‡ California Air Resources Board, 1001 “I” Street, P.O. Box 2815, Sacramento, California 95812, United States S Supporting Information *

ABSTRACT: Aldehyde emissions were measured from two heavy-duty trucks, namely 2000 and 2008 model year vehicles meeting different EPA emission standards. The tests were conducted on a chassis dynamometer and emissions were collected from a constant volume dilution tunnel. For the 2000 model year vehicle, four different fuels were tested, namely California ultralow sulfur diesel (CARB ULSD), soy biodiesel, animal biodiesel, and renewable diesel. All of the fuels were tested with simulated city and high speed cruise drive cycles. For the 2008 vehicle, only soy biodiesel and CARB ULSD fuels were tested. The research objective was to compare aldehyde emission rates between (1) the test fuels, (2) the drive cycles, and (3) the engine technologies. The results showed that soy biodiesel had the highest acrolein emission rates while the renewable diesel showed the lowest. The drive cycle also affected emission rates with the cruise drive cycle having lower emissions than the urban drive cycle. Lastly, the newer vehicle with the diesel particulate filter had greatly reduced carbonyl emissions compared to the other vehicles, thus demonstrating that the engine technology had a greater influence on emission rates than the fuels.

■

INTRODUCTION There has been considerable interest in developing biofuels to reduce dependence on fossil fuels and reduce the global warming impact of transportation fuels. Biodiesel and renewable diesel have emerged as biofuels that can be used with existing diesel engine technology with little or no modifications. Biodiesel fills a niche in the heavy-duty vehicle fleet where alternative vehicles and fuels are scarce with the notable exception of compressed natural gas. Biodiesel is generally produced from plant oil or animal fat feedstocks where the fatty acids are transesterified with methanol to create fatty acid methyl esters (FAME). The result is a fuel that retains some characteristics of the fatty acids from which they were derived. The fuel has many of the same physical properties as petroleum diesel, hence it is adaptable to existing engine technology. Renewable diesel is also created from biological fatty acids, but the fatty acids are converted to hydrocarbons by a hydrotreating process. As with any new technology, it is important to assess potential impacts of new fuels before they are widely adopted in case there are any unexpected negative effects. The main concern of biodiesel is a potential increase in NOx emissions.1 However, other pollutants may also increase during the vehicle fleet conversion from petroleum diesel to biodiesel. The biodiesels retain many unsaturated sites in the fatty acids which may result in different combustion chemistry. In particular, emissions of unsaturated aldehydes are likely to increase in © 2012 American Chemical Society

biodiesel compared to petroleum diesel. Heated cooking oils, which share most of the same fatty acids as biodiesel, emit large amounts of acrolein (2-propenal),2−4 although these emission rates may not continue at the higher combustion temperatures in diesel engines. The different biodiesel formulations need to be characterized under the same experimental protocols to identify which formulations have the lowest emission rates. It is also important to test these biofuels against diesel fuel on vehicles meeting different emissions standards to determine emission differences within each engine technology group. Lastly, the biodiesel emission tests need to be conducted under realistic driving conditions, such as stop-and-go city traffic and high speed expressway travel. Biodiesel and biodiesel blends have been studied extensively in recent years. Carbonyl data are rarely reported,5−7 although some engine dynamometer reports give detailed carbonyl data (e.g., refs 8−13). In particular, some of the main carbonyls of concern, such as acrolein, are either not reported or the studies utilized analytical methods (e.g., DNPH) that are known to be problematic for acrolein.14 The unsaturated aldehydes are precisely the type of chemicals that are expected to increase in Received: Revised: Accepted: Published: 8382

April 26, 2012 June 25, 2012 July 2, 2012 July 2, 2012 dx.doi.org/10.1021/es301659u | Environ. Sci. Technol. 2012, 46, 8382−8388

Environmental Science & Technology

Article

Table 1. Recovery (%) of Labeled Standards Added to Sample Collection Matrix for the Different Sampling Episodes Labeled by the Biodiesel Type and Engine Being Tested labeled standard

soy (C-15) (n = 16)

animal (C-15) (n = 20)

renewable (C-15) (n = 22)

soy (MBE 4000) (n = 26)

gas-phase samplesa acetaldehyde-d4 acrolein-d4 glyoxal-d2 benzaldehyde-d6

4.8 96.3 93.4 96.0

5.1 91.2 103 97.2

10.4 138 208 86.4

20.3 416 899 82.3

filter samplesb acetaldehyde-d4 acrolein-d4 glyoxal-d2 benzaldehyde-d6

102 106 128 105

100 89.7 103 97.2

84.1 95.6 60.1 88.4

121 124 94.7 90.0

a

The gas-phase samples were spiked with the labeled standards prior to sample collection. bThe filter samples were spiked after sample collection to avoid blow-off into the gas-phase samples.

biodiesel due to the unsaturated sites in the fatty acids. To address these information gaps, CARB conducted a major study to specifically compare the emissions impacts of biofuels use in California, with a focus on both NOx and toxic emissions.15 In this study, the aldehyde emissions from heavy-duty biodiesel fueled trucks were determined using a chassis dynamometer to simulate realistic driving conditions. Four different fuels were tested, namely CARB ULSD (which was the control petroleum fuel), soy biodiesel, animal biodiesel, and renewable diesel. The biodiesel and renewable diesel fuels were tested as pure fuels and as a 50% blend with the CARB ULSD. To simulate emissions under realistic conditions, all of the fuels were tested with both city stop-and-go and cruise drive cycles using the model year 2000 truck. Lastly, one newer 2008 year vehicle with a diesel oxidation catalyst/diesel particulate filter (DOC/ DPF) was tested with the soy biodiesel and CARB ULSD to assess the emission differences between these two fuels when used in a truck with newer emission control technology.

Figure 1. Emission rates (μg/km) of three representative carbonyls during the “UDDS” drive cycle with the 2000 model year Caterpillar C-15 and the 2008 model year MBE4000 engines. The error bars represent the standard deviation of emission rates from separate chassis dynamometer runs (n = 3 for ULSD and MBE4000 trials and n = 2 for all other trials).

■

MATERIALS AND METHODS The emission sampling was conducted using a heavy-duty chassis dynamometer at the California Air Resources Board’s Heavy-Duty Emissions Test Facility in Los Angeles.15,16 Briefly, the truck was strapped to the chassis with a total inertial weight of between 19,050 kg or 26,300 kg depending on the drive cycle. The exhaust stack of the truck was connected to primary dilution tunnel operated in a constant flow mode. The dilution air was HEPA filtered and charcoal purified prior to entering the dilution tunnel. Sampling ports collected air from the dilution tunnel for various physical and chemical analyses. For the aldehyde analysis, a second sampling port was connected to the dilution air source downstream of the charcoal trap to obtain the background concentrations of carbonyls entering the dilution tunnel. The emission samples were then “blank subtracted” to determine the carbonyls arising from the vehicle emissions. The blank subtraction was weighted by the volume of the dilution air in the emission samples. Four fuels were the focus of the experiment. The first was the CARB ULSD, which served as the petroleum fuel to which the biodiesel and renewable diesel fuels were compared. A soybeanbased biodiesel was the second fuel while an animal-based

biodiesel was the third fuel (donated by the National Biodiesel Board, Jefferson City, MO). Both the soybean and animal biodiesel fuels were largely FAME formulations. The last fuel was NExBTL renewable diesel (Neste Oil, Espoo, Finland), which is different from biodiesel formulas since the feedstock is hydrotreated to remove oxygen from the fatty acids to produce a pure hydrocarbon fuel.15,17 The basic fuel characteristics are presented in Table S1 in the Supporting Information. The fuels were tested as either pure (100%) test fuels or as a 50% blend with the CARB ULSD fuel. Each sampling campaign focused on one of the biofuels. To avoid systematic biases, the test drives were interspersed (e.g., first ULSD, first 50% blend, first 100% bio, second ULSD, second 50% blend, etc.). The vehicle was reconditioned with a cruise drive cycle after each fuel change to minimize memory effects in the engine. Two trucks were tested in this study. The first vehicle was a 2000 Freightliner Truck equipped with a 2000 Caterpillar C-15 engine which met EPA 2000 year standards. This truck lacked any post-combustion emission control measures. This vehicle was tested with all four test fuels. The second vehicle was a 8383

dx.doi.org/10.1021/es301659u | Environ. Sci. Technol. 2012, 46, 8382−8388

Environmental Science & Technology

Article

Table 2. UDDS Emission (mean ± SD) from the C-15 Vehicle in μg/kma ULSD (n = 3)

soy 50% (n = 2)

soy 100% (n = 2)

animal 50% (n = 2)

animal 100% (n = 2)

renew. 50% (n = 2)

renew. 100% (n = 2)

ND 7.4 ± 8.1b not reportedc not reportedc 23.7 ± 15.2 22.1 ± 19.4b 20.3 ± 18.0b

ND 21.5 ± 3.5 not reportedc not reportedc 12.9 ± 1.9 34.4 ± 9.3 ND

ND 15.2 ± 1.7 not reportedc not reportedc 14.5 ± 2.9 37.8 ± 3.6 14.0 ± 19.8

ND 14.1 ± 1.2 not reportedc not reportedc 20.5 ± 7.3 25.9 ± 36.5 ND

ND 12.2 ± 2.4 not reportedc not reportedc 19.0 ± 5.5 53.9 ± 7.0 15.5 ± 22.0a

ND ND not reportedc not reportedc 32.3 ± 14.7 41.0 ± 18.0 60.2 ± 3.2

ND 2.0 ± 2.8b not reportedc not reportedc 31.1 ± 7.7 36.3 ± 7.6 43.4 ± 61.4b

unsaturated aldehydes acrolein (2-propenal) methacrolein (2-methyl-2-propenal) crotonaldehyde (2-butenal) 2-methyl-2-butenal 3-methyl-2-butenal 2-hexenal 2,4-hexadienal 2-heptenal 2,4-heptadienal 4-decenal

83.9 ± 23.3 6.5 ± 1.9 41.2 ± 4.7 5.5 ± 2.2 13.0 ± 3.2 2.5 ± 2.2b 24.1 ± 12.6 6.2 ± 2.8 3.5 ± 4.4b 10.1 ± 9.0b

248 ± 12.8 13.4 ± 0.3 100 ± 11.4 10.8 ± 1.7 19.6 ± 1.4 10.3 ± 1.6 820 ± 7.0 12.7 ± 3.1 ND ND

191 ± 45.9 6.7 ± 1.3 65.9 ± 13 7.7 ± 2.1 11.6 ± 2.2 4.3 ± 6.1b 59.1 ± 9.6 11.3 ± 4.3 ND ND

132 ± 43.8 7.7 ± 2.2 53.8 ± 11.8 5.3 ± 1.3 12.4 ± 3.8 ND 28.2 ± 5.8 6.0 ± 2.3 9.2 ± 1.9 14.4 ± 2.3

108 ± 3.4 3.9 ± 0.1 37.4 ± 5.2 3.5 ± 0.5 7.6 ± 0.2 4.6 ± 0.2 22.4 ± 1.6 5.3 ± 0.2 8.1 ± 0.3 5.6 ± 7.9b

61.7 ± 7.0 4.5 ± 0.3 45.3 ± 2.2 4.5 ± 0.2 21.9 ± 0.4 5.6 ± 0.3 18.4 ± 1.7 2.4 ± 0.1 2.2 ± 0.6 2.9 ± 4.0

72.1 ± 10.6 4.6 ± 0.7 40.8 ± 6.3 4.4 ± 0.3 21.7 ± 2.7 4.9 ± 1.0 15.2 ± 2.7 4.5 ± 2.6 2.0 ± 0.5 17.4 ± 2.1

aromatic aldehydes benzaldehyde o,m-tolualdehyde p-tolualdehyde 2-ethylbenzaldehyde 3,4-dimethylbenzaldehyde 4-methoxybenzaldehyde 1-naphthaldehyde

57.0 ± 9.1 15.6 ± 3.3 14.2 ± 2.5 3.5 ± 1.9 9.3 ± 2.8 1.6 ± 1.4b 34.1 ± 10.9

79.6 ± 7.8 29.2 ± 2.7 25.4 ± 3.3 4.1 ± 0.2 16.1 ± 2.7 3.1 ± 0.3 34.5 ± 1.9

46.4 ± 15.8 12.7 ± 2.7 12.1 ± 2.6 2.5 ± 0.5 9.2 ± 1.9 2.4 ± 0.7 27.8 ± 0.6

72.7 ± 21.8 18.0 ± 4.2 16.2 ± 3.3 5.3 ± 1.0 10.0 ± 2.1 4.4 ± 0.9 18.9 ± 6.3

34.5 ± 1.5 7.3 ± 0.6 7.3 ± 0.6 4.0 ± 0.8 5.5 ± 0.9 1.6 ± 2.2a 12.2 ± 0.1

40.4 ± 7.5 12.3 ± 0.2 12.6 ± 0.2 1.9 ± 0.2 6.5 ± 0.1 2.7 ± 0.4 32.5 ± 1.9

15.2 ± 21.6b 6.0 ± 0.9 6.8 ± 0.7 1.1 ± 0.1 2.8 ± 0.2 2.3 ± 0.3 9.9 ± 1.1

misc. carbonyls glyoxal (ethanedial) methyl glyoxal (2-oxopropanal)

5310 ± 2050 1030 ± 304

6160 ± 371 1360 ± 83.3

5750 ± 2060 1260 ± 11.7

5510 ± 357 1080 ± 22.7

5760 ± 357 870 ± 10.6

2630 ± 53.6 677 ± 31.1

3050 ± 74.6 777 ± 17.6

saturated aldehydes butanal pentanal hexanal heptanal octanal nonanal decanal

The average fuel economy for these tests was 2.83 ± 0.095 km/L, so emission rates in units of “μg/L of fuel burned” can be estimated by multiplying the table values by 2.83. bAverage contains one replicate that was not detected. The nondetected sample was treated as having a value of zero for both the average and standard deviation. cHexanal and heptanal are not reported due to high reagent blanks.

a

UDDS and cruise drive cycles, respectively. The fuel economy of the 2008 model year vehicle was 2.31 ± 0.096 km/L for the UDDS drive cycle.15 Aldehyde samples were collected using paired mist chambers using procedures as described in refs 18,19 (Standard Operating Procedure provided in the Supporting Information). In short, the method pulls air through two mist chambers containing a 0.1 M bisulfite solution that traps the aldehydes as sulfonate adducts. Isotopically labeled standards are added to the mist chambers before sample collection to account for any sample blow-off, incomplete derivatization or analyte adhesion to glassware. After sample collection, the sulfite solution is added to a test tube containing hydrogen peroxide, pentafluorobenzylhydroxyl amine (PFBHA), and hexane to derivatize the carbonyls. After four days of derivatization time, the analytes were extracted with hexane and quantified using a Agilent 6890N gas chromatograph coupled with an Agilent 5973 mass selective detector. The GC was equipped with a DB-5 fused silica capillary column (30 m, 0.25 mm i.d., 0.25 μm film thickness). The detector was operated in the negative chemical ionization mode with a methane reagent gas. The only modification to the standard protocol was to add a 47 mm Zeflor filter upstream of the mist chambers to keep particles

2008 Freightliner Truck equipped with a MBE4000 engine that met EPA 2007 year standards. This vehicle was equipped with a DOC/DPF to decrease emissions. This vehicle was only tested with soy biodiesel blends and CARB ULSD. The DPF was actively regenerated once per day after a fuel change to reduce carryover between sample runs and to ensure the DPF would not regenerate during a test run. This is a more frequent DPF regeneration rate than in normal vehicle operation, so the DPF should have been well within operational specifications during our tests. Two different drive cycles were tested in this experiment, namely the Urban Dynamometer Driving Schedule (UDDS) and the CARB Heavy, Heavy-Duty Diesel Truck 50 mph cruise cycle (hereafter “cruise”). The UDDS simulates urban driving with many starts and stops. This cycle lasts 17.7 min and simulates driving 8.94 km at an average speed of 30.3 km/h The cruise drive cycle has a high, consistent speed (88.6 km/h) for most of the 12.6 min drive cycle, resulting in a distance of 16.9 km at an average speed of 78.7 km/h The total vehicle weights, namely vehicle and simulated cargo, for the UDDS and cruise cycles were ∼19 050 and ∼26 300 kg, respectively.15 The average fuel economy for the 2000 model year vehicle across all fuels tested was 2.83 ± 0.095 and 3.00 ± 0.074 km/L for the 8384

dx.doi.org/10.1021/es301659u | Environ. Sci. Technol. 2012, 46, 8382−8388

Environmental Science & Technology

Article

from fouling the mist chambers. The filter was also analyzed for carbonyls by adding the filter to the derivatization test tube along with 20 mL of the bisulfite solution to keep the sulfite/ H2O2 reaction balanced. The results presented herein are the sum of the gas and particulate phases, although most of the smaller aldehydes were predominately detected in the gas-phase.

Table 3. UDDS Emission (mean ± SD) from the MBE4000 Vehicle in μg/km ULSD (n = 3) saturated aldehydes butanal pentanal hexanal heptanal octanal nonanal decanal

■

RESULTS Quality Control. The quality control program using the labeled carbonyls showed that the analytical method was generally effective for most of the carbonyls except for acetaldehyde. (Table 1) Acetaldehyde was effectively derivatized, extracted, and quantified in blanks, the calibration curves and spiked filter samples, so the low recovery of acetaldehyde from the gas-phase samples was most likely due to blow-off during sample collection. Therefore, acetaldehyde is not reported for this study due to the uncertainty of the acetaldehyde measurements even though it was the single most abundant chemical detected. The peculiar and interesting result was the enhanced response of acrolein-d4 and glyoxal-d2 in the gas-phase soy MBE4000 tests and, to a much lesser extent, in the renewable diesel tests. This enrichment was not observed in the filter samples or the field blanks (gas and filter) processed at the same time. Since the labeled spike was added to the mist chamber solution before sample collection, the labeled chemicals contacted the sample air. In contrast, the filter samples were spiked after sample collection to avoid blow-off into the gas collection system. The elevated recovery of these two compounds was observed in both the dilution air and the exhaust air samples, but not the field blanks. Therefore, it would appear that a component of the dilution air or sampling configuration caused an increase in the derivatization and/or partitioning equilibrium in these samples. The most likely explanation was contamination from the DNPH cartridges that were added to the sampling system during these later runs as a “T” branch off of the sampling line after the filter but before the mist chambers. If the vacuum pump on the mist chambers was stronger than the DNPH cartridge pump (which it is believed to be), then there was potential to back-flush DNPH into the mist chambers and that would change the trapping and/or derivatization conditions. It is worth noting that benzaldehyde-d6 was not affected, but it is the least sensitive chemical to changing derivatization conditions.18 Subsequent tests with newly purchased standards and reagents verified the accuracy of the standards used in the MBE4000 tests. Ultimately, the concept behind using an isotopic standard for quantification of analytes (isotope dilution method) is that any process that affects the labeled standard should affect the unlabeled analyte at the same rate. Thus, the calculation of the relative response factor causes these systematic biases to cancel out and the results should still be valid. Therefore, the results from the MBE4000 tests and renewable diesel gas phase should still be accurate, but the excessive recovery of acrolein-d4 and glyoxal-d2 indicated that some aspect of the sampling system was behaving in an unexpected fashion. Comparison of Different Fuels on the 2000 Model Year Caterpillar C-15 Engine. The carbonyl results showed that the different fuel blends gave different carbonyl emissions for the UDDS drive cycle. The first observation was that acrolein emissions were higher for soy and animal blends than the ULSD fuel (Figure 1, Table 2). This was largely expected since plant and animal fatty acids contain more unsaturated sites that petroleum alkanes, so they would be expected to generate more unsaturated aldehydes. Typically, the acrolein

unsaturated aldehydes acrolein methacrolein crotonaldehyde 2-methyl-2-butenal 3-methyl-2-butenal 2-hexenal 2,4-hexadienal 2-heptenal 2,4-heptadienal 4-decenal

soy 50% (n = 3)

ND ND ND ND not reporteda not reporteda not reporteda not reporteda ND 6.96 ± 6.03b ND 8.88 ± 8.63b b 17.3 ± 16.6 33.7 ± 35.3b

soy 100% (n = 3) ND ND not reporteda not reporteda 8.14 ± 7.27 8.7 ± 5.23b 87.6 ± 59.3

39.7 ± 13.4 0.32 ± 0.16 1.44 ± 0.30 0.14 ± 0.09 ND ND ND ND ND ND

27.9 ± 1.76 ND ND ND ND ND ND ND 0.12 ± 0.19b ND

65.8 ± 16.8 ND 1.69 ± 1.00 0.07 ± 0.07b ND ND ND ND 0.14 ± 0.15b 0.44 ± 0.41b

aromatic aldehydes benzaldehyde 2.56 ± 1.63 o,m-tolualdehyde 0.22 ± 0.08 p-tolualdehyde ND 2-ethylbenzaldehyde ND 3,4-dimethylbenzaldehyde 0.09 ± 0.04 4-methoxybenzaldehyde 0.14 ± 0.09 1-naphthaldehyde 0.35 ± 0.44b

1.31 ± 0.61 0.12 ± 0.08 ND ND 0.07 ± 0.1 0.03 ± 0.02b ND

1.4 ± 0.57 ND ND ND 0.03 ± 0.04b ND ND

ND ND

ND ND

misc. carbonyls glyoxal methylglyoxal

ND ND

a

Hexanal and heptanal are not reported due to high reagent blanks. Average contains one replicate that was not detected. The nondetected sample was treated as having a value of zero for both the average and standard deviation.

b

emissions from the soy blends were double that of the ULSD fuel. The animal-based biodiesel had acrolein emission rates that were approximately 25% to 50% higher than the ULSD fuel. This is supported by the analysis of the soy and animal biodiesel fuels which show 85% of the soy biodiesel fatty acid esters to be unsaturated as compared to 46% for the animal biodiesel. In addition, the animal biodiesel has only 42% as many double bonds as the soy biodiesel.20 The renewable diesel did not show any increase in acrolein emissions when compared to the ULSD. It is worth noting that the renewable diesel largely consists of saturated alkanes,15 so it was expected to have similar acrolein emissions compared to standard diesel. One observation was the intermediate fuel blends (e.g., 50% soy) tended to have slightly higher concentrations of the carbonyls compared to the two individual fuels that comprised the blend. The aromatic aldehydes showed a different emission pattern than acrolein (Figure 1). In general, the concentrations of aromatic aldehydes were lower in the pure biodiesel fuels compared to the ULSD fuel. As a group, the aromatic aldehydes are likely to arise from the oxidation of aromatic 8385

dx.doi.org/10.1021/es301659u | Environ. Sci. Technol. 2012, 46, 8382−8388

Environmental Science & Technology

Article

Table 4. Cruise Emission (mean ± SD) from the C-15 Vehicle in μg/kma ULSD (n = 3)

soy 50%b (n = 0)

soy 100% (n = 2)

animal 50% (n = 2)

animal 100% (n = 2)

renew. 50% (n = 2)

renew. 100% (n = 2)

ND 3.5 ± 3.0c not reportedd not reportedd 7.3 ± 0.6 14.3 ± 13.2c ND

ND 7.3 ± 0.6 not reportedd not reportedd 6.6 ± 0.6 18.7 ± 5.3 ND

ND 10.0 ± 5.2 not reportedd not reportedd 7.8 ± 0.2 ND ND

ND 8.4 ± 0.1 not reportedd not reportedd 9.3 ± 1.0 ND ND

ND 8.0 ± 3.9 not reportedd not reportedd 15.6 ± 0.5 39.8 ± 34.5 5.4 ± 7.6c

ND 9.3 ± 1.2 not reportedd not reportedd 10.6 ± 5.0 18.8 ± 1.4 17.7 ± 25.0c

unsaturated aldehydes acrolein methacrolein crotonaldehyde 2-methyl-2-butenal 3-methyl-2-butenal 2-hexenal 2,4-hexadienal 2-heptenal 2,4-heptadienal 4-decenal

62.0 ± 22.3 4.4 ± 1.5 24.2 ± 5.8 4.0 ± 2.2 7.5 ± 1.6 3.5 ± 1.1 12.2 ± 6.3 3.1 ± 1.7 3.2 ± 3.0 ND

99.4 ± 15.3 4.3 ± 1.4 34.3 ± 2.5 3.9 ± 0.7 5.7 ± 0.6 1.6 ± 2.2c 30.0 ± 0.5 5.5 ± 0.1 ND ND

102 ± 23.5 6.1 ± 0.7 37.1 ± 3.2 4.3 ± 0.4 8.2 ± 1.2 4.2 ± 0.5 15.8 ± 0.4 4.5 ± 0.1 7.1 ± 0.7 6.5 ± 9.1

77.6 ± 11.6 3.4 ± 0.4 25.9 ± 2.7 3.5 ± 0.1 5.8 ± 0.3 4.5 ± 0.8 14.9 ± 0.3 5.1 ± 0.2 7.4 ± 0.4 5.7 ± 8.1c

50.9 ± 1.7 4.0 ± 0.5 28.7 ± 2.0 2.6 ± 0.8 10.9 ± 3.8 2.5 ± 0.7 7.8 ± 3.8 1.6 ± 1.2 1.4 ± 0.4 1.1 ± 0.3

70.8 ± 6.1 4.2 ± 0.6 31.6 ± 3.2 3.5 ± 0.8 13.4 ± 1.1 4.5 ± 0.6 8.6 ± 0.2 2.5 ± 0.3 1.8 ± 0.1 2.5 ± 0.8

aromatic aldehydes benzaldehyde o,m-tolualdehyde p-tolualdehyde 2-ethylbenzaldehyde 3,4-dimethylbenzaldehyde 4-methoxybenzaldehyde 1-naphthaldehyde

34.3 ± 6.6 9.9 ± 2.5 8.7 ± 2.1 2.1 ± 1.4 5.7 ± 2.5 2.2 ± 0.7 14.2 ± 1.1

24.5 ± 6.1 6.5 ± 0.2 6.1 ± 0.2 1.1 ± 0.4 4.3 ± 1.4 1.3 ± 0.2 11.9 ± 4.7

57.5 ± 6.4 10.9 ± 0.5 9.8 ± 0.2 4.0 ± 0.4 6.4 ± 0.4 1.3 ± 1.9c 9.6 ± 2.4

30.7 ± 4.2 5.3 ± 0.2 5.7 ± 0.1 3.9 ± 0.1 4.7 ± 0.1 2.9 ± 0.1 8.4 ± 1.7

18.4 ± 1.1 5.8 ± 2.0 5.5 ± 2.0 0.7 ± 0.2 2.6 ± 1.0 1.0 ± 0.2 14.2 ± 5.0

22.2 ± 8.2 5.2 ± 0.7 5.5 ± 1.3 1.1 ± 0.4 2.0 ± 0.6 1.8 ± 0.7 7.8 ± 3.4

misc. carbonyls glyoxal methylglyoxal

1840 ± 565 462 ± 149

2940 ± 1480 491 ± 206

444 ± 35.8 235 ± 3.7

1110 ± 870 277 ± 171

512 ± 273 223 ± 5.5

733 ± 311 260 ± 47.0

saturated aldehydes butanal pentanal hexanal heptanal octanal nonanal decanal

The average fuel economy for these tests was 3.00 ± 0.074 km/L, so emission rates in units of “μg/L of fuel burned” can be estimated by multiplying the table values by 3.00. bNo samples of the Soy 50 cruise were collected. cAverage contains one replicate that was not detected. The nondetected sample was treated as having a value of zero for both the average and standard deviation. dHexanal and heptanal are not reported due to high reagent blanks. a

components in the fuel and/or lubrication oil, so fuels with lower aromatic fractions would be expected to have lower emissions of the aromatic aldehydes. Petroleum fuels can have significant aromatic hydrocarbon fractions while the biodiesels have considerably lower aromatic content. The renewable diesel had the lowest aromatic aldehyde emissions since the fuel had a very low aromatic fraction in the fuel of 0.4% compared to CARB ULSD of 18.7% 15 (Table S1 in Supporting Information). In all cases, the blends had higher aromatic aldehyde emissions compared to the 100% biofuel. The most abundant aldehydes quantified in this analysis, which excludes formaldehyde and acetaldehyde, were glyoxal and methyl glyoxal. These two highly oxidized compounds showed effectively no differences between ULSD, soy and animal biodiesel, while the renewable diesel had emission rates that were approximately half that of the other three fuels. Unlike the prior compounds, the 50% blends had about the same emissions as the 100% biofuels. Comparison with a DPF-Equipped Engine Technology with Soy-Based Biodiesel. Two vehicles with different engine technologies were tested with the soy biodiesel and ULSD fuels and blends. The model year 2000 Caterpillar C-15

engine did not have post-combustion emission controls while the model year 2008 MBE4000 engine was equipped with EGR and DOC/DPF. The aldehyde emission rates from the MBE4000 were considerably lower than the Caterpillar C-15 for all fuels tested. (Table 3 and Figure1) and only a few compounds were detected at concentrations greater than the background dilution air. The DOC/DPF appears very effective at lowering carbonyl emissions. In contrast to the 2000 vehicle results, the intermediate fuel blend had slightly lower emissions of acrolein compared to either the ULSD or soy 100 fuels (Figure 1). However, the sample size is too small to be conclusive. Overall, the data clearly demonstrates that the influence of engine technology greatly exceeds the influence of biodiesel fuels for aldehyde emission rates. Comparison between Drive Cycles. Emissions from the four test fuels were measured on two drive cycles, namely the UDDS and the cruise drive cycles. The emissions, quantified as μg/km, from the cruise drive cycle were lower than the UDDS drive cycle for all test fuels (Table 4, Figure 2). The lower emissions are probably due to the consistent high speed of the cruise compared to the start−stop−idle of the UDDS drive cycle. The general trends of carbonyl emissions between the 8386

dx.doi.org/10.1021/es301659u | Environ. Sci. Technol. 2012, 46, 8382−8388

Environmental Science & Technology

Article

test fuels for the cruise were similar to the UDDS results. The soy biodiesel tended to have higher emissions and the renewable diesel had the lowest emissions. The intermediate blends were more similar to the pure components of the blend than was observed in the UDDS tests.

■

DISCUSSION Biodiesel has been promoted as a means to reduce the global warming impact of transportation fuels, but it needs to be thoroughly evaluated to ensure there are not any unwanted side effects. The main concern of widespread use of biodiesel is increased NOx emissions, which was the primary focus of the overarching CARB project.15 However, there is also interest in acrolein since acrolein often ranks high among the list of ambient hazardous air pollutants. Acrolein is generally perceived as having a major vehicular source as a both a primary pollutant and as a secondary oxidation product of vehicle emissions. This study tested emissions from three different biodiesel fuels in addition to the typical ULSD fuel. The results showed acrolein emissions were higher for the soy biodiesel formulations compared to the other fuels, which was expected since vegetable oils are known to emit acrolein when heated or partly combusted.2−4 The other fuels were fairly comparable in terms of acrolein emissions. In general, most of the fuels were fairly comparable in terms of carbonyl emissions with differences rarely exceeding a factor of 2. The renewable diesel generally had lower overall emissions than the other fuels. While the different fuels showed modest differences in emission rates, the comparison of engine technology showed significant differences in emission rates. Overall, the 2008 model MBE4000 engine with the DOC/DPF had substantially lower emissions across the board compared to the 2000 model year Caterpillar C-15 engine that lacked any emission controls. This demonstrates that the DOC/DPF was very effective at removing the gaseous carbonyls from the exhaust, which has been observed in prior studies as well.8,10 These results also clearly show that the engine technology is more important in terms of controlling emissions than the different fuels. The one concern about the DPF is the potential emission pulse when the filter is being regenerated, which involves adding fuel directly to the filter to heat it up and burn off the trapped particulates. The test matrix in this study deliberately avoided sampling during a regeneration event by forcing a regeneration event after each fuel change. The emissions during the DPF regeneration will need to be assessed in the future to give a more complete presentation of the emissions from vehicles with DPFs. A review of the literature on carbonyl emissions from biodiesel engines is complicated by many factors including: testing platform (engine dynamometer or chassis dynamometer), size of engine tested (diesel generator, passenger vehicle, or heavyduty vehicle), analytical methodology (DNPH cartridge, DNPH impinger, or sulfite/PFBHA derivatization), and units in which the data are presented (ng/L, μg/km and % change from ULSD). Therefore, only qualitative trends in biodiesel relative to petroleum diesel can be reported. In general, the consensus appears to be that biodiesel increases emissions of acrolein,5,6,11−13,21−25 although there are reports of lower acrolein emissions 10 or divergent trends depending on conditions.7 The results from this current study support the consensus of increasing acrolein emissions and the magnitude of the increase in vegetable-derived biodiesel is comparable to other studies. Another observation in the current study was that the emissions of aromatic aldehydes was higher in the ULSD tests

Figure 2. Emission rates (μg/km, plotted on the same scale as Figure 1) of three representative carbonyls during the “cruise” drive cycle with the 2000 model year Caterpillar C-15 engine. No samples were collected for the soy 50% runs. The MBE4000 engine was not tested with the cruise drive cycle. The error bars represent the standard deviation of emission rates from separate chassis dynamometer runs (n = 3 for ULSD and n = 2 for all other trials).

compared to the 100% biodiesel tests, although some intermediate blends (soy 50% and animal 50%) had higher emissions. The literature is not consistent on the influence of biodiesel on aromatic aldehyde emissions, with some groups reporting lower emissions,10−12,24 others reporting higher emissions,13,21,23 and some groups reporting divergent trends depending on the particular aromatic aldehyde.5−7 Biodiesel typically has effectively no aromatic content, hence it would be expected to have lower aromatic aldehyde emissions. However, the differences in aromatic aldehydes emission rates between the ULSD and the biofuels is far smaller than the differences in the aromatic composition of the fuel. For example, the renewable diesel had aromatic aldehyde emissions that were about 2 to 3-fold lower than the ULSD diesel for the UDDS drive cycle, but the aromatic content of the fuel (Supporting Information Table S1) is approximately 50-fold lower. These data suggest that the aromatic aldehydes may be the result of the combustion of other materials (e.g., lubrication oil) or by a different combustion processes other than the simple oxidation of aromatics in the fuel. 8387

dx.doi.org/10.1021/es301659u | Environ. Sci. Technol. 2012, 46, 8382−8388

Environmental Science & Technology

■

Article

(12) Guarieiro, L. L. N.; de Souza, A. F.; Torres, E. A.; de Andrade, J. B. Emission profile of 18 carbonyl compounds, CO, CO(2), and NO(x) emitted by a diesel engine fuelled with diesel and ternary blends containing diesel, ethanol and biodiesel or vegetable oils. Atmos. Environ. 2009, 43 (17), 2754−2761. (13) Cardone, M.; Prati, M. V.; Rocco, V.; Seggiani, M.; Senatore, A.; Vitolo, S. Brassica carinata as an alternative oil crop for the production of biodiesel in Italy: Engine performance and regulated and unregulated exhaust emissions. Environ. Sci. Technol. 2002, 36 (21), 4656−4662. (14) Goelen, E.; Lambrechts, M.; Geyskens, F. Sampling intercomparisons for aldehydes in simulated workplace air. Analyst 1997, 122 (5), 411−419. (15) Durbin, T. D.; Miller, J. W.; Johnson, K.; Hajbabaei, M.; Kado, N. Y.; Kobayashi, R.; Liu, X.; Vogel, C. F. A.; Matsumura, F.; Wong, P. S.; Cahill, T. M.; Okamoto, R.; Mitchell, A. CARB assessment of the emissions from the use of biodiesel as a motor vehicle fuel in California “Biodiesel characterization and NOx mitigation study; California Air Resources Board: Sacramento, CA, 2011. (16) Herner, J. D.; Hu, S. H.; Robertson, W. H.; Huai, T.; Collins, J. F.; Dwyer, H.; Ayala, A. Effect of Advanced Aftertreatment for PM and NO(x) control on heavy-duty diesel truck emissions. Environ. Sci. Technol. 2009, 43 (15), 5928−5933. (17) Aatola, H.; Larmi, M.; Sarjovaara, T.; Mikkonen, S. Hydrotreated vegetable oil (HVO) as a renewable diesel fuel:trade-off between NOx, particulate emission, and fuel consumption of a heavy duty engine; SAE International, Technical Paper No. 2008−01−2500: 2008. (18) Seaman, V. Y.; Charles, M. J.; Cahill, T. M. A sensitive method for the quantification of acrolein and other volatile carbonyls in ambient air. Anal. Chem. 2006, 78 (7), 2405−2412. (19) Spada, N.; Fujii, E.; Cahill, T. M. Diurnal cycles of acrolein and other small aldehydes in regions impacted by vehicle emissions. Environ. Sci. Technol. 2008, 42 (19), 7084−7090. (20) Gieleciak, R.; Hager, D.; Fairbridge, C. Advanced analysis of the CARB biodiesels (B100), biodiesel blend (B20), and ULSD Samples; Natural Resources Canada, CanmetENERGY-Devon: 2012. (21) Peng, C. Y.; Yang, H. H.; Lan, C. H.; Chien, S. M. Effects of the biodiesel blend fuel on aldehyde emissions from diesel engine exhaust. Atmos. Environ. 2008, 42 (5), 906−915. (22) Fontaras, G.; Karavalakis, G.; Kousoulidou, M.; Ntziachristos, L.; Bakeas, E.; Stournas, S.; Samaras, Z. Effects of low concentration biodiesel blends application on modern passenger cars. Part 2: Impact on carbonyl compound emissions. Environ. Pollut. 2010, 158 (7), 2496−2503. (23) Fontaras, G.; Karavalakis, G.; Kousoulidou, M.; Tzamkiozis, T.; Ntziachristos, L.; Bakeas, E.; Stournas, S.; Samaras, Z. Effects of biodiesel on passenger car fuel consumption, regulated and nonregulated pollutant emissions over legislated and real-world driving cycles. Fuel 2009, 88 (9), 1608−1617. (24) Correa, S. M.; Arbilla, G. Carbonyl emissions in diesel and biodiesel exhaust. Atmos. Environ. 2008, 42 (4), 769−775. (25) Fontaras, G.; Kousoulidou, M.; Karavalakis, G.; Bakeas, E.; Samaras, Z. Impact of straight vegetable oil-diesel blends application on vehicle regulated and non-regulated emissions over legislated and real world driving cycles. Biomass Bioenergy 2011, 35 (7), 3188−3198.

ASSOCIATED CONTENT

S Supporting Information *

A detailed standard operating procedure for carbonyl collection and analysis, along with details of our acrolein quality assurance program. A table of the basic fuel characteristics and the comprehensive results tables that contain additional minor chemicals. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank the California Air Resources Board for both financial and logistical support and, in particular, Tom Durbin, Norm Kado, Lex Mitchell, William Robertson, Kwangsam Na, Thomas Ladzinki, Ralph Rodas, George Gatt, and Keshav Sahay, who were vital for study design and sample generation. Additionally, we thank the National Biodiesel Board for donating the fuels for this project.

■

REFERENCES

(1) U. S. Environmental Protection Agency. A Comprehensive Analysis of Biodiesel Impacts on Exhaust Emissions, Draft Technical Report, Report # EPA420-P-02−001. Oct 2002. http://www.epa.gov/ otaq/models/analysis/biodsl/p02001.pdf. (2) Seaman, V. Y.; Bennett, D. H.; Cahill, T. M. Indoor acrolein emission and decay rates resulting from domestic cooking events. Atmos. Environ. 2009, 43 (39), 6199−6204. (3) Da Silva, T. O.; Pereira, P. A. D. Influence of time, surface-tovolume ratio, and heating process (continuous or intermittent) on the emission rates of selected carbonyl compounds during thermal oxidation of palm and soybean oils. J. Agric. Food Chem. 2008, 56 (9), 3129−3135. (4) Fullana, A.; Carbonell-Barrachina, A. A.; Sidhu, S. Comparison of volatile aldehydes present in the cooking fumes of extra virgin olive, olive, and canola oils. J. Agric. Food Chem. 2004, 52 (16), 5207−5214. (5) Karavalakis, G.; Bakeas, E.; Stournas, S. Influence of oxidized biodiesel blends on regulated and unregulated emissions from a diesel passenger car. Environ. Sci. Technol. 2010, 44 (13), 5306−5312. (6) Karavalakis, G.; Boutsika, V.; Stournas, S.; Bakeas, E. Biodiesel emissions profile in modern diesel vehicles. Part 2: Effect of biodiesel origin on carbonyl, PAH, nitro-PAH and oxy-PAH emissions. Sci. Total Environ. 2011, 409 (4), 738−747. (7) Karavalakis, G.; Stournas, S.; Bakeas, E. Light vehicle regulated and unregulated emissions from different biodiesels. Sci. Total Environ. 2009, 407 (10), 3338−3346. (8) Ballesteros, R.; Monedero, E.; Guillen-Flores, J. Determination of aldehydes and ketones with high atmospheric reactivity on diesel exhaust using a biofuel from animal fats. Atmos. Environ. 2011, 45 (16), 2690−2698. (9) He, C.; Ge, Y. S.; Tan, J. W.; You, K. W.; Han, X. K.; Wang, J. F.; You, Q. W.; Shah, A. N. Comparison of carbonyl compounds emissions from diesel engine fueled with biodiesel and diesel. Atmos. Environ. 2009, 43 (24), 3657−3661. (10) Ratcliff, M. A.; Dane, A. J.; Williams, A.; Ireland, J.; Luecke, J.; McCormick, R. L.; Voorhees, K. J. Diesel particle filter and fuel effects on heavy-duty diesel engine emissions. Environ. Sci. Technol. 2010, 44 (21), 8343−8349. (11) Yuan, C. S.; Lin, Y. C.; Tsai, C. H.; Wu, C. C.; Lin, Y. S. Reducing carbonyl emissions from a heavy-duty diesel engine at US transient cycle test by use of paraffinic/biodiesel blends. Atmos. Environ. 2009, 43 (39), 6175−6181. 8388

dx.doi.org/10.1021/es301659u | Environ. Sci. Technol. 2012, 46, 8382−8388