Influence of Oxidized BiodieselBlends on Regulated andUnregulated Emissions from aDiesel Passenger CarG E O R G I O S K A R A V A L A K I S , * , †

E V A N G E L O S B A K E A S , ‡ A N DS T A M O S S T O U R N A S †

Laboratory of Fuels Technology and Lubricants, School ofChemical Engineering, National Technical University ofAthens, 9 Iroon Polytechniou Str. Zografou Campus, 157 80,Athens, Greece, and Laboratory of Analytical Chemistry,Chemistry Department, National and Kapodistrian Universityof Athens, Panepistimioupolis, 15771, Athens, Greece

Received March 15, 2010. Revised manuscript receivedMay 6, 2010. Accepted May 13, 2010.

This paper investigates the effects of biodiesel blends onregulated and unregulated emissions from a Euro 4 dieselpassenger car, fitted with a diesel oxidation catalyst and a dieselparticlefilter(DPF).Emissionandfuelconsumptionmeasurementswere conducted for the New European Driving Cycle (NEDC)and the Artemis driving cycles. Criteria pollutants, along withcarbonyl, polycyclic aromatic hydrocarbon (PAH) and nitrate PAHand oxygenate PAH emissions, were measured and recorded.A soy-based biodiesel and an oxidized biodiesel, obtainedfrom used frying oils, were blended with an ultra low sulfurdiesel at proportions of 20, 30, and 50% by volume. The resultsshowed that the DPF had the ability to significantly reduceparticulate matter (PM) emissions over all driving conditions.Carbon monoxide (CO) and hydrocarbon (HC) emissions werealso reduced with biodiesel; however, a notable increase innitrogen oxide (NOx) emissions was observed with biodieselblends. Carbon dioxide (CO2) emissions and fuel consumptionfollowed similar patterns and increased with biodiesel. Theinfluence of fuel type and properties was particularly noticeableon the unregulated pollutants. The use of the oxidizedbiodiesel blends led to significant increases in carbonylemissions, especially in compounds which are associatedwith potential health risks such as formaldehyde, acetaldehyde,and acrolein. Sharp increases in most PAH compounds andespecially those which are known for their toxic and carcinogenicpotency were observed with the oxidized blends. The presenceof polymerization products and cyclic acids were the mainfactors that influenced the PAH emissions profile.

IntroductionBiodiesel is chemically synthesized via the transesterificationfrom vegetable oils, used frying oils, and animal fats. Biodieselproperties are similar to those of diesel fuel; it has low sulfurcontent, is free of aromatic compounds, is nontoxic andreadily biodegradable, and possesses a higher cetane number,

higher flash point, and better lubricity performance. Despiteits many advantages, biodiesel has poor oxidation stabilitywith respect to petroleum diesel. Fuel stability may be affectedby the type of feedstock, the presence of naturally occurringantioxidants, and the storage conditions (1). Oxidationstability is of great importance in the context of possibleproblems with engine parts, as well as the impact ofemissions. The main oxidation products are peroxides andhydroperoxides. During further degradation, these productsform shorter-chain compounds such as low molecular weightacids, aldehydes, ketones, and alcohols. Further reactions ofthe unstable hydroperoxide species with another fatty acidchain may form high molecular weight materials, such asdimer or trimer acids which may lead to filter blocking,injector failures, and deposit formation (2, 3).

The use of biodiesel for reducing the environmentalimpact of diesel emissions has been widely investigated. Infact, many studies have centered their research on the criteriapollutants, such as carbon monoxide (CO), unburnedhydrocarbons (HC), nitrogen oxides (NOx), and particulatematter (PM). Many authors have shown that the addition ofbiodiesel can reduce most of the aforementioned pollutants,with the exemption of NOx (4–6). It has been demonstratedthat the positive or adverse effect of biodiesel fuels on PMemissions varied significantly among vehicles, engine tech-nology, and test cycle (7). A number of authors showed thatPM emissions for the New European Driving Cycle (NEDC)may be adversely affected by biodiesel, a phenomenon whichis mainly attributed to certain physicochemical parametersof biodiesel and to the cold-start conditions (8, 9). In theliterature, there are two diverse interpretations concerningNOx emissions of biodiesel. Some studies reported higherNOx production (10, 11), whereas others have shown lowerNOx emissions in respect to diesel fuel (12, 13). The increasein NOx emissions is still not very well explained, but severalparameters, including fuel type, fuel quality, fuel spraycharacteristics, operating conditions, and engine technologyare implicated (14–16).

Limited information is available regarding the effects ofbiodiesel on emissions of carbonyl compounds and polycyclicaromatic hydrocarbons (PAHs). Concerning carbonyl emis-sions, there are some divergences when considering theresults obtained with diesel fuel and biodiesel blends. Manystudies showed a clear increasing trend of these emissionswhen biodiesel was used (17, 18); however, some studiesreveal decreases or even insignificant differences (19, 20). Itshould be noted that carbonyl compounds from vehicularexhaust are of great importance since some species are toxic,mutagenic, and even carcinogenic to the human body. Theyalso play a critical role to the tropospheric chemistry, as theyare important precursors to free radicals (HOx), ozone, andperoxyacylnitrates (21). PAH emissions are released duringincomplete combustion of fossil fuels, are widely distributedin the atmosphere, and are one of the first pollutants to havebeen identified as suspected carcinogens (22). Nitrated andoxygenated PAHs, which have elicited the most concern dueto their mutagenic and carcinogenic properties, are alsoproducts of fuel combustion (23, 24). The information givenin literature about the effect of biodiesel on PAH emissionsis limited and often contradictory. The majority of authorshave observed some decrease in PAH emissions withbiodiesel, although a noticeable dependence on engineoperation conditions is usually acknowledged (25–27). Onthe other hand, some authors found some increase withbiodiesel, which may be attributed to the test cycle and thefuel chemical structure (19, 28).

* Corresponding author phone: +30 210 7723213; fax: +30 2107723163; e-mail: karavalg@mail.ntua.gr.

† National Technical University of Athens.‡ National and Kapodistrian University of Athens.

Environ. Sci. Technol. 2010, 44, 5306–5312

5306 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 13, 2010 10.1021/es100831j 2010 American Chemical SocietyPublished on Web 05/24/2010

All of the above show that biodiesel impact on engineemissions could be quite significant. Few studies are availableon modern passenger cars, employed common-rail enginesystems, and after-treatment technologies, and even fewerstudies report results, which are not necessarily representativeof actual driving conditions, making it difficult to assess thefuel impact on diesel car fleet emissions. This work has threemain objectives: (i) to evaluate the impact of fuel sourcematerial on the formation of exhaust emissions from amodern passenger car representative of the current Europeanfleet, (ii) to investigate the effect of the use of oxidized fuelon the formation of carbonyl and PAH emissions, and (ii) tostudy the influence of biodiesel concentration and drivingcycle on the exhaust emissions.

Experimental SectionTest Fuels and Vehicle. A total of seven fuels were evaluatedin this study. An ultra low sulfur diesel meeting the currentEuropean Union (EU) fuel quality requirements for dieselvehicles (EN 590:2008) was used as reference fuel and tocreate blends with two types of methyl esters. Soy-basedmethyl ester (SME) and used frying oil methyl ester (UFOME)were blended with the reference diesel at proportions of 20,30, and 50% v/v. The neat biodiesels were examined accordingto the automotive fatty acid methyl ester (FAME) standardEN 14214 (Table S1-S2 in the Supporting Information). Itshould be mentioned, that UFOME was an oxidized fuel,since it was naturally aged during long-term storage. Themain quality properties of the diesel fuel and its blends withboth biodiesels are given in the Supporting Informationsection (Table S2-S3).

A 2007 model year Subaru Forester 2.0D XS (SUV type),equipped with a common-rail direct injection diesel engineand meeting Euro 4 emission standards, was used in thisstudy. Emissions in this vehicle were controlled by a dieseloxidation catalyst (DOC) and a silicon carbide (SiC) dieselparticulate filter (DPF). All emission tests were performedwith the vehicle in its original configuration. The technicalspecifications of the vehicle are listed in Table S3-S4 in theSupporting Information.

Driving Cycles and Measurement Protocol. In order toinvestigate the impact of biodiesel on the exhaust emissionsand fuel consumption, the vehicle was driven on a chassisdynamometer over the certification NEDC and the nonleg-islated Artemis driving cycles. The Artemis cycles are dis-tinguished into an urban (Urban), a rural (Road), and amotorway (Motorway) part, each representative of thecorresponding driving condition. The speed vs time profilesof the applied driving cycles can be found elsewhere (9).

The daily measurement protocol started with the NEDC,which is a cold-start driving cycle. This comprises two parts:an urban part (UDC), where the engine starts from roomtemperature, and an extra-urban part (EUDC), which aimsat testing the car at higher than urban speeds. The NEDCwas then followed by the three Artemis cycles. This protocolwas repeated twice per fuel blend, while two sets with thereference fuel were conducted at the beginning and end ofthe campaign. Prior to each measurement, the vehicle wasconditioned for about 350-400 km before testing whenevera fuel change was required.

Exhaust Sampling and Emission Analyzers. Emissionmeasurements were conducted following the Europeanregulations (Directive 70/220/EEC and amendments). Gas-eous and PM mass were sampled according to the constantvolume sampling (CVS) technique, with a dilution tunnel. Aschematic of the sampling system and detailed informationregarding the emission analyzers are given in ref 27.

Carbonyls and PAHs Analysis. A detailed descriptionregarding the sampling and analysis of carbonyls and PAHsis given in the Supporting Information section.

Results and DiscussionFuel Consumption and CO2 Emissions. Figure 1a-b showsthe experimental results obtained for the CO2 emissions (a)and fuel consumption (b). Regarding CO2, some increaseswere observed with biodiesel over all driving conditions.Higher increases were found over the NEDC, which may beattributed to the cold-start UDC phase of the cycle. Thisphenomenon indicates a possible drop in engine efficiencywhen biodiesel under low-speed load conditions is used.The biodiesel impact on CO2 emissions ranged on averagefrom 1 to 6% over the NEDC and the Artemis cycles.

Fuel consumption, expressed in L/100 km, presentedsimilar patterns with CO2 emissions. An increase in fuelconsumption proportional to the difference in energy contentof the fuels was observed for tested fuels. This increase wasin the order of 1 to 6% on average over all cycles for thebiodiesel blends. It should be stressed that the oxidized blendsproduced slightly higher consumption when compared tothe SME blends. This observation was probably based uponthe fact that the oxidized biodiesel contained a higher amountof oxygen (due to the formation of oxidation products) and,hence, presented lower energy content than the SME.

FIGURE 1. (a-b) CO2 emissions and fuel consumption for allfuel/cycle combinations.

VOL. 44, NO. 13, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 5307

Regulated Emissions. In order to facilitate result pre-sentation, the regulated emissions of NOx, PM, HC, and COare presented in Figure 2a-d. Regarding NOx emissions, anincreasing trend was observed when increasing biodieselconcentration over all driving cycles, with the exception ofSME-50 blends which resulted in some reductions over theNEDC and Artemis Urban for no obvious reason, since theresults were repetitive. When biodiesel is used, NOx emissionlevels over the NEDC were found to be above the Euro 4specification limits (0.25 g/km). The increase was in the orderof 9 and 8% on average for SME and UFOME blends,respectively. Biodiesel application during the Artemis drivingcycles did not change NOx profile; however, the emissionlevels were higher than those of the NEDC. The highest NOx

emissions were observed for the Artemis Urban cycle. Thosewere 3.1-3.7 higher than equivalent emissions observed forNEDC. The aforementioned observations support the hy-pothesis that the engine control is optimized to use regularautomotive diesel, and its calibration was based upon theoperating points of the NEDC. The lower NOx emission levelsfor the NEDC, as opposed to Artemis cycles, were as expected.NEDC is characterized by low vehicle speed and low engineload and, thus, low exhaust gas temperature. As a higherload induces a higher combustion temperature and NOx

emissions increase with temperature, NOx emissions arehigher at higher loads. Additionally, the presence of oxygenalso enchases combustion temperature resulting in higherNOx formation and, thus, higher exhaust temperature.

The impact of biodiesel source material and type on NOx

emissions may be potentially explained by several mecha-nisms. It has been reported that the increased number ofdouble bonds and, thus, the lower cetane number, alongwith the oxygen content in the ester molecule, may affectNOx formation (5, 14, 29). Under the present test conditions,both biodiesels were mainly composed of unsaturated fatty

esters and, therefore, an increase in NOx may result from therole of the double bonds in the combustion chemistry.

PM emission results are shown in Figure 2b. Largereductions in PM were achieved with biodiesel blends overall driving cycles. These reductions can be explained by theincrease in oxygen content in the fuel which contributes tomore complete fuel oxidation even in locally rich zones andthe lack of sulfur and aromatics since these compoundsincrease the soot nucleation rate (30). Another majorcontributing factor was the exhaust configuration of thevehicle (DOC+DPF), which led in full oxidation of PM insidethe catalyst. In general, PM emissions were found well belowthe Euro 4 limit (0.025 g/km) over all driving cycles butsurprisingly above the future Euro 5 limit (0.005 g/km). Thehighest reductions were observed over the NEDC and rangedfrom -7 to -23% and -9 to -24% for the SME and UFOMEblends, respectively. Smaller reductions were achieved overthe Artemis cycles, which were in the order of -15, -4, and-3% on average for all biodiesel blends over Urban, Road,and Motorway, respectively. However, the emission levelsobtained during Road and Motorway operation were lowerthan those of the NEDC. These results suggest that thecombined effect of the cold-start UDC phase and the lowervolatility of biodiesel adversely influenced the formation ofPM emissions during the NEDC. This could be a consequenceof the poorer combustion, of lower fuel vaporization at lowtemperature, which was the case with the biodiesel blends,and also of the reduced efficiency of the catalytic converter,which was below the light-off temperature in the startingphase of the cycle (8, 9, 29, 31).

A trend toward higher HC emission levels was observedwith biodiesel over the legislated NEDC (Figure 2c). Thisphenomenon may be due to the lower volatility of thebiodiesel blends and the cold-start effect during the UDCoperation. A different behavior was observed for both

FIGURE 2. (a-d) NOx (a), PM (b), HC (c), and CO (d) emission measurement results for the tested fuels over the NEDC and theArtemis driving cycles.

5308 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 13, 2010

biodiesel blends over the Artemis cycles. In general, emissionlevels seem to be reduced as the mean driving cycle powerincreases. Similar to HC, CO emissions decreased as thecontent of biodiesel increased over all driving conditions(Figure 2d). Higher CO emission levels were observed overthe NEDC which can be attributed to the poor catalystefficiency; while during Artemis operation, these emissionswere noticeably lower. Again, it should be noted that theincrease of the mean load and speed of the cycle, along withthe oxygen availability of biodiesel, positively affected COemissions (32). In addition, during Artemis operation, thecatalytic activity is higher due to the increased exhausttemperature and, thus, leads to lower HC and CO emissions.

Carbonyl Emissions. Thirteen carbonyl compounds (al-dehydes and ketones) were identified in the exhaust gases,and the concentrations measured over all driving cycles arelisted in Tables S4-S7 and S5-S8 in the Supporting Informa-tion. Consistent with other studies (33–35), low molecularweight compounds such as formaldehyde, acetaldehyde,acrolein, and propionaldehyde were found to be the mostabundant carbonyls emitted. However, heavier compoundssuch as crotonaldehyde, methacrolein, and butyraldehydewere present in the exhaust in relatively high concentrations.The addition of biodiesel, independent of its origin, providedsignificant increases in emissions of carbonyl compoundsover all cycles. A clear trend of increase in carbonyl emissionswas observed with the increase in biodiesel content. On theother hand, aromatic aldehydes produced discordant results.Benzaldehyde was reduced with biodiesel as expected, sinceits formation depends mainly from the aromatic content ofthe fuel (34). However, p-tolualdehyde and hexanaldehydeincreased with biodiesel, an observation which indicates thatthe formation of these compounds may be influenced byother parameters than the aromatic content.

Under the present test conditions, the level of formal-dehyde emissions significantly increased over reference levelsfor both types of biodiesel blends. The highest increases overthe NEDC were 109 and 54% for UFOME-50 and SME-50,respectively. During Artemis operation, the blends of UFOMEresulted in average increases of 38, 42, and 38% over Urban,Road, and Motorway, respectively. The same trend was foundfor the SME blends where the average increases were 15, 23,and 27% over Artemis Urban, Road, and Motorway, respec-tively. Acetaldehyde emissions followed a similar pattern withformaldehyde and increased with biodiesel fraction in thefuel. The average increases for the UFOME blends were inthe order of 26, 22, 43, and 50%, while for the SME blendsthe increases were 12, 12, 30, and 35% over NEDC, Urban,Road, and Motorway, respectively. The higher level of thesecompounds with biodiesel may be attributed to the presenceof short-chain esters that favor formation of the shortestchain aldehydes (namely formaldehyde and acetaldehyde)during combustion, since formaldehyde in vehicle exhaustis mainly produced from the incomplete combustion ofsaturated aliphatic hydrocarbons (36). This was probablythe case of used frying oil methyl ester, which was initiallycomposed of unsaturated methyl esters, where during therepeated deep frying process the formation of short-chaincomponents was favored from the breakdown of unsaturatedfatty acids (20). Another major contributing factor for thehigher carbonyl emission levels was the presence of oxygenin the ester molecule (4, 18).

A strong increase was observed in acrolein emissions forall fuel/cycle combinations. A clear trend toward higheremissions with the application of UFOME blends wasobserved as regards the SME blends. The average increasesfor the UFOME blends were in the order of 19, 17, 35, and40%, while for the SME blends these increases were 17, 13,29, and 32% over the NEDC, Urban, Road, and Motorway,respectively. Acrolein emissions are mainly originated by the

oxidation of glycerol, glycerides, and fatty acid residuespresent in the biodiesel (36). This result is in agreement withthe higher amount of glycerol and glycerides content foundin UFOME as compared to SME, which met the strictspecifications.

Figure 3 shows the total carbonyl emissions for all fuel/cycle combinations. Results indicate that the presence ofthe UFOME blends led to significantly higher emissions thanthe corresponding SME blends and diesel fuel. This phe-nomenon may be explained by the fact that the parent oilalready had an amount of carbonyls and carboxylic acids,which were formed during the thermal stressing of the oiland finally remained in the methyl ester during transesteri-fication. In line with the above, a considerable amount ofaldehydes and ketones, especially those of high molecularweight, would be present in the fuel due to its oxidized nature(37, 38). As the oxidation and auto-oxidation mechanismsproceed, the hydroperoxides which are formed during theprimary oxidation stage decompose to ultimately formaldehydes (secondary oxidation products). A more in depthanalysis of the individual carbonyl compounds shows thatthe use of UFOME blends led to higher emission levels ofheavier compounds than the blends of SME. Therefore, it isreasonable to assume that the application of UFOME wouldbe more prone to higher formation of carbonyl emissionswhen compared to SME.

Concerning the influence of the driving cycle on carbonylcompounds, notable differentiations were observed betweenthe employed cycles. During operation over the NEDC, higheremissions levels were obtained when compared to theArtemis cycles. This phenomenon can be ascribed to thecold-start effect of the UDC phase and the partial deactivationof the oxidation catalyst. The exhaust concentration of allcarbonyls were lower over the transient Artemis cycles, whichwas due to the increased exhaust temperatures and, thus,the higher performance of the oxidation catalyst. On the otherhand, the increased combustion efficiency, due to the higheraverage speed and load of such driving modes, possibly ledto reduced carbonyl emissions as these compounds aremainly products of incomplete combustion (39).

PAH, Nitro-PAH, and Oxy-PAH Emissions. Although theDPF are very effective in capturing PM and leaving engineexhaust gases nearly clean of PM, there is no clear evidenceof their effect on PAH emissions. A total number of 12 PAHs,4 nitro-PAHs, and 6 oxy-PAHs were identified and quantifiedin the vehicle’s exhaust. Detailed information on these

FIGURE 3. Total carbonyl emissions for all fuel/cyclecombinations.

VOL. 44, NO. 13, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 5309

emissions is shown in Tables S6-S9 and S7-S11 in theSupporting Information. It is evident from the results thatthe application of UFOME blends led to higher PAH emissionscompared to those of diesel fuel. On the contrary, the blendsof SME produced lower PAH emissions for all fuel/cyclecombinations. It should be noted that the use of DPFpositively influences the PAH profile in the exhaust, since itis assumed that these species were either absorbed in theDPF or probably converted to CO2. Another critical observa-tion is that the concentrations of both lower and highermolecular weight PAHs were found in similar levels. Thesefindings are not consistent with the results reported inprevious studies, which showed that PAH levels are higherin the absence of DPF while light PAHs are usually thepredominant compounds in diesel exhaust (26, 40).

Heavier PAH compounds such as those of indeno[1,2,3-c,d]pyrene and benzo[g,h,i]perylene were only detected forUFOME blends during all driving conditions. For the SMEblends, these species remained almost undetectable. How-ever, the concentrations of these microcontaminants werelower than those of light PAHs. It is possible that PAHs withhigher ring number are retained more efficiently in DPFs(41). The increased emissions of these species might be dueto pyrosynthesis of lower molecular weight aromatic com-pounds to larger PAHs and to the contribution of the lubricantoil (42). Low molecular-weight PAHs (containing 3-5 aro-matic rings), such as those of phenanthrene, anthracene,fluoranthene, and pyrene, were found in higher concentra-tions with the use of UFOME blends in respect to diesel fueland SME blends. The higher levels of light PAHs suggest thatthese compounds were pyrolyzed from incomplete combus-tion of the fuel (22, 43, 44). With increasing market share ofDPF vehicles, it is expected that PAH profiles in urban airwill change. However, under the present test conditions, PAHcompounds which are known for their toxic, mutagenic, andcarcinogenic properties, such as chrysene, benzo[a]an-thracene, and benzo[a]pyrene were significantly increasedwith the application of UFOME blends. This phenomenonmay be attributed to fuel composition and the de novoformation of these species in the DPF.

Of the nitro-PAH compounds analyzed, only 1-nitro-pyrene and 6-nitro-benzo[a]pyrene were found in quantifi-able levels in the particle phase. 6-Nitro-benzo[a]pyreneemissions were reduced for all fuel/cycle combinations whencompared to diesel fuel. On the other hand, emission levelsof 1-nitro-pyrene showed discordant results. The use ofUFOME blends led to strong increases as compared to dieselfuel, while the use of SME blends provided some reductionswith the exemption of a marginal increase over Artemis Road.It is possible that DPFs can support nitration chemicalreactions, maybe due to the high NOx emissions constantlypassing through the filter. In any case, further research isrequired in order to strengthen this hypothesis.

Oxygenated PAH levels were found in significant loweramounts than their parent PAHs. It should be mentionedthat four of the six oxy-PAH compounds in the test matrixwere found in quantifiable levels. The prominent oxy-PAHsemitted were anthraquinone, benzanthrone, benz[a]an-thracene-7,12-dione, and 9-fluorenone. The relatively highlevels of anthraquinone, benzanthrone, and benz[a]an-thracene-7,12-dione can be attributed to the fact that thesecompounds are the most stable fragments of oxidized PAHs(45). The general picture showed that the addition of biodieselhad a negative effect on the emissions of these compounds.This was particularly noticeable with the use of UFOMEblends, which led in sharp increases over all driving condi-tions. Despite the fact that the SME blends presented loweroxy-PAH emissions compared to diesel fuel, their emissionlevels increased with increasing biodiesel content. In general,the higher levels of oxy-PAH emissions may be related to the

oxygenated group in the methyl ester. These observationsgenerate scepticism on the true impact of biodiesel on oxy-

FIGURE 4. (a-c). Total PAH, nitro-PAH, and oxy-PAH emissionsfor all fuel/cycle combinations.

5310 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 13, 2010

PAH emissions because of the specific toxicity of quinoidand other oxygenated compounds (46).

Figure 4a-c shows the total PAH, nitro-PAH, and oxy-PAH emissions, which are the sum of the concentrations ofthe 12 PAH, 4 nitro-PAH, and 6 oxy-PAH components,respectively. Total PAH emissions confirm the negative andbeneficial performance of UFOME and SME blends, respec-tively. It is probable that fuel composition and quality makea major contribution to the PAHs found in the exhaust of thevehicle. As mentioned previously, the combined effect ofbiodiesel source material and its oxidized nature led to suchincreases in PAH emissions. It is reasonable to assume thatthe biodiesel obtained from used frying oils would containdimers, trimers, polymerization products, and cyclic acidsoriginated from breakdown between two fatty acid chains inthe same triacylglyceride molecule (via, e.g., Diels-Alderreaction). These diesters can be formed during thermalstressing, and they would be present in the final fuel and,thus, in the exhaust (45, 47). The case of SME presents somedifferences when compared to UFOME. The PAH emissionreduction may be attributed to the presence of excess oxygenin biodiesel and the absence of aromatic and polyaromaticcompounds in the fuel (25).

A clear correlation was observed between PAH emissionsand the applied driving conditions. The exhaust concentrationsof all PAHs were quite low over Artemis Road and Motorwaycycles when compared to NEDC and Artemis Urban, i.e.,simulating driving at low average speeds in city conditions.The observed reductions in PAH emissions can be attributedto the higher average speed and engine load during these drivingconditions, which increases exhaust temperatures and thusresults in better oxidation of these compounds both inside thevehicle catalyst and the sampling system (39, 48). The higherPAH emissions obtained over the NEDC may be due to thecold-start effect and the reduced catalyst efficiency, along withthe lower fuel vaporization at low-temperature, which was thecase with the biodiesel blends (45).

Toxicity. The health risk associated with inhalatoryexposure to PAHs is commonly assessed on the basis ofbenzo[a]pyrene, which is known for its high carcinogenicand toxic properties (47). The carcinogenic effect of eachindividual PAH can be determined by means of a conversionfactor (toxic equivalent factor, TEF). TEFs for individual PAHswere used to estimate human health risk associated withinhalatory exposure to PAHs (44). Figure 5 shows the general

toxicity of the tested fuels over the several driving cycles. Asit can be observed, there is a clear reduction in the overalltoxicity of SME blends compared to diesel fuel. On the otherhand, TEF was significantly increased for the UFOME blendsover the legislated NEDC and Artemis Urban, which meanthat these emissions have higher health risks than conven-tional fuel emissions. This phenomenon was attributed tothe increased levels of the highly carcinogenic compoundsof benzo[a]anthracene and benzo[a]pyrene during thesedriving modes. This result suggests that the application ofan oxidized biodiesel obtained from used frying oils elimi-nates the benefits of the biofuel with respect to the soy-based biodiesel. It was also evident from the results that theoverall PAH toxicity was influenced by the driving cycle. Thelower toxicity for all fuel blends during Road and Motorwaydriving implies that the emitted PAH compounds were oflow molecular weight, which have lower toxicities comparedto heavier species. On the contrary, the formation of largePAHs with higher toxicities was favored by low-load condi-tions and the cold-start effect.

AcknowledgmentsThis paper is dedicated to the memory of Professor StamosStournas.

Supporting Information AvailableIncluding the main fuel quality properties, description ofsampling, and analysis of carbonyls and PAHs, as well as theindividual carbonyl and PAH compounds emitted for all fuel/cycle combinations.This material is available free of chargevia the Internet at http://pubs.acs.org.

Literature Cited(1) Knothe, G. Some aspects of biodiesel oxidative stability. Fuel

Process. Technol. 2007, 88, 669–677.(2) Schober, S.; Mittelbach, M. The impact of antioxidants on

biodiesel oxidation stability. Eur. J. Lipid Sci. Technol. 2004,106, 382–389.

(3) Bouaid, A.; Martinez, M.; Aracil, J. Long storage stability ofbiodiesel from vegetable and used frying oils. Fuel 2007, 86,2596–2602.

(4) Cardone, M.; Prati, M. V.; Seggiani, M.; Senatore, A.; Vitolo, S.Brassica carinata as an alternative oil crop for the productionof biodiesel in Italy: Engine performance and regulated andunregulated exhaust emissions. Environ. Sci. Technol. 2002, 36,4656–4662.

(5) Knothe, G.; Sharp, C. A.; Ryan, T. W. Exhaust emissions ofbiodiesel, petrodiesel, neat methyl esters, and alkanes in a newtechnology engine. Energy Fuels 2006, 20, 403–408.

(6) Durbin, T. D.; Norbeck, J. M. Effects of biodiesel blends andArco EC-diesel on emissions from light heavy-duty dieselvehicles. Environ. Sci. Technol. 2002, 36, 1686–1691.

(7) Fontaras, G.; Kousoulidou, M.; Karavalakis, G.; Tzamkiozis, T.;Pistikopoulos, P.; Ntziachristos, L.; Bakeas, E.; Stournas, S.;Samaras, Z. Effects of low concentration biodiesel blendapplication on modern passenger cars. Part 1: Feedstock impacton regulated pollutants, fuel consumption and particle emis-sions. Environ. Pollut. 2010, 158, 1451–1460.

(8) Martini, G.; Astorga, C.; Farfaletti, A. Effect of biodiesel fuels onpollutant emissions from EURO 3 LD diesel vehicles. Transportand Air Quality Unit; Institute for Environment and Sustain-ability, EC-Joint Research Centre; JRC Scientific and TechnicalReports (EUR collection): http://publications.jrc.ec.europa.eu/repository/handle/111111111/155, 2007.

(9) Fontaras, G.; Karavalakis, G.; Kousoulidou, M.; Tzamkiozis, T.;Ntziachristos, L.; Bakeas, E.; Stournas, S.; Samaras, Z. Effects ofbiodiesel on passenger car fuel consumption, regulated andnon-regulated pollutant emissions over legislated and real-worlddriving cycles. Fuel 2009, 88, 1608–1617.

(10) Szybist, J. P.; Song, J.; Alam, M.; Boehman, A. L. Biodieselcombustion, emissions and emission control. Fuel Process.Technol. 2007, 88, 679–91.

(11) Canakci, M. Combustion characteristics of a turbocharged DIcompression ignition engine fueled with petroleum diesel fuelsand biodiesel. Bioresour. Technol. 2007, 98, 1167–1175.

FIGURE 5. Toxicity equivalent factors (TEFs) for diesel fuel andits blends with biodiesel.

VOL. 44, NO. 13, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 5311

(12) Dorado, M. P.; Ballesteros, E.; Arnal, J. M.; Gomez, J.; Lopez, F. J.Exhaust emissions from a Diesel engine fueled with transes-terified waste olive oil. Fuel 2003, 82, 1311–1315.

(13) Szybist, J. P.; Boehman, A. L.; Taylor, J. D.; McCormick, R. L.Evaluation of formulation strategies to eliminate the biodieselNOx effect. Fuel Process. Technol. 2005, 86, 1109–1126.

(14) McCormick, R. L.; Graboski, M. S.; Alleman, T. L.; Herring, A. M.;Tyson, K. S. Impact of biodiesel source material and chemicalstructure on emissions of criteria pollutants from a heavy-dutyengine. Environ. Sci. Technol. 2001, 35, 1742–1747.

(15) Ban-Weiss, G. A.; Chen, J. Y.; Buchholz, B. A.; Dibble, R. W. Anumerical investigation into the anomalous slight NOx increasewhen burning biodiesel; A new (old) theory. Fuel Process.Technol. 2007, 88, 659–667.

(16) Fernando, S.; Hall, C.; Jha, S. NOx reduction from biodieselfuels. Energy Fuels 2006, 20, 376–382.

(17) Pang, X.; Shi, X.; Mu, Y.; He, H.; Shuai, S.; Chen, H.; Li, R.Characteristics of carbonyl compounds emission from a diesel-engine using biodiesel-ethanol-diesel as fuel. Atmos. Environ.2006, 40, 7057–7065.

(18) Correa, S. M.; Arbilla, G. Carbonyl emissions in diesel andbiodiesel exhaust. Atmos. Environ. 2008, 42, 769–775.

(19) Turrio-Baldassarri, L.; Battistelli, C. L.; Conti, L.; Crebelli, R.;Berardis, B.; Iamicelli, A. L.; Gambino, M.; Iannaccone, S.Emission comparison of urban bus engine fueled with dieseloil and biodiesel blend. Sci. Total Environ. 2004, 327, 147–162.

(20) Peng, C. Y.; Yang, H. H.; Lan, C. H.; Chien, S. M. Effects of thebiodiesel blend fuel on aldehyde emissions from diesel engineexhaust. Atmos. Environ. 2008, 42, 906–915.

(21) Bakeas, E. B.; Argyris, D. I.; Siskos, P. A. Carbonyl compoundsin the urban environment of Athens, Greece. Chemosphere 2003,52, 805–813.

(22) Ravindra, K.; Sokhi, R.; Van Grieken, R. Atmospheric polycyclicaromatic hydrocarbons: Source attribution, emission factorsand regulation. Atmos. Environ. 2008, 42, 2895–2921.

(23) Miet, K.; Le Menach, K.; Flaud, P. M.; Budzinski, H.; Villenave,E. Heterogeneous reactivity of pyrene and 1-nitropyrene withNO2: Kinetics, product yields and mechanism. Atmos. Environ.2009, 43, 837–843.

(24) Tsapakis, M.; Stephanou, E. G. Diurnal cycle of PAHs, nitro-PAHsand oxy-PAHs in a high oxidation capacity marine backgroundatmosphere. Environ. Sci. Technol. 2007, 41, 8011–8017.

(25) Bagley, S. T.; Gratz, L. D.; Johnson, J. H.; McDonald, J. F. Effectsof an oxidation catalytic converter and a biodiesel fuel on thechemical, mutagenic, and particle size characteristics of emissionsfrom a diesel engine. Environ. Sci. Technol. 1998, 32, 1183–1191.

(26) Correa, S. M.; Arbilla, G. Aromatic hydrocarbons emissions in dieseland biodiesel exhaust. Atmos. Environ. 2006, 40, 6821–6826.

(27) Karavalakis, G.; Stournas, S.; Bakeas, E. Light vehicle regulatedand unregulated emissions from different biodiesels. Sci. TotalEnviron. 2009, 407, 3338–3346.

(28) Zhou, L.; Atkinson, S. Characterising vehicle emissions fromthe burning of biodiesel made from vegetable oil. Environ.Technol. 2003, 24, 1253–1260.

(29) Lapuerta, M.; Armas, O.; Fernandez, J. R. Effect of biodieselfuels on diesel engine emissions. Prog. Energy Combust. Sci.2008, 34, 198–223.

(30) Lapuerta, M.; Rodrıguez-Fernandez, J.; Agudelo, J. R. Dieselparticulate emissions from used cooking oil biodiesel. Bioresour.Technol. 2008, 99, 731–740.

(31) Tinaut, F. V.; Melgar, A.; Briceno, Y.; Horrillo, A. Performanceof vegetable derived fuels in diesel engine vehicles. Int. Congr.Combust. Sci. 2005 [PTNSS Kongres], Poland.

(32) Lujan, J. M.; Bermudez, V.; Tormos, B.; Pla, B. Comparativeanalysis of a DI diesel engine fuelled with biodiesel blends duringthe European MVEG-A cycle: Performance and emissions (II).Biomass Bioenergy 2009, 33, 948–956.

(33) Karavalakis, G.; Stournas, S.; Bakeas, E. Effects of diesel/biodieselblends on regulated and unregulated pollutants from a pas-senger vehicle operated over the European and the Athensdriving cycles. Atmos. Environ. 2009, 43, 1745–1752.

(34) He, C.; Ge, Y.; Tan, J.; You, K.; Han, X.; Wang, J.; You, Q.; Shah,A. N. Comparison of carbonyl compounds emissions from dieselengine fueled with biodiesel and diesel. Atmos. Environ. 2009,43, 3657–3661.

(35) Grosjean, D.; Grosjean, E.; Gertler, A. W. On-road emissions ofcarbonyls from light-duty and heavy-duty vehicles. Environ.Sci. Technol. 2001, 35, 45–53.

(36) Guarieiro, L. L. N.; Pereira, P. A.d. P.; Torres, E. A.; da Rocha,G. O.; de Andrade, J. B. Carbonyl compounds emitted by a dieselengine fuelled with diesel and biodiesel-diesel blends: samplingoptimization and emissions profile. Atmos. Environ. 2008, 42,8211–8218.

(37) Waynick, A. J. Characterization of biodiesel oxidation andoxidation products. The Coordinating Research Council, Task1 Results, CRC Project No. AVFL-2b, 2005 (http://www.nrel.gov/docs/fy06osti/39096.pdf).

(38) Berrios, M.; Gutierrez, M. C.; Martin, M. A.; Martin, A. Obtainingbiodiesel from spanish used frying oil: Issues in meeting the EN14214 biodiesel standard. Biomass Bioenergy 2010, 34, 312–318.

(39) Bikas, G.; Zervas, E. Nonregulated pollutants emitted from Euro3 diesel vehicles as a function of their mileage. Energy Fuels2007, 21, 2731–2736.

(40) Lin, Y. C.; Lee, W. J.; Wu, T. S.; Wang, C. T. Comparison of PAHand regulated harmful matter emissions from biodiesel blendsand paraffinic fuel blends on engine accumulated mileage test.Fuel 2006, 85, 2516–2523.

(41) Heeb, N. V.; Schmid, P.; Kohler, M.; Gujer, E.; Zennegg, M.;Wenger, D.; Wichser, A.; Ulrich, A.; Gfeller, U.; Honegger, P.;Zeyer, K.; Emmenegger, L.; Petermann, J.-L.; Czerwinski, J.;Mosimann, T.; Kasper, M.; Mayer, A. Secondary effects ofcatalytic diesel particulate filters: conversion of PAHs versusformation of nitro-PAHs. Environ. Sci. Technol. 2008, 42, 3773–3779.

(42) Lea-Langton, A. ; Li, H.; Andrews, G. E. Comparison of particulatePAH emissions for diesel, biodiesel and cooking oil using a heavyduty DI diesel engine. SAE Tech. Pap. Ser. 2008, No. 2008-01-1811.

(43) Rhead, M. M.; Hardy, S. A. The sources of polycyclic aromaticcompounds in diesel engine emissions. Fuel 2003, 82, 385–393.

(44) Lim, M. C. H.; Ayoko, G. A.; Morawska, L.; Ristovski, Z. D.;Jayarantne, E. R. Effect of fuel composition and engine operatingconditions on polycyclic aromatic hydrocarbon emissions froma fleet of heavy-duty diesel buses. Atmos. Environ. 2005, 39,7836–7848.

(45) Karavalakis, G.; Fontaras, G.; Ampatzoglou, D.; Kousoulidou,M.; Stournas, S.; Samaras, Z.; Bakeas, E. Effects of lowconcentration biodiesel blends application on modern pas-senger cars. Part 3: Impact on PAH, nitro-PAH, and oxy-PAHemissions. Environ. Pollut. 2010, 158, 1584–1594.

(46) McDonald, G. D.; Barr, E. B.; White, R. K. Design, characteriza-tion, and evaluation of a small-scale diesel exhaust exposuresystem. Aerosol Sci. Technol. 2004, 38, 62–78.

(47) Ballesteros, R.; Hernandez, J. J.; Lyons, L. L. An experimentalstudy of the influence of biofuel origin on particle-associatedPAH emissions. Atmos. Environ. 2010, 44, 930–938.

(48) Brandenberger, S.; Mohr, M.; Grob, M.; Neukom, H. P. Con-tribution of unburned lubricating oil and diesel fuel to particulateemission from passenger cars. Atmos. Environ. 2005, 39,6985–6994.

ES100831J

5312 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 13, 2010