Effect of Water/Fuel Emulsions anda Cerium-Based CombustionImprover Additive on HD and LDDiesel Exhaust EmissionsA R I A N N A F A R F A L E T T I ,C O V A D O N G A A S T O R G A ,G I O R G I O M A R T I N I , * U R B A N O M A N F R E D I ,A N N E M U E L L E R , M A R I A R E Y ,G I O V A N N I D E S A N T I ,A L O I S K R A S E N B R I N K , A N DB O R . L A R S E N *

EU Joint Research Centre Ispra, Institute for Environmentand Sustainability, Emissions and Health Unit,21020 Ispra (VA), Italy

One of the major technological challenges for the transportsector is to cut emissions of particulate matter (PM) andnitrogen oxides (NOx) simultaneously from diesel vehicles tomeet future emission standards and to reduce theircontribution to the pollution of ambient air. Installation ofparticle filters in all existing diesel vehicles (for new vehicles,the feasibility is proven) is an efficient but expensiveand complicated solution; thus other short-term alternativeshave been proposed. It is well known that water/diesel (W/D) emulsions with up to 20% water can reduce PM andNOx emissions in heavy-duty (HD) engines. The amount ofwater that can be used in emulsions for the technicallymore susceptible light-duty (LD) vehicles is much lower, dueto risks of impairing engine performance and durability.The present study investigates the potential emissionreductions of an experimental 6% W/D emulsion with EURO-3LD diesel vehicles in comparison to a commercial 12%W/D emulsion with a EURO-3 HD engine and to a Cerium-based combustion improver additive. For PM, theemulsions reduced the emissions with -32% for LDvehicles (mass/km) and -59% for the HD engine (mass/kWh). However, NOx emissions remained unchanged, andemissions of other pollutants were actually increasedfor the LD vehicles with +26% for hydrocarbons (HC), +18%for CO, and +25% for PM-associated benzo[a]pyrenetoxicity equivalents (TEQ). In contrast, CO (-32%), TEQ(-14%), and NOx (-6%) were reduced by the emulsion forthe HD engine, and only hydrocarbons were slightlyincreased (+16%). Whereas the Cerium-based additivewas inefficient in the HD engine for all emissions exceptfor TEQ (-39%), it markedly reduced all emissions for the LDvehicles (PM -13%, CO -18%, HC -26%, TEQ -25%)except for NOx, which remained unchanged. The presenteddata indicate a strong potential for reductions in PMemissions from current diesel engines by optimizing thefuel composition.

IntroductionParticulate matter (PM) is a major environmental problemin urban environments. A number of studies have pointedto adverse health effects of particulate matter with a diameterbelow 10 µm (PM10). Limit values for PM mass concentrationsin ambient air, expressed as PM10, are frequently exceededin major cities, and the World Health Organization (WHO)has worked out that in Europe more than 100 000 peoplesuffer premature death every year from effects of air pollution,with PM10 being of major importance (1). A recent majorpublic health impact study of PM10 in 19 European cities,on a population of 32 million inhabitants, has estimated thatreducing long-term exposure to PM10 concentrations by only5 µg/m3 in these cities may prevent between 3300 and 7700early deaths annually. (2). What is more, a recent majorepidemiological study in 23 North-American metropolitanareas has demonstrated that PM10 pollution even plays arole in post neonatal infant mortality, such as the suddendeath syndrome and respiratory diseases (3). Exhaust fromheavy-duty (HD) and light-duty (LD) diesel vehicles is animportant source of PM, and, even though emission standardshave been designed to curb the pollution from these trafficsources, they are still counted as the major contributors toPM in urban environments (4). Epidemiological studies ineight major European cities conclude that cardiac admissionsare likely to be mainly attributable to diesel exhaust (5).

In the diesel sector, engine technology is in continuousdevelopment, and a number of after-treatment systems havebeen found, which, in combination with an enhanced fuelquality, reduce emissions of PM, such as the oxidative catalystand the particle trap. Today, the oxidative catalyst isincorporated in all modern LD diesel vehicles. The particletrap is a technically feasible solution not yet implementedin all diesel vehicles. To meet future, stricter emissionstandards, the particle trap seems unavoidable for LD andcould be one of the future options for HD vehicles. However,there is a notable time lag between the introduction of a newemission standard, or a new engine technology, and therenewal of the vehicle fleet, and, as a short term alternativeto the particle trap, new fuel formulations may be employedto abate particle emissions. The emulsion of conventionaldiesel fuel and water, produced by addition of a small amountof water and an appropriate surfactant to the fuel, has recentlygained popularity for HD diesel engines, especially in thepublic transport sector. The cooling effect obtained by thepresence of water in the combustion chamber is particularlyefficient in suppressing the formation of NOx and may alsoplay a role for PM. Additionally, the improved nebulizationin the combustion chamber during the injection phase, byso-called micro-explosions of the emulsion, and the delay itcauses in the ignition may play a role. Finally, water may alsohave an influence on the chemical reactions forming soot(6-13). LD diesel vehicle emulsion formulations have beenproposed. However, a reduced amount of water in theemulsion is a necessity for this type of engine to avoidimpairing performance and durability. With the addition ofless water to the emulsions for LD applications, it can beexpected that gains obtained in emission reduction wouldbe lower. To date, this has not been fully investigated. Anothershort-term alternative to the particle trap is the use of metal-based additives (14). For the present study, we have testeda cerium-containing multifunction additive package withCetane improver, detergent, and a cold-flow propertiesimprover. This additive represents the typical compositionof similar products on the market and is designed to improvethe combustion process (14, 15).

* Address correspondence to either author. Phone: +39-0332-789293 (G. M., vehicle testing); +39-0332-789647 (B.R.L., chemicalmeasurements). Fax: +39-0332-789259 (G.M.); +39-0332-789259(B.R.L.). E-mail: giorgio.martini@jrc.it (G.M.); bo.larsen@jrc.it (B.R.L.).

Environ. Sci. Technol. 2005, 39, 6792-6799

6792 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 17, 2005 10.1021/es048345v CCC: $30.25 2005 American Chemical SocietyPublished on Web 07/27/2005

Diesel exhaust contains a group of regulated constituents,that is, PM, carbon monoxide (CO), hydrocarbons (HC), andNOx, and a number of compounds that are not specified bythe emission standards. The analysis of such unregulatedcompounds as individual species in diesel exhaust isimportant, either due to their carcinogenic properties, whichis the case for polycyclic aromatic hydrocarbons (PAH, 16),or due to their potential for photochemical ozone formation,such as the volatile organic compounds (VOC, 17). Underthe framework of collaboration between Regione Lombardia,Italy, and the European Union Joint Research Centre, theeffect of alternative, reformulated, and modified fuels onpollutant emissions from diesel vehicles is investigated inour laboratories. In this paper, the effects of emulsions andfuel additives on emissions of PM, NOx, CO, PAH, and VOCfrom LD and HD diesel engines are studied. Emission factorsare presented for each pollutant for a fleet of seven commonEURO-3 LD diesel vehicles as mass per driven km on thedynamometer with the New European Driving Cycle (NEDC,18) and as mass per produced kWh by a EURO-3 HD dieselengine (commonly found in the European HD fleet), duringthe European Stationary Cycle (ESC, 18) on the dynamic testbench.

Experimental SectionTest Vehicles and Engines. Seven different EURO-3 LD dieselvehicles, with an approximate 2 L displacement and exhaustgas recirculation (EGR), were used and compared to acommon EURO-2 LD diesel vehicle, representing the com-mon rail system, the unit injector technology, and the oldrotary pump technology (detailed information can be foundin Table S1 of the Supporting Information). As representativeof the European heavy-duty sector, a EURO-3 HD enginewith a 10 L displacement 6 cylinder, no EGR, and a maximumpower of 316 kW was selected.

Test Fuels, Emulsions, and Additives. Whereas diesel/water (D/W) emulsions for heavy-duty applications have beenmarketed for some years in various countries, D/W emulsionsfor light-duty applications are not currently available. In thepresent study, we used a commercial HD emulsion (GECAM)and two experimental emulsions specially formulated forLD vehicles.

The HD emulsion contained 12% water. The LD emulsionswere comprised of two commercial diesel fuels (fuel 1 and2) and two experimental emulsions (detailed informationcan be found in Table S2 of the Supporting Information)with a water content of 6%. A larger PM reduction effectwould be expected by using higher water content. However,preliminary tests have demonstrated that more than 6%blends significantly increase the risk of impairing engineperformance and durability. The two base fuels of which theemulsions were made had very different physical properties:fuel 1 had a density close to the upper limit of the legislativerange (0.820-0.845 g/cm3), while fuel 2 was close to the lowerlimit. Both fuels were low in sulfur content, especially fuel1 (less than 10 ppm). The two fuels differed also in thedistillation curve and for the aromatic content. The twoemulsions were produced using the two base fuels and anidentical formulation; however, the final blends of the twoemulsions differed for the actual values of the Cetaneimprover and water content. The base fuel for the HDemulsion study (fuel 3) had a low content of sulfur, a relativelylow density (0.826 kg/L), and a relatively high Cetane number(53.5). The addition of water resulted in an increased density(0.848 kg/L) and a decreased Cetane number, the later ofwhich was countered by the additions of approximately 2 gof Cetane improver per kg of base fuel. Preliminary testsconfirmed that it was not possible to reach the same poweroutput with emulsions as with the water free base fuel. Hence,to ensure comparability of the emission results, the base fuel

was tested using the same ESC cycle derived from themaximum power curve of emulsions.

For the additive studies, the test fuels (fuel 3 for HD andfuel 4 for LD) were commercial, low sulfur diesel fuels havinga relatively low density and a relatively low content ofaromatics as compared to the market average. The AMF-ALFA additive is made of two main components: 85% multi-function package (Cetane improver, cold properties en-hancer, detergent) and 15% combustion improver catalyst,consisting of a mixture of organo-metallic compounds basedmainly on Cerium. The additive was provided to the JRCwith the two components separated; they were then blendedand added to the fuel, following the recommendations ofthe producer. After the two components of the additive hadbeen blended together, the additive was added to the fuel at0.35% vol. Detailed information on the test fuels can be foundin Table S2 of the Supporting Information.

Measurements of Regulated Emissions. Regulated pol-lutant emissions from LD vehicles were measured using achassis dynamometer and a conventional constant volumesampling (CVS) system with a critical flow Venturi. The CVSwas equipped with four critical orifices that allow the selectionof the most appropriate flow rate. The roller bench of thechassis dynamometer was a 48” single roller type. To followthe legislative cycle, the driver was assisted by a driver aidsystem. CO, NOx, and PM emissions were measured ondiluted exhausts, while HC was measured continuously onraw exhaust using a heated sampling line. All exhaustsampling was carried out on diluted exhausts in a dilutiontunnel (average dilution ratio approximately 100). Theregulated emissions were measured as follows: carbonmonoxide (CO) with a nondispersive infrared (NDIR) ana-lyzer, total unburned hydrocarbons (HC) with a flameionization detector (FID), and oxides of nitrogen (NOx) witha chemiluminiscense analyzer (CLA) using a NO2 to NOconverter. Particulate samples were collected according tothe legislative procedure for diesel vehicles using Teflon-coated glass fiber filters (Pallflex T60A20), and the mass wasdetermined by weighing. Regulated pollutant emissions fromthe HD engine were measured using a full flow dilution tunneland a CVS system. Particulate mass was measured byweighing, and the diluted exhaust was sampled from thesecondary dilution tunnel.

Measurements of Unregulated Emissions. The charac-terization of unregulated pollutants in diesel exhaust wascarried out as described in detail in the Supportive Informa-tion and the cited references: the mass/size distribution wasmeasured using a 12-stage low-pressure impactor (LPI); theparticulate-associated PAH was determined by gas chro-matography-mass spectrometry (GC-MS) based upon themethods EPA TO13 and the ISO/DIN 12844 (19) and validatedby a successful participation in an interlaboratory comparisonusing the reference material (NIST SRM 1650 soot) and PMfrom vehicle exhaust (20). For data reduction, the toxicityequivalency factor (TEF) approach was used, in which eachindividual PAH is assigned a toxicity rating relative to benzo-[a]pyrene that is set to unity. The benzo[a]pyrene toxicityequivalents (TEQ) of a given sample are calculated as thesum of the concentrations multiplied by TEF over all of themeasured compounds (21). The TEF values used are listedin the Supporting Information; individual ozone precursorVOCs were sampled from the dilution tunnel and ac-cumulated for each test-cycle in Tedlar bags and analyzedby dual column GC with flame ionization detection (22). Astandard mixture obtained from the National Physics Labo-ratory (UK) containing known amounts of all determinedVOCs (23) was used for response factor calculations and fordetermination of retention times. This mixture contains all30 ozone precursor VOCs (C2-C9 hydrocarbons) specified inthe European Ozone Air Quality Directive 2002/3/EC. The

VOL. 39, NO. 17, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 6793

Maximum Incremental Reactivity (MIR) approach (17) wasused to estimate the potential impact of the emitted VOCmixture on ozone formation. Details on VOCs and MIR aregiven in the Supporting Information. Cerium was analyzedby high-resolution inductively coupled plasma mass spec-trometry (HR-ICP-MS) to study the fate of the metallic ceriumcontained in the combustion improver additive. The instru-mentation and the sample preparation followed proceduresdescribed for ambient particulate elsewhere (24) with theacquisition settings for cerium (25).

Statistical Analysis. The results of the measurements ofregulated and unregulated emissions were subjected toanalysis of variance (ANOVA) using the STATISTICA softwareversion 6 from StatSoft (Tulsa, OK). For the HD study,individual ANOVAs were conducted with the factors “emul-sion” (with/without) and “additive” (with/without) to analyzethe effect on the emissions per produced kWh (PM/NOx/CO/HC/TEQ). Each cell consisted of 3-4 repetitions of theemission measurement. For the LD study, individual ANOVAswere conducted with the factors “emulsion” (with/without)and “additive” (with/without) to analyze the effect on theemissions per driven km (PM/NOx/CO/HC/TEQ). In theemulsion study, the base fuel type (fuel 1/fuel 2) was usedas an additional factor. The inter test-vehicle variation was2-30 times higher than the intra test-vehicle variance; thusthe average emission result for each test vehicle was used asrepetitions in the ANOVA cells (n ) 3 for the emulsion study;n ) 4 for the additive study). All ANOVAs revealed significantinteractive effects between factors, which imply that the effectof emulsion or additive was not the same for all types ofemissions. Thus, it was necessary to conduct two-sided t-teststo evaluate the mean effect for each individual emission. Nointeractions were found for the factor “fuel type”, and thusdata were pooled for fuel 1 and fuel 2 (n ) 6) before t-testing.In Figure 1, the results of the statistical tests are plotted asthe mean effect of using emulsions or additive, the 95%confidence intervals as bars, and the t-test p-value when pis less or equal to 0.05.

Results and DiscussionAll regulated emissions measured in the present study (Table1) were below the limits of the EURO-3 standards, which forHD engines are 100 mg/kWh for PM, 660 mg/kWh for HC,2100 mg/kWh for CO, and 5000 mg/kWh for NOx and for theLD vehicles are 50 mg/km for PM, 640 mg/km for CO, and560 mg/km for HC + NOx. Not even in the cases where theuse of W/D emulsions increased emissions were these limitsexceeded. The emission factors for unregulated individualcompounds measured in the present study using base fuelwithout emulsion or combustion improver additive areshown in Table 2.

Effects of W/D Emulsion - HD Engine. The results onthe regulated and unregulated emissions are presented inTable 1. The measurements were highly reproducible withcoefficients of variance in the order of 1-2% for the regulatedemissions and 4-13% for the TEQ emissions, which comparesfavorably with literature (e.g., 26, 27). The 95% confidenceinterval of the TEQ emissions was somewhat higher than the10-18% variability typical for the analytical method for PAHs(19, 20), which also reflects variability in the combustionprocesses controlling PAH emissions. The effect of theemulsion on the emissions is shown in Figure 1. A significantdrop in emissions was obtained for PM (-59%) and CO(-32%) together with a minor reduction in emissions of NOx

(-6%). The only drawback was a slight increase in theemission of hydrocarbon (+16%). A number of studies haveshown the same trend of beneficial effects by D/W emulsionson particulate emissions and NOx emissions, but withvariability in the magnitude of the effect (7-12). For CO,however, no clear conclusion can be drawn from thesestudies. A very comprehensive study has been publishedrecently by the U.S. Environmental Protection Agency (13)in which emissions data were obtained from a wide rangeof conditions including engine type and model year, on andoff road applications, with and without after treatmentemission controls. With the investigated commercial W/Demulsion (20% water), emissions of PM and NOx were reducedon average, by 58% and 14%, respectively, and ozoneprecursor reactive organic gas (ROG) emissions were in-creased by 87%. In a multi-media assessment of these results,it has been pointed out that the emissions of ROG by use ofthe W/D emulsion are about 29% of the NOx emissions indiesel exhaust, which in other words means that for each tonROG increased, NOx will be reduced by 3.4 tons (28). Thus,when evaluating the emission effects on an absolute basis,mass emission reductions for NOx by use of emulsions wasgreater than mass emission increases of total HC.

The PAH emission factors (Table 2) are lower thanemissions we have previously found for a 10L EURO-3 HDD-DI engine (sum of PAH, 20 ( 1.1 µg/kWh; TEQ, 0.13 ( 0.07µg/kWh; average ( 95% confidence interval of modes 3, 5,and 8 of the ECE R-49 procedure) and 2 orders of magnitudelower than emissions from a EURO-1 type 6.6.L HD engine(37) (sum of PAH, 311 ( 12 µg/kWh; TEQ recalculated fromthe original data, 16 ( 0.6 µg/kWh - average ( SD of tworeplicates with the transient US HD Federal Test Procedure).The sum of PAH adsorbed to the emitted PM was increasedfrom 111 ( 18 to 158 ( 13 µg/g by use of emulsion, with anenrichment for the 4-6 ring compounds (data not shown)resulting in an increase in TEQ concentration on the PMfrom 1.1 ( 0.12 to 2.3 ( 0.14 µg/g. However, the overall effectof increasing the PAH concentrations on the particles andat the same time reducing the particle emission by use ofemulsion was a slight cutback of the TEQ emissions (ng/kWh) of 14% that, however, resulted statistically insignificant(p ) 0.22). The only other study, which addresses PAH, thatthe authors of the present paper are aware off is an earlystudy of the experiences with water in gas oil diesel fuels (7),which showed that employing emulsion fuels in aged, direct

FIGURE 1. Effect (fleet average ( 95% confidence interval) ofemulsion (dark gray) and combustion improver additive (light gray)on emissions from a EURO-3 HD diesel engine (top) and EURO-3 LDdiesel vehicles (bottom). The t-test p-value is indicated, when lowerthan 0.05.

6794 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 17, 2005

injection diesel engines is environmentally beneficial becauseof a substantial reduction in PM, NOx, and PAH emissionsat the expense of a slight increase in CO and HC emissions.Many studies have demonstrated that a quantifiable rela-tionship between diesel fuel features, composition of emis-sions, and their biological effects exists (e.g., 29-33).Westerholm and Li, 1994, found that the fuel content of PAHsand aromatics (and sulfur) have the strongest influence onPAH emission (29), and, today, it is generally accepted thatthe main sources of PAHs in diesel exhaust are noncombustedPAHs (34) from fuel, pyrosynthesis during combustion, andmodification of one PAH into another (35). The levels of PAHsand TEQ concentrations on the PM emitted in the presentstudy by the EURO-3 HD engine without using emulsion arecomparable to the levels we have previously found for a 10L EURO-3 HD D-DI engine (211 µg/g PAH - sum of samecompounds as in the present study; 1.4 µg/g TEQ) and thelevels published by others (36) for a 12 L EURO-2 HD D-DIengine using a purified diesel fuel (161 µg/g PAH - sum ofsame compounds as in the present study; 0.85 µg/g TEQ -recalculated from the original data). These levels are sig-nificantly lower than levels published with standard dieselsfuels (36) for a 12 L EURO-2 HD D-DI engine (946 µg/g PAH

- sum of same compounds as in the present study; 4.6 µg/gTEQ - recalculated from the original data) and published(33) for a 7.8 L EURO-2 HD D-DI engine (4.0 µg/g TEQ -recalculated from the original data), and they are orders ofmagnitudes lower than emissions from a EURO-1 type 6.6L HD engine (37) (400-750 µg/g PAH - sum of samecompounds as in the present study; 20-65 µg/g TEQ -recalculated from the original data) and from a EURO-1 type7.3 L HD engine (38) (400 µg/g PAH - sum of samecompounds as in the present study).

Effects of W/D Emulsion - LD Vehicles. A seen in Table1, the emission results were repeatable over the vehicle fleetused in this study (cars 1-3) with relative 95% confidenceintervals in the order of 5-10% for PM, 15-40% for HC,30-40% for CO, 35-40% for NOx, and 25-50% for the TEQemissions. The measurement reproducibility for individualcars was in the order of 2-4% for PM, 2-7% for CH, 1-4%for CO, 1-2% for NOx, and 10-15% for the TEQ emissions.The effect of the emulsions on the emissions was onlymarginally influenced by the quality of the base diesel fueldespite the differences in the basic properties. The onlydifferences were seen for TEQ and, to a small extent, for PM,for which fuel 2 and emulsion 2 produced lower emissions

TABLE 1. Emissiona of PM, HC, CO, NOx, and B(a)P TEQ from EURO-3 LD Diesel Vehicles and HD Engine with and withoutEmulsions and Metal Additive Combustion Improver

regulated emissions (mg/km) from LD diesel vehicles regulated emissions (mg/kWh) from HD diesel

base fuel +emulsion fuel 4 +additive fuel 3b +emulsion fuel 3 +additive

PM fuel 1 40 ( 3 26 ( 1 33 ( 11 28 ( 11 88 ( 2 36 ( 1 69 ( 1 70 ( 1fuel 2 35 ( 3 21 ( 2

HC fuel 1 44 ( 7 59 ( 7 74 ( 53 54 ( 48 146 ( 2 167 ( 2 115 ( 2 120 ( 3fuel 2 53 ( 21 65 ( 22

CO fuel 1 376 ( 130 457 ( 125 518 ( 348 416 ( 305 665 ( 7 454 ( 3 715 ( 7 745 ( 10fuel 2 371 ( 100 446 ( 130

NOx fuel 1 510 ( 180 525 ( 180 415 ( 150 431 ( 195 4435 ( 80 4175 ( 40 4185 ( 70 4150 ( 40fuel 2 514 ( 170 488 ( 190

unregulated emissions (ng/km) from LD diesel unregulated emissions (ng/kWh) from HD diesel

base fuel +emulsion base fuel +additive fuel 3 +emulsion fuel 3 +additive

TEQ, B(a)P equiv fuel 1 449 ( 230 632 ( 340 139 ( 10 92 ( 23 98 ( 11 84 ( 5 70 ( 9 43 ( 2fuel 2 256 ( 120 305 ( 80

a All values indicate the fleet average ( 95% confidence interval. b This test was made with the same max power curve as for emulsion; seetext for explanations.

TABLE 2. Emission Factors for Unregulated Individual Compounds from EURO-3 LD Vehicles and HD Engine

EURO-3 LD vehiclesa EURO-3 HD engineb EURO-3 LD vehiclesc

particle-bound polyaromatic hydrocarbons gas-phase volatile organic compounds

µg/km µg/kWh mg/km mg/km

F 0.46 ( 0.13 0.11 ( 0.04 ethane 0.9 ( 0.34 n-pentane 0.03 ( 0.04Phen 4.5 ( 1.7 2.6 ( 0.93 ethene 11 ( 4.3 trans-2-pentene 0.03 ( 0.08A 0.11 ( 0.07 0.06 ( 0.09 propane 0.17 ( 0.07 cis-2-pentene <0.02Fl 1.6 ( 0.42 2.2 ( 0.50 propene 5.5 ( 1.6 2-methylpentane 0.16 ( 0.32P 1.4 ( 0.41 2.7 ( 0.85 acetylene 2.1 ( 0.79 3-methylpentane+c-hex 0.27 ( 0.23B(a)A 0.10 ( 0.03 0.18 ( 0.06 isobutane <0.02 isoprene 0.04 ( 0.12Chr 0.41 ( 0.16 0.51 ( 0.17 n-butane 0.15 ( 0.23 n-hexane <0.02B(b)Fl 0.13 ( 0.03 0.14 ( 0.03 trans-2-butene 0.29 ( 0.32 n-heptane 0.03 ( 0.09B(k)Fl 0.10 ( 0.04 0.009 ( 0.013 1-butene 1.2 ( 0.77 benzene 0.61 ( 0.23B(a)P 0.11 ( 0.06 0.007 ( 0.005 isobutene 0.43 ( 0.13 toluene 0.31 ( 0.17Ind(123cd)P 0.08 ( 0.06 0.13 ( 0.03 cis-2-butene 0.03 ( 0.07 ethylbenzene 0.11 ( 0.32diB(ah)A 0.02 ( 0.03 0.005 ( 0.005 propyne <0.02 m+p-xylene 0.07 ( 0.19B(ghi)Per 0.14 ( 0.06 0.14 ( 0.01 isopentane 0.05 ( 0.06 o-xylene <0.02

1,3-butadiene 0.07 ( 0.10sum of PAH 9.1 ( 3.4 8.7 ( 2.2B(a)P-TEQ 0.28 ( 0.17 0.090 ( 0.017 sum of VOC 23.7 ( 9.4

a Average ( 95% confidence interval for seven vehicles (car 1-Ccar 7) and three diesel fuels (fuel 1, fuel 2, fuel 4). b Average ( 95% confidenceinterval for eight repetitions (fuel 3). c Average ( 95% confidence interval for three vehicles (car 1-car 3) and two diesel fuels (fuel 1, fuel 2).

VOL. 39, NO. 17, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 6795

than fuel 1 and emulsion 1, attributable to its lower contentof polyaromatics and perhaps its lower density.

The effect of the emulsion on the emissions is shown inFigure 1. A significant drop (-32%) in the PM mass emissions(mg/km) was obtained without changing the particle sizedistribution (data shown in information) for the LD vehicles.This reduction is somewhat smaller than the one obtainedfor the HD engine and is most probably explained by thelower water content of the W/D emulsion used for the LDdiesel vehicles. In contrast to the HD engine, the emulsionhad no effect on NOx for LD diesel vehicles and even increasedemissions of CO (+18%), TEQ (+25%), and hydrocarbons(+26%). We are not aware of any other published study inwhich W/D emulsions are used for LD diesel vehicles, whichmakes it difficult to evaluate our results. The PAH emissionfactors (sum of PAH, 9.1 ( 3.4 µg/km; TEQ, 0.28 ( 0.17 µg/km) are somewhat lower than the emissions we obtained forthe EURO-2 vehicle, car 8 (sum of PAH, 11 ( 3.4 µg/km; TEQ,0.36 ( 0.17 µg/km, data not shown), and those we havepreviously determined (20) for a 2.5 L EURO-2 LD IDI engine(no catalyst) and a 1.9 L EURO-2 LD DI engine (oxy-catalyst)fueled with a CEC RF-73-A-93 reference fuel (19-120 µg/kmPAH; 0.13-1.2 µg/km TEQ). Emission factors for EURO-1and older types of LD vehicles are orders of magnitudeshigher, that is, from 160 to 1800 µg/km for the sum of thesame individual PAH as in the present study and from 14 to50 µg/km for TEQ (recalculated data for the LD-diesel vehiclefleet of Denver, Colorado region summer 1996 and winter1997 (39)) and from 450 to 3600 µg/km for the sum of thesame individual PAH as in the present study and from 3 to117 µg/km for TEQ for four 2.8-3.1 L EURO-1 type LD-dieselvehicles fueled with Brazilian commercial diesel (40).

To better understand the TEQ data, PAH emissions weremeasured separately for the urban phase and the extra-urbanphase of the European Driving Cycle (Figure 3, top). Thesame was also done for the regulated emissions (data notshown), which revealed that the stop-go driving conditionsof the urban phase provoked incomplete combustion of thediesel fuel. In all three test vehicles, the incomplete combus-tion resulted in a drastic increase in CO emissions (factor >30) and HC emissions (factor 5-20), without affectingsignificantly the emitted amount of NOx and PM per drivenkm. However, as it appears in Figure 3 (top), the quality ofthe PM was deteriorated inasmuch as the profile of theassociated PAH was shifted toward drastically higher con-centrations of the most toxic congeners (factors 10-30). Thedeteriorating effect on the exhaust of the incompletecombustion was most pronounced for fuel 1 with its highercontent of polyaromatics and was accentuated by the use ofemulsions (Figure 3, center). For the combined EuropeanDriving Cycle, the sum of PAH adsorbed to the emittedparticles was increased from 328 ( 51 µg/g (fleet average forboth base fuels ( 95% confidence interval) to 509 ( 109 µg/g

by use of emulsion, with the enrichment of toxic congenersresulting in an increase in TEQ concentration on the PMfrom 8.7 ( 3.4 µg/g (fleet average for both base fuels ( 95%confidence interval) to 18.0 ( 7.2 µg/g. This explains theoverall effect of a 25% increase in benzo[a]pyrene TEQemissions (ng/km) despite the cutback of the PM emissionsby 32%.

The different effect of emulsions for the LD vehicles ascompared to the HD engine is interesting and may simplybe due to the use of different injection systems (unit injector,common rail vs unit pump) or a potential deactivation bywater of the oxidation catalyst, which was only used in theLD vehicles. Nevertheless, the latter possibility seems to beruled out by the fact that the temperature profile of theexhaust gas, which is highly influenced by the catalyst activity,was practically identical for all individual tests independentof emulsions or fuel quality (data not shown). EGR is designedto lower NOx emissions in diesel engines on the expense ofa minor increase in PM emissions (9, 12), while the water inthe emulsion lowers mainly PM. EGR was used in the LDvehicles but not in the HD engine, which may also explainthe different response to emulsions. However, it is possiblethat the fact that the HD engine was tested with a static cycleas opposed to the LD vehicles may be the main reason forthe different response. The considerably higher levels of PAHs

FIGURE 2. Emissions (fleet average ( 95% confidence interval) ofozone precursor VOCs from EURO-3 LD diesel vehicles with andwithout emulsions.

FIGURE 3. Emissions (overall mean for fuels and fleet ( 95%confidence interval) of particle associated PAH from EURO-3 LDvehicles. Top: effect of driving cycle (base fuels). Center: effectof emulsion (combined driving cycle). Bottom: effect of combustionimprover additive (combined driving cycle).

6796 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 17, 2005

and TEQ concentrations on the PM emitted by the EURO-3LD vehicles as compared to the HD engine point to the latterexplanation and is, most likely, generated by the incompletecombustion during the stop-go driving conditions of theurban phase for the LD vehicles (326 ( 33 µg/g PAHascompared to 111 ( 18 µg/g PAH and 7.1 ( 2.6 µg/g TEQas compared to 1.1 ( 0.11 µg/g TEQ - data obtained withoutemulsion). The levels emitted by the EURO-3 LD vehiclesare similar to those we determined for the EURO-2 LD vehicle,car 8 (384 ( 54 µg/g PAH; 2.9 ( 0.1 µg/g TEQ, average ( 95%confidence interval of triplicate measurements, data notshown) and those we have previously determined (20) for a2.5 L EURO-2 LD IDI engine (no catalyst) and a 1.9 L EURO-2LD DI engine (oxy-catalyst) fuelled with a CEC RF-73-A-93reference fuel (313 ( 214 µg/g PAH; 2.9 ( 2.6 µg/g TEQ,average ( SD). As expected, published levels for types of LDengines older than the EURO-2 types are much higher, thatis, 460-1890 µg/g for the sum of the same individual PAHas in the present study and 40-53 µg/g for TEQ (recalculateddata for the LD-diesel vehicle fleet of Denver, Colorado regionsummer 1996 and winter 1997 (39)) and 4700 ( 3200 µg/gfor the sum of the same individual PAH as in the presentstudy and 66 ( 36 µg/g for TEQ (recalculated average ( 95%confidence interval of four 2.8-3.1 L EURO-1 type LD-dieselvehicles fueled with Brazilian commercial diesel (40)).

The emission factors for benzene (0.28 ( 0.17 mg/km)and 1,3-butadiene (0.07 ( 0.10 mg/km) for the EURO-3 LDvehicles fall in a range that would be expected whencomparing with typical emission factors published by theUK National Atmospheric Emission Inventory (44) forEURO-2 LD vehicles (benzene, 0.43 ( 0.27 mg/km; 1,3-butadiene, 0.20 ( 0.14 mg/km), for EURO-1 LD vehicles(benzene, 0.58 ( 0.42 mg/km; 1,3-butadiene, 0.28 ( 0.20mg/km), and for pre-Euro LD vehicles (benzene, 1.8 ( 0.9mg/km; 1,3-butadiene, 0.88 ( 0.41 mg/km). To betterunderstand the impact of the W/D emulsion on HC emis-sions, gas chromatographic analysis was performed onsamples collected in Tedlar bags for the complete NEDCtesting cycles. The results revealed only small differences inthe composition of the ozone precursor VOCs with andwithout use of emulsion (Figure 2), which translated the 26%increase in HC emissions into a 29% increased ozoneformation potential of the exhaust, that is, from 0.19 ( 0.07g O3/km to 0.25 ( 0.08 g O3/km (fleet average for both basefuels ( 95% confidence interval) as calculated by the MIRapproach (17) not considering the increased CO emissions.Among the VOCs, benzene and 1,3-butadiene are classifiedas carcinogenic to humans (41). Even though VOCs, benzene,and 1,3-butadiene in ambient air are traditionally attributedto gasoline emissions, some urban areas are stronglyinfluenced by mixed LD and HD traffic, and significantamounts of VOCs, including benzene and 1,3-butadiene inthe air, have been apportioned to these types of sources (22).The gas chromatographic analyses showed that the 26%increase in HC emissions caused by the use of the W/Demulsion corresponded to a similar increase (23%) in benzeneemissions from 0.61 ( 0.14 mg/km to 0.76 ( 0.15 (fleet averagefor both base fuels ( 95% confidence interval), with a slightlyhigher effect for fuel 1 than fuel 2 attributable to the highercontent of aromatic compounds in the former. The 26%increase in HC emissions did not result in any significantchange in the 1,3-butadiene emissions.

When the positive effect of the W/D emulsion in reducingPM emissions is weighed up against the lack of effects forNOx and the negative effects on emissions of CO, HC, andbenzo[a]pyrene toxicity equivalents, it remains unclearwhether the use of emulsions can be recommended for LDapplications.

Effect of Combustion Improver Additive. Cerium-basedadditives have previously been proven effective in reducing

emissions from ante-Euro type HD diesel engines (14) andhave been investigated in combination with electronicexhaust gas recirculation for achieving modern HD dieselemission standards (15). Today, they are incorporated incommercial systems for on-line diesel trap regeneration (42).To our best knowledge, no studies have been published onthe use of fuel-borne Cerium-based additives with modernLD vehicles. As seen in Table 1, the same high level ofreproducibility was obtained in the additive study as in theemulsion study for the HD engine. However, for the LDvehicle fleet (cars 4-7), a somewhat inferior repeatabilitywas observed with relative 95% confidence intervals in theorder of 30-40% for PM, 35-45% for CH, 35-40% for CO,40-50% for NOx, and 7-25% for the TEQ emissions, althoughthe reproducibility for individual cars was better withcoefficients of variance in the order of 2-4% for PM, 2-7%for HC, 1-4% for CO, 1-2% for NOx, and 3-30% for the TEQemissions.

With the HD engine, the combustion improver additivehad only a marginal effect on the regulated emissions (PM,+1.3%; CO, +4.2%; HC, +4.2%; NOx, -0.8%), but a significantreduction was obtained in adsorbed PAH from 122 ( 35 to42 ( 10 µg/g (fleet average ( 95% confidence interval).Moreover, the toxic congeners were depleted, which causeda reduction in TEQ concentrations on the PM from 1.0 (0.14 to 0.61 ( 0.03 µg/g (fleet average ( 95% confidenceinterval). The overall effect of the combustion improveradditive was a 39 ( 8% cutback in the benzo[a]pyrene TEQemissions (ng/kWh).

As seen in Figure 1, the additive markedly improved theregulated and unregulated emissions for the investigatedEURO-3 LD vehicles (cars 4-7): PM (-13%), CO (-18%),HC (-26%), TEQ (-25%), except for NOx, which remainedunchanged. It is well known that an increase in the Cetanenumber leads to a more complete combustion of diesel fuelsand a lower emission of CO and HC. Thus, parts of thebeneficial effects of the additive may simply be explained bythe Cetane improver contained in the additive. Yet, thereduction in PM and the associated PAH point to the factthat also the catalytic effect of the cerium, contained in theadditive, plays an important role. For the EURO-2 LD vehicle(car 8), emissions are significantly improved by the additivefor PM (-17%) and the associated TEQ (-41%), whereas CO(-2%), HC (-4%), and NOx (-10%) were only slightlyimproved. This type of vehicle is not fitted with any exhaustafter-treatment system, which strongly indicates that thebeneficial effect of the additive is obtained in the combustionchamber.

As for the emulsion study, the mass/size distribution ofthe emitted particulates was measured using the 12-stagelow-pressure impactor (data not shown), and it was dem-onstrated that for the whole fleet (cars 4-7) the reductionin PM obtained by the combustion improver additive wasobtained over the complete size range of particles, which isof uttermost importance for the evaluation of potential healthimpact of the emitted particles, inasmuch as particles in the10-100 nm range have the greatest deposition in the alveolarregion of the respiratory tract (43).

As seen in Figure 3 (bottom), the combustion improveradditive reduced the PAH emissions for all compoundsresulting in a reduction in the sum of PAH adsorbed to theemitted particles from 337 ( 57 to 245 ( 79 µg/g (fleet average( 95% confidence interval) and even shifted the emissionprofile toward a depletion of the more toxic 5-6 ringcongeners, resulting in a reduction in TEQ concentration onthe PM from 5.5 ( 2.2 to 3.9 ( 1.2 µg/g (fleet average ( 95%confidence interval). The magnitude of this effect variedstrongly between the different vehicles/engine technologiesand between the different operating conditions of the engine.For the complete EURO-3 LD vehicle fleet (cars 4-7), the

VOL. 39, NO. 17, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 6797

additive significantly reduced the PM emission levels overthe extra-urban part of the cycle (-23%, fleet average), buta smaller effect was observed over the urban driving phase;for a few vehicles, PM emissions even slightly increased withthe use of additive. This may be explained by the fact thatthe combustion of the additive itself produces ash, whichcontributes to the total mass of PM collected on the filter,and which may counteract the effect of a reduction ofcarbonaceous matter resulting in a combined increase ofthe particulate total mass. Results of HR-ICP-MS analyses ofthe cerium concentration in the emitted PM revealed residuesof 22 ( 18 mg/g cerium, corresponding to a recovery of 33( 12% of the cerium supplied to the combustion chamberwith the consumed fuel.

The different behavior of the additive over the urban andextra-urban driving phases may be due to a lower combustionimproving efficiency at lower exhaust gas temperaturesoccurring during the urban phase, which at the tailpipe outletremained below 90-100 °C during the urban part of the NEDCcycle. To investigate the influence of the exhaust temperature,some ad hoc tests were carried out using nonstandard cycles(data not shown). The results confirmed that the effect of theadditive depends on the exhaust temperature and that it ismore efficient when the exhaust temperature is higher.

In addition to the described tests, the effects of the additiveon emissions after a short mileage accumulation wereinvestigated by running the vehicles for about 300 km onroad, using fuel with additive. After the mileage accumulation,the emission tests were repeated and the results comparedto those obtained with the fuel plus additive before themileage accumulation. No significant change in emissionswas observed.

Further Perspectives. PM traps are the first and mostimportant measure to abate PM emissions from dieselengines and can reach extremely high (>99%) efficiencies.However, it appears from the results in the present paperthat the use of W/D emulsion may be a good alternative inthe transition period for HD engines, inasmuch as itsignificantly reduces emissions during the European Sta-tionary Cycle of PM, the associated benzo[a]pyrene toxicityequivalents, and CO and gives a minor reduction in emissionsof NOx. The only drawback is a slight increase in the emissionsof hydrocarbons. If these results can be confirmed in futuretests using a dynamic testing cycle, W/D emulsions may bea recommendable short term alternative abatement strategyto particle traps. Future tests should be done with HD engineswith and without after-treatment systems, such as the oxy-catalyst and exhaust gas recirculation.

The use of a 6% W/D emulsion as fuel for EURO-3 LDvehicles, which is about the upper limit to maintain properfunctioning in these more susceptible engine technologies,efficiently reduced emissions of PM during the New EuropeanDriving Cycle. However, in contrast to what happened forthe HD engine, the emulsion had no effect on NOx for LDdiesel vehicles and even increased emissions of CO, benzo-[a]pyrene toxicity equivalents, and hydrocarbons. Withoutconsidering the effects of increased CO emissions, theincrease in HC emissions corresponds to a 29% higher ozoneformation potential of the exhaust. A better short-termalternative abatement strategy to particle traps for LD vehiclesmay turn out to be a cerium-based combustion improverfuel additive, which yielded marked reductions in emissionsof PM, CO, HC, and benzo[a]pyrene toxicity equivalents.However, cerium accumulation of up to 2 wt % in the emittedPM was the unavoidable consequence of the use of thisadditive, and the possible effects thereof on human healthand the environment should be evaluated before thisabatement strategy can be fully recommended.

AcknowledgmentsThe VELA staff is acknowledged for their skillful technicalassistance, in particular, Mr. R. Colombo, Mr. M. Sculati,and Mr. G. Lanappe for the vehicle testing; Mr. A. Brunella,for the engine testing; Mr. M. Duane and Mr. V. Forcina forthe chemical analysis; Mr. P. Trincherini for the HR-ICP-MSanalysis; and Mr. K. Douglas for linguistic revision of themanuscript. CAM Tecnologie is thanked for the donation offuels, emulsions, and test vehicles.

Supporting Information AvailableDetailed information on materials and methods. This materialis available free of charge via the Internet at http://pubs.acs.org.

Literature Cited(1) WHO. Health Aspects of Air Pollution with Particulate Matter,

Ozone and Nitrogen Dioxide; Bonn, Germany, 2003.(2) Medina, S.; Plasencia, A.; Ballester, F.; Muecke, H. G.; Schwartz,

J. Apheis: public health impact of PM10 in 19 European cities.J. Epidemiol. Commun. Health 2004, 58, 831.

(3) Kaiser, R.; Romieu, I.; Medina, S.; Schwartz, J.; Krzyzanowski,M.; Nino Kunzli, N. Air pollution attributable postneonatal infantmortality in U. S. metropolitan areas: a risk assessment study.Environ. Health: A Global Access Science Source 2004, 3, 1.

(4) EPA. Health Assessment Document for Diesel Engine ExhaustEPA/600/8-90/057F; National Center for Environmental As-sessment, Washington, DC, 2002.

(5) Le Tertre, A.; Medina, S.; Samoli, E.; Forsberg, B.; Michelozzi,P.; Boumghar, A.; Vonk, J. M.; Bellini, A.; Atkinson, R.; Ayres, J.G.; Sunyer, J.; Schwartz, J.; Katsouyanni, K. J. Epidemiol.Commun. Health 2002, 56, 773.

(6) Kadota, T.; Yamasaki, H. Recent advances in the combustionof water fuel emulsion. Progr. Energy Combust. Sci. 2002, 2, 385.

(7) Kovacs, A. Experiences with water in gas oil diesel fuels. Pet.Coal 1995, 37, 17.

(8) Crookes, R. J.; Kiannejad, F.; Nazha, M. A. A. Systematicassessment of combustion characteristics of biofuels andemulsions with water for use as diesel engine fuels. EnergyConvers. Manage. 1997, 38, 1785.

(9) Marwan, A. A.; Rajakaruna, N. H.; Wagstaff, S. A. The Use ofEmulsion, Water Induction and EGR for Controlling Diesel EngineEmissions; SAE Technical Papers 2001-01-1941, SAE, Warrendale,PA, 2001.

(10) Janakiraman, G.; Johnson, J. H.; James, S. T.; Cuong, P. W.; Khan,T. H. A.; Leddy, D. G. Oxidation Catalytic Converter andEmulsified Fuel Effects on Heavy-Duty Diesel Engine Emissions;SAE Technical Papers 2002-01-1277, SAE, Warrendale, PA, 2002.

(11) Hall, D.; Thorne, C.; Goodier, S. An Investigation Into the Effectof a Diesel/Water Emulsion on the Size and Number Distributionof the Particulate Emissions From a Heavy-Duty Diesel Engine;SAE Technical Paper 2003-01-3168, SAE, Warrendale, PA, 2003.

(12) Bertola, A.; Boulouchos, K. Influence of Water-Diesel FuelEmulsions and EGR on Combustion and Exhaust Emissions ofHeavy-Duty, Di-Diesel Engines Equipped With Common-RailInjection System; SAE Technical Paper 2003-01-3146, SAE,Warrendale, PA, 2003.

(13) EPA. Impacts of Lubrizol’s PuriNOx Water/Diesel Emulsion onExhaustEmissions fromHeavy-Duty Engines Draft TechnicalReport. Assessment and Standards Division Office of Trans-portation and Air Quality. U.S. Environmental ProtectionAgency, EPA420-P-02-007, December, 2002.

(14) Lemaire, J.; Magdi Khair, M. Effect of Cerium Fuel Additive onthe Emissions Characteristics of a Heavy-Duty Diesel Engine;SAE Technical Papers 1994-2067, SAE, Warrendale, PA, 1994.

(15) Psaras, D.; Summers, J. C.; Das, P. K.; Ceynow, K. P.; Florek, B.;Khair, M. K.; Disilverio, W. D. Achieving the 2004 Heavy-DutyDiesel Emissions Using Electronic Egr and a Cerium-Based Fuel-Borne Catalyst; SAE Technical Paper 970189, SAE, Warrendale,PA, 1997.

(16) IARC. Monographs on the Evaluation of the Carcinogenic Riskof Chemicals to Humans. Volume 34. Polynuclear AromaticCompounds; International Agency for Research on Cancer:Lyon, France, 1984.

(17) Carter, W. P. L. Development of ozone reactivity scales for volatileorganic compounds. J. Air Waste Manage. Assoc. 1994, 44, 881.

6798 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 17, 2005

(18) Concawe. Motor vehicle emission regulations and fuel speci-fications, Part 1, 2002/203 update, Report no. 9/04, Concawe,Brussels, 2003.

(19) Astorga, C.; Suurballe, A.; Bech Olsen, M.; Vialaton, D.; Dilara,P.; De Santi, G.; Larsen, B. R. Development of a protocol for theanalysis of toxic PAH, azarenes, and their nitro-, hydroxyl-, andoxo-derivatives in emission exhaust and ambient PM. EUR 20560EN. European Communities, Brussels, 2003.

(20) Larsen, B. R.; Astorga, C.; Baglio, D.; Brussol, C.; Duane, M.;Kotzias, D. Characterisation of engine exhaust particulatefingerprints and the contribution to air quality. EUR 19047 EN.European Communities, Brussels, 2000.

(21) EPA. Provisional guidance for Quantitative Risk Assessment ofPolycyclic Aromatic Hydrocarbons. U.S. Environmental Protec-tion Agency, Washington, 1989.

(22) Latella, A.; Stani, G.; Cobelli, L.; Duane, M.; Junninen, H.; Astorga,C.; Larsen, B. R. Semicontinuous GC analysis and receptormodelling for source apportionment of ozone precursor hy-drocarbons in Bresso, Milan, 2003. J. Chromatogr., A 2005, 1071,39.

(23) Larsen, B. R., Ed. Proceedings of the International Expert Meetingon Measurement of Ozone Precursor Volatile Organic Com-pounds (VOCs), Ispra (VA), Italy, March 6-7, 2003; EUR 21056EN. European Communities, Brussels, 2004.

(24) Lanza, F.; Trincherini, P. R. The determination of cadmium,lead and vanadium by high-resolution ICP-MS in Antarctic snowsamples. Ann. Chim. 2000, 90, 61.

(25) Buchmann, J. H.; de Souza Sarkis, J. E.; Rodrigues, C. Deter-mination of metals in plant samples by using a sector fieldinductively coupled plasma mass spectrometer. Sci. TotalEnviron. 2000, 263, 221.

(26) Sharp, C.; Howell, S.; Jobe, J. The effect of biodiesel fuel ontransient emissions from modern diesel engines, Part II:Unregulated emissions and chemical characterization. SAETechnical Paper No. 2000-01-1968, SAE, Warrendale, PA, 2000.

(27) Norbeck, J. M. Evaluation of factors that affect diesel exhausttoxicity; Final Report contract No. 94-312, California AirResources Board, 1998.

(28) CARB. Multi-Media Assessment of Lubrizol’s PuriNOx Water/Diesel Emulsion; California Environmental Protection Agency,California, 2004.

(29) Westerholm, R.; Li, H. A multivariate statistical analysis of fuel-related polycyclic aromatic hydrocarbon emissions from heavy-duty diesel vehicles. Environ. Sci. Technol. 1994, 28, 965.

(30) Westerholm, R. N.; Almen, J.; Lee, H.; Rannug, J. U.; Egeback,K. E.; Gragg, K. Chemical and biological characterization ofparticulate-, semivolatile-, and gas-phase-associated com-pounds in diluted heavy-duty diesel exhausts: a comparisonof three different semivolatile-phase samplers. Environ. Sci.Technol. 1991, 25, 332.

(31) Johansen, K.; Gabrielsson, P.; Stavnsbjerg, P.; Bak, F.; Andersen,E.; Autrup, H. Effect of upgraded diesel fuels and oxidationcatalysts on emission properties, especially PAH and genotoxicity;

SAE Technical Paper No. 973001, SAE, Warrendale, PA, 1997.(32) EPA. Health Assessment Document for Diesel Engine Exhaust,

Washington, DC, 2002.(33) Turrio-Baldassarri, L.; Battistellia, C. L.; Contia, L.; Crebellia, R.;

De Berardisa, B.; Iamicelia, A.; Gambino, M.; Iannaccone, S.Emission comparison of urban bus engine fueled with dieseloil and ‘biodiesel’ blend. Sci. Total Environ. 2004, 327, 147.

(34) Tancell, P. J.; Rhead, M. N.; Trier, C. J.; Bell, M. A.; Fussey, D.E. The sources of benzo[a]pyrene in diesel exhaust emissions.Sci. Total Environ. 1995, 162, 179.

(35) Concawe. Polycyclic Aromatic hydrocarbons in automotiveexhaust emissions and fuels, Report 98/55, Bruxelles, Belgium,1998.

(36) Westerholm, R. N.; Christensen, A.; Tornqvist, M.; Ehrenberg,L.; Rannug, U.; Sjogren, M.; Rafter, J.; Soontjens, C.; Almen, J.;Gragg, K. Comparison of exhaust emissions from Swedishenvironmental classified diesel fuels (MK1) and europeanprogram on emissions fuels and engine technologies (EPEFE)reference fuels: a chemical and biological characterization withviewpoint on cancer risk. Environ. Sci. Technol. 2001, 35, 1748.

(37) Chao, M.-R.; Lin, T.-C.; Chao, H.-R.; Chang, F.-H.; Chen, C.-B.Effects of methanol-containing additive on emission charac-teristics from a heavy-duty diesel engine. Sci. Total Environ.2001, 279, 167.

(38) Durbin, T. D.; Collins, J. R.; Norbeck, J. M.; Smith, M. R. Effectsof Biodiesel, Biodiesel Blends, and a Synthetic Diesel onEmissions from Light Heavy-Duty Diesel Vehicles. Environ. Sci.Technol. 2000, 34, 349.

(39) Cadle, S. H.; Mulawa, P. A.; Hunsanger, E. C.; Nelson, K.; Ragazzi,R. A.; Barrett, R.; Gallagher, G. L. Composition of Light-DutyMotor Vehicle Exhaust Particulate Matter in the Denver,Colorado Area. Environ. Sci. Technol. 1999, 33, 2328.

(40) de Abrantes, R.; de Assuncao, J. V.; Pesquero, C. R. Emission ofpolycyclic aromatic hydrocarbons from light-duty diesel vehiclesexhaust. Atmos. Environ. 2004, 38, 1631.

(41) IARC. Monographs on the evaluation of carcinogenic risks tohumans. Benzene; International Association for Research onCancer: Lyone, France, 1982.

(42) Pattas, K.; Samaras, Z.; Roumbos, A.; Lemaire, J.; Mustel, W.;Rouveirolles P. Regeneration of Dpf At Low Temperatures Withthe Use of a Cerium-Based Fuel Additive; SAE Technical PaperNo. 960135, SAE, Warrendale, PA, 1996.

(43) Hofmann, W.; Martonen, T. B.; Graham, R. C. J. Aerosol Med.1989, 2, 49.

(44) UK National Atmospheric Emission Inventory; UK Departmentof the Environment, Transport and the Regions, 2004.

Received for review October 25, 2004. Revised manuscriptreceived June 10, 2005. Accepted June 23, 2005.

ES048345V

VOL. 39, NO. 17, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 6799