Secondary Effects of CatalyticDiesel Particulate Filters:Conversion of PAHs versusFormation of Nitro-PAHsN O R B E R T V . H E E B , * , † P E T E R S C H M I D , ‡

M A R T I N K O H L E R , ‡ E R I K A G U J E R , ‡

M A R K U S Z E N N E G G , ‡ D A N I E L A W E N G E R , ‡

A D R I A N W I C H S E R , ‡ A N D R E A U L R I C H , ‡

U R S G F E L L E R , † P E T E R H O N E G G E R , §

K E R S T I N Z E Y E R , § L U K A S E M M E N E G G E R , §

J E A N - L U C P E T E R M A N N , |

J A N C Z E R W I N S K I , | T H O M A S M O S I M A N N , ⊥

M A R K U S K A S P E R , ⊥ A N DA N D R E A S M A Y E R #

Empa, Swiss Federal Laboratories for Materials Testing andResearch, Laboratory for Solid State Chemistry and Catalysis,Laboratory for Analytical Chemistry, Laboratory for AirPollution/Environmental Technology, Uberlandstrasse 129,CH-8600 Dübendorf, Switzerland, UASB, University of AppliedSciences Biel, Laboratory for Exhaust Emission Control,Gwerdtstrasse 5, CH-2560 Nidau, Switzerland, MatterEngineering AG, Bremgarterstrasse 62, CH-5610 Wohlen,Switzerland, and TTM, Technik Thermischer Maschinen,Fohrhölzlistr. 14b, CH-5443 Niederrohrdorf, Switzerland

Received October 25, 2007. Accepted February 26, 2008.

Diesel particulate filters (DPFs) are a promising technologyto detoxify diesel exhaust. However, the secondary combustionof diesel soot and associated compounds may also inducetheformationofnewpollutants.Dieselsoot isratedascarcinogenicto humans and also acts as a carrier for a variety of genotoxiccompounds such as certain polycyclic aromatic hydrocarbons(PAHs) or nitrated PAHs (nitro-PAHs). Furthermore, diesel exhaustcontains considerable amounts of nitric oxide (NO), whichcan be converted to more powerful nitrating species like nitrogendioxide (NO2), nitric acid (HNO3), and others. This mix ofcompounds may support nitration reactions in DPFs. Hereinwe report effects of two cordierite-based, monolithic, wall-flowDPFs on emissions of genotoxic PAHs and nitro-PAHs andcompare these findings with those of a reporter gene bioassaysensitive to aryl hydrocarbons (AHs). Soot combustion waseither catalyzed with an iron- or a copper/iron-based fuel additive(fuel-borne catalysts). A heavy duty diesel engine, operatedaccording to the 8-stage ISO 8178/4 C1 cycle, was used as testplatform. Emissions of all investigated 4- to 6-ring PAHswere reduced by about 40–90%, including those rated ascarcinogenic. Emissions of 1- and 2-nitronaphthalene increasedby about 20–100%. Among the 3-ring nitro-PAHs, emissionsof 3-nitrophenanthrene decreased by about 30%, whereas

9-nitrophenanthrene and 9-nitroanthracene were found onlyafter DPFs. In case of 4-ring nitro-PAHs, emissions of3-nitrofluoranthene, 1-nitropyrene, and 4-nitropyrene decreasedby about 40–60% with DPFs. Total AH-receptor (AHR) agonistconcentrations of diesel exhaust were lowered by 80–90%,when using the iron- and copper-based DPFs. The tested PAHsaccounted for <1% of the total AHR-mediated response,indicating that considerable amounts of other aryl hydrocarbonsmust be present in filtered and unfiltered exhaust. Weconclude that both DPFs detoxified diesel exhaust with respecttototalarylhydrocarbons, includingtheinvestigatedcarcinogenicPAHs, but we also noticed a secondary formation of selectednitro-PAHs. Nitration reactions were found to be stereoselectivewith a preferential substitution of hydrogen atoms at peri-positions. The stereoisomers obtained are related to combustionchemistry, but differ from those formed upon atmosphericnitration of PAHs.

IntroductionDiesel exhaust is a heterogeneous mixture of soot particles,condensate droplets, and numerous compounds, whichremain in the gas phase, dissolve in droplets, or adsorb ontoparticles. The toxicity of diesel exhaust has been investigatedfor several decades (1, 2). Based on the proven genotoxicactivity of dozens of its constituents, diesel exhaust isconsidered mutagenic and carcinogenic to humans. This hasconsiderable consequences for the application of dieselengines at workplaces like mining or tunnel constructionand for indoor-use (3). Diesel particulate filters (DPFs) haveevolved to a promising technology to detoxify diesel exhaust.Three major classes of DPFs can be distinguished: (i) porousor fibrous substrates coated with catalysts, typically noblemetals; (ii) uncoated substrates, which accumulate fuel-bornecatalysts, typically transition metal oxides; and (iii) uncoatedfilters, which use active regeneration, for example burners.In Switzerland, now widely accepted procedures (VERTprocedures) were developed to assess DPF technology suit-able for retrofitting large construction engines (>18 kW) forworkplace application. In brief, these procedures include (i)testing of filtration efficiency and penetration of nanoparticles(10–400 nm), (ii) on-site evaluation of durability and failurerates, and (iii) an assessment of toxic secondary emissionssuch as polychlorinated dibenzodioxins/furans (PCDD/Fs)and genotoxic polycyclic aromatic hydrocarbons (PAHs) andnitro-PAHs. The overall effects of DPFs on aryl hydrocarbons(AHs) were tested with a reporter gene assay sensitive toAH-receptor (AHR) agonists such as PCDD/Fs, PAHs, andnitro-PAHs (4). PAHs and nitro-PAHs will be addressed inthis contribution and the risks of a DPF-induced PCDD/Fformation have been assessed before (5).

Among other anthropogenic activities, evaporation andincomplete combustion of gasoline and diesel fuels areimportant PAH sources. Substantial amounts are found infuels (6–8), vehicle exhaust (9), and road-tunnel air (7, 10, 11).Diesel exhaust in particular, is a source of different genotoxiccompounds (12). Bioassay-directed chemical analyses leadto the identification of different carcinogenic and mutagenicPAHs, nitro-PAHs, and nitro-oxy-PAHs (13–15). Mainly 4- to7-ring PAHs were identified to be most active. Besides thesecritical PAHs, other PAHs can act as precursors for genotoxicderivatives. For example, inactive pyrene can be convertedto mutagenic nitro- or dinitropyrenes (16).

Nitration of PAHs occurs during fuel combustion orcatalytic exhaust gas treatment, as discussed herein, but

* Corresponding author phone: +41 44 823 4257; fax: +41 44 8234041; e-mail: norbert.heeb@empa.ch.

† Laboratory for Solid State Chemistry and Catalysis.‡ Laboratory for Analytical Chemistry.§ Laboratory for Air Pollution/Environmental Technology.| University of Applied Sciences Biel.⊥ Matter Engineering AG.# TTM, Technik Thermischer Maschinen.

Environ. Sci. Technol. 2008, 42, 3773–3779

10.1021/es7026949 CCC: $40.75 2008 American Chemical Society VOL. 42, NO. 10, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 3773

Published on Web 04/10/2008

similar reactions also take place in the atmosphere (17).Airborne nitro-PAHs appear to be responsible for a substantialproportion of the total direct mutagenicity of respirableparticles (17, 18). An impressive number of potential nitrationproducts have to be expected, when considering the largevariety of PAHs and alkylated PAHs found in diesel exhaust.Furthermore, several nitrating species support such trans-formation reactions (17, 18).

Exposure of electron-rich PAHs to electrophilic oxides ofnitrogen leads to nitro-PAH formation. Nitration reactionscan occur in the gas-phase with electrophiles like NO2, NO3,and N2O5. This chemistry may require an additional activationstep at ambient conditions, e.g., the attack of another radicallike OH. But nitration reactions also proceed in solution orat particle surfaces in presence of nitrating species likenitrosyl- (NO+) or nitronium (NO2

+) ions, or nitrous acid(HNO2), and nitric acid (HNO3). Thus, the lower troposphereis a complex nitration reactor, supporting gas-, liquid-, andsolid-phase chemistry.

DPFs have to be considered as chemical reactors as well.They are operated at significantly higher temperatures andreactant concentrations than those found in ambient air.Furthermore, DPFs trap compounds and elongate theirexposure against reactive species. DPFs may also accumulatecompounds acting as catalysts, e.g., metal ions from fueladditives, lubrication oils, and catalytic coatings, or nitratingspecies such as HNO2 and HNO3, and DPFs equally supportgas-, liquid-, and solid-phase chemistry.

With these similarities, we expect that DPFs must influencePAH- and nitro-PAH-profiles of diesel exhaust. We hypoth-esized that DPFs either (i) lower PAH and nitro-PAHemissions, assuming that both classes of compounds areadsorbed in DPFs, (ii) convert PAHs to nitro-PAHs, if nitrationchemistry is supported in DPFs, or (iii) induce the de novoformation of both PAHs and nitro-PAHs via soot decom-position and subsequent nitration. Thus, depending on therelevance of these processes, DPFs may either reduce orincrease the genotoxic potential of diesel exhaust.

Experimental SectionEngine, Test Cycle, Fuel, Lubricant. All engine tests werecarried out on the test stands of the University of AppliedSciences in Biel (UASB, Switzerland) using a heavy duty dieselengine with direct fuel injection (Liebherr, type 914 T, 6.11L, 4 cylinders, 105 kW, Bulle, Switzerland). The engine wasoperated in the eight-stage ISO 8178/4 C1 test cycle, whichconsists of four load stages at maximum revolutions-per-minute (RPM), three load stages at intermediate RPM (60%of max. RPM), an idling phase, and eight transients. Eachstage is held for 10 or 15 min, resulting in a total cycle timeof 100 min. The base fuel for all experiments was a commercialdiesel (class D, SN 181190–1:2000) with a density of 824.3kg/m3, a cetane number of 56.0, and a sulfur content of 16mg/kg. The content of PAHs was 3%, as determined by HPLC.The fuel neither contained iron, copper, nor chlorine abovedetection limits of 0.1, 0.1, and 2 µg/g, respectively, as deter-mined by wavelength dispersed X-ray fluorescence spec-trometry (WD-XRF, PW 2400, Phillips, Netherlands) and byinductively coupled plasma optical emission spectrometry(ICP-OES, Vista Pro, Varian, Orlando, FL). A chlorine contentof 120 µg/g was determined for the lubricant (Universal SAE,15W40) by WD-XRF analysis, but iron- and copper levelswere below 0.1 µg/g (ICP-OES). Further details are given in(5).

Particulate Filters, Catalysts, Test Configurations. Twonew, uncoated, cordierite-based, monolithic, wall-flow DPFs(Greentop, 100 cells per square inch, 22.8 L, Grävenwiesbach,Germany) were used in combination with an iron- and acopper/iron-based fuel additive (ITN, Krakow, Poland). Bothadditives were diluted with reference fuel to final iron- and

copper/iron-concentrations of 4.5 and 9.0 ( 0.5/7.5 ( 0.7µg/g, respectively, as determined by ICP-OES. Four additionalfuel blends were mixed with 1,6-dichlorohexane (Fluka,Buchs, Switzerland) as chlorine dopant (5). However, nodifferences were noticed with respect to PAH and nitro-PAHemissions with this dopant. Therefore, mean PAH and nitro-PAH emission factors are discussed herein. Experimentalconditions of all test configurations are given as SupportingInformation (Table S1). All premixed fuel blends weredelivered from barrels, minimizing cross contamination andallowing fast adaptation of engine and DPFs, both precon-ditioned for one hour at the first stage of the ISO 8178/4cycle (max. torque, full load) after each fuel change.

Exhaust Sampling, Chemical Analysis, AH-Receptor Bio-assay. Analyses of the major combustion products carbondioxide (CO2), carbon monoxide (CO), total hydrocarbons(THC), nitrogen monoxide (NO), and nitrogen oxides (NOx)have been described elsewhere (5) and results are given inTable S1. For the analysis of non- and semivolatile tracecompounds such as PAHs, nitro-PAHs, and PCDD/Fs, 5–7m3 of undiluted exhaust were collected through an all-glasssampling arrangement consisting of a sampling probe, acooler, a condensate separator, a glass fiber filter, and a two-stage adsorber unit (XAD-2). A water/isopropyl-alcoholcooling bath was used to keep the condensate below 4 °Cduring sampling, and exhaust temperatures remained below30 °C all the time. Mass flow proportional aliquots of theexhaust were taken during two consecutive runs (200 min)covering both, steady-state and transient operation modes.Prior to sampling, the employed glass apparatus wasintensively cleaned and heated to 450 °C.

Aliquots of a mixture of perdeuterated-PAHs containingnaphthalene, phenanthrene, pyrene, fluoranthene, chrysene,benz(a)anthracene, benzo(b)fluoranthene, benzo(k)fluoran-thene, benzo(a)pyrene, indeno(1,2,3-cd)pyrene, and D9-1-nitropyrene (CIL, Andover, MA) were added to the samples asquantification standards. A mixture of unlabeled nitro-PAHscontaining 1-nitronaphthalene, 2-nitronaphthalene, 3-nitro-phenanthrene, 9-nitrophenanthrene, 9-nitroanthracene, 3-ni-trofluoranthene, 1-nitropyrene, and 4-nitropyrene (Dr. Ehren-storfer AG, Augsburg, Germany) were mixed with an aliquot ofD9-1-nitropyrene and used as quantification standard. Separa-tion of PAHs and nitro-PAHs was obtained by gas chroma-tography (Fisons Instruments HRGC Mega 2, Rodano, Italy) ona capillary column (PS086, 20 m × 0.28 mm, 0.15 µm) anddetection and identification was achieved by high resolutionmass spectrometry (Thermo Finnigan MAT 95, Bremen,Germany) in electron ionization mode (GC/EI-HRMS).

Compounds inducing AHR-mediated gene expressionwere detected using the DR-CALUX assay, an in vitroreporter gene assay. This assay is based on stably trans-fected H4IIE rat hepatoma cells with an AHR-controlledluciferase reporter gene construct. Detailed descriptionsof the procedures for cell cultivation, cell exposure withexhaust samples, and quantification of the luciferaseactivity are given elsewhere (4).

Quality Assurance, Recovery Rates, Nitration Artifacts.To determine overall recovery rates and nitration artifacts,aliquots of 13C6-naphthalene, 13C6-phenanthrene, 13C3-pyrene,and 13C12-1,2,3,4,5,6-hexachlorodibenzodioxin (CIL) werespiked on quartz swab and placed in the condensate separatorprior to sampling. Mean overall recoveries of 26 ( 9%, 34 (5%, 59 ( 5%, and 57 ( 11% were obtained, respectively,covering sampling, workup, cleanup, and analysis. Asexpected, recoveries for naphthalene, the most volatile PAHwith a vapor pressure of 10 Pa at 25 °C, were low, but sufficientmaterial could be obtained to study nitration artifacts.Recoveries for non- and semivolatile AHs with vapor pres-sures <10-3 Pa at 25 °C, such as 4- to 6-ring PAHs, nitro-PAHs, and PCDD/Fs, were found to be comparable and in

3774 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 10, 2008

the expected range of ∼60%. Thus, large changes of thereported PAH patterns during sampling are not expected.

Nitration of PAHs can occur during combustion, in theDPF, in ambient air, but also during sampling and cleanup.The latter two would be considered as unwanted artifacts.Different measures can minimize such risks. Rapid dilutionof exhaust in a dilution tunnel is one option, but dilution inair represents an additional contamination source for tracecompounds such as PAHs, nitro-PAHs, and PCDD/Fs, anda soot-loaded dilution tunnel acts as a sink for such non-and semivolatile compounds. We therefore preferred torapidly cool the raw exhaust below 30 °C and to keep thecondensate <4 °C in the described sampling device (FigureS2). This validated filter/condenser method has become aninternational standard (EN-1948–1) and is widely used tosample non- and semivolatile trace compounds (19). How-ever, even with these precautions, nitration artifacts can notbe excluded. We therefore tested the extent of nitration duringsampling and cleanup with 13C-labeled naphthalene, phenan-threne, and pyrene. In case of naphthalene, neither 13C6-labeled 1-nitronaphthalene nor 13C6-labeled 2-nitronaph-talene was detected. On average, less than 1% of the given13C6-phenanthrene was transformed to 13C6-3-nitrophenan-threne, but no 13C6-9-nitrophenanthrene was found. About6 ( 5% and 2 ( 2% of the 13C3-pyrene was transformed to13C3-1-nitro-pyrene in tests without and with DPFs, respec-tively. No other 13C3-labeled nitro-pyrenes such as 4-nitro-pyrene were detected. We conclude that some nitrationartifacts occurred, but were small and did not bias ourfindings and conclusions.

Results and DiscussionA 6.11 L heavy duty diesel engine was operated with referencefuel (ref) or with iron- (Fe) or copper/iron-doped (Cu) fuelswith (F) or without DPFs. Two cordierite-based, monolithicDPFs of identical origin and dimensions were tested.Variations of engine load, consumed fuel, and exhaust gas

flow were small. On average, 25.7 ( 0.2 L of fuel wereconsumed during one pass of the ISO8178/4 cycle at a meanengine load of 55.3 ( 0.1 kW resulting in 532 ( 12 m3 of dryexhaust. Experimental setup and emission factors of majorcombustion products are given as supplementary data (TableS1) and are described elsewhere (5). Effects of both DPFs onPCDD/F emissions have been discussed before (5). The samesamples were also tested against a reporter gene assay,sensitive to aryl hydrocarbon receptor (AHR) agonists (4).Parts of the AHR data are also discussed herein and given asSupporting Information (Tables S2 and S3).

Three general findings are of relevance for the followingdiscussion of specific PAH and nitro-PAH data: (i) theapplication of both DPFs lowered THC emissions by about40% (Table S1) (5), (ii) both DPFs reduced total AHR agonistconcentrations by about 80–90% (Table S2) (4), and (iii) bothDPFs neither affected NO emissions, nor did they induce asecondary formation of NO2 (Table S1) (5). In fact, NO2/NOx

ratios decreased from 0.09 ( 0.01 to 0.03 ( 0.01, whenapplying DPFs, indicating that some of the engine-out NO2

is converted in both DPFs.Conversion of PAHs. Figure 1 includes the chemical

structures of the investigated PAHs. Pyrene (1) and fluo-ranthene (2) are not carcinogenic, but are both precursorsfor nitropyrenes and nitrofluoranthenes, of which severalisomers are classified as direct-acting mutagens. PAHs 3-8are classified as carcinogenic to humans by the World HealthOrganization (WHO) (20).

Figure 2 displays mean, minimum, and maximum PAHemission factors (µg/L) in the ISO8178/4 C1 cycle. Comparedto the reference fuel (ref), effects of the iron- (Fe) and thecopper/iron- (Cu) doped fuels were small, but all PAHs werereduced by 40–90% when applying DPFs (Table S2). At leasttwo physicochemical properties determine the extent of PAHconversion: volatility and reactivity. Less volatile or highlyreactive PAHs are retained in DPFs more efficiently thanothers. For example, pyrene (1) emissions of 72–89 and 14–20

FIGURE 1. Chemical structures of investigated PAHs and nitro-PAHs. Pyrene (1), fluoranthene (2), chrysene (3), benz(a)anthracene (4),benzo(b)fluoranthene (5), benzo(k)fluoranthene (6), benzo(a)pyrene (7), indeno(1,2,3-cd)pyrene (8), 1-nitronaphthalene (9), 2-nitronaphthalene(10), 3-nitrophenanthrene (11), 9-nitrophenanthrene (12), 9-nitroanthracene (13), 3-nitrofluoranthene (14), 1-nitropyrene (15), 4-nitropyrene (16).Compounds 3–8 are carcinogenic and compounds 14 and 15 are mutagenic according to WHO (12, 20).

VOL. 42, NO. 10, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 3775

µg/L were obtained without and with DPFs (F), resulting inconversion efficiencies of ∼80%. Fluoranthene (2) emissionswere reduced from 21-27 to 12-16 µg/L when applying DPFs,corresponding to efficiencies of only 40%. The 2-fold lowervapor pressure of pyrene (6.0 × 10-4 Pa at 25 °C) comparedto fluoranthene (1.2 × 10-3 Pa) explains the higher filtrationefficiency, but differences in reactivities may also beimportant.

At 100–500 °C relevant proportions of semivolatile 4-ringPAHs are present in the gas phase and therefore still canpenetrate DPFs. With increasing ring-number, PAHs areretained more efficiently in DPFs. Emission factors of thecarcinogenic 4- to 6-ring PAHs 3-8 are one or 2 orders ofmagnitude lower than those of pyrene and fluoranthene(Figure 2). Both DPFs clearly reduced emissions of thesecarcinogenic PAHs by 60–90%.

Benz(a)anthracene (4) and chrysene (3) are 4-ring PAHswith rather different vapor pressures of 2.8 × 10-5 and 8.5× 10-7 Pa at 25 °C. A conversion efficiency of ∼90% wasfound for benz(a)anthracene (4), even so its vapor pressureis about 30 times higher than the one of chrysene (3), whichwas retained in both DPFs with 70% efficiency. We assumethat benz(a)anthracene (4) is more reactive than chrysene(3) under the given conditions. These examples illustrate thecompeting processes of adsorption versus degradation thatoccur in DPFs.

The data also indicate that DPFs change PAH profiles ofdiesel exhaust. This should be considered, if ratios of certainPAHs are used as indicators for diesel traffic. For example,the fluoranthene/pyrene ratios increased from about 1:3 to1:1 and the benz(a)anthracene/chrysene ratios decreasedfrom 1:2 to 1:3 when applying DPFs. With increasing marketshares of DPF vehicles, we expect that PAH profiles of urbanair will change. Similar effects were observed, when intro-ducing the three-way catalyst technology (TWC) to gasoline-fueled vehicles. Alkylbenzenes are converted more efficientlythan benzene, which can even form de novo in TWCs undercertain conditions (21–24). Consequently, benzene/alkylbenzene ratios increased in exhaust gas and in urban airwith the implementation of the TWC technology (25, 26).

Stereoselective Formation of Certain Nitro-PAHs. Nitro-PAHs can form from precursor PAHs via gas-, liquid-, orsolid-phase chemistry. Different nitrating species such asgaseous NO2 or NO3 radicals or less volatile compounds likeHNO2 or HNO3 may be involved. Other radicals may activateprecursor PAHs in a first step, followed by addition of thenitrating species. In principle, DPFs can support all theseprocesses. At temperatures up to 500 °C, even larger PAHsmay volatilize, leading to gas-phase reactions not occurringin the atmosphere. Furthermore, prolonged exposure maysupport slower reactions. Typically, vapor pressures of nitro-PAHs are 2-3 orders of magnitude lower than those of relatedPAHs, and their reactivities are assumed to be higher.Consequently, conversion efficiencies are expected to behigher, unless nitro-PAHs do form de novo in DPFs. DPFscan support nitration reactions, but they can also catalyzenitro-PAH-degradation. It remains to be shown which of thecompeting processes dominates in each case.

Figure 1 displays the chemical structures of the investi-gated 2- to 4-ring nitro-PAHs (9-16). Included are 3-nitro-fluoranthene (14) and 1-nitropyrene (15), both direct-actingmutagens in the Ames Salmonella typhimurium TA98reversion assay (27). Nitropyrenes 15 and 16 representprecursors for mutagenic dinitropyrenes, like 1,3-, 1,6-, and1,8-dinitropyrenes, which are among the most potent direct-acting mutagens known (27). Figure 3 displays mean,minimum, and maximum emission factors (µg/L) of nitro-PAHs 9-16. In general, nitro-PAH emissions of the iron- andthe copper/iron-catalyzed DPFs are similar. Emissions ofthe most volatile 2-ring nitro-PAHs 9 and 10 are more than1 order of magnitude higher than those of 3- and 4-ring nitro-PAHs (11-16). Emissions of 1-nitronaphthalene (9) increasedfrom 15-21 to 19-30 µg/L without and with DPFs (F),respectively, those of 2-nitronaphthalene (10) rose from28-39 to 44-77 µg/L. When considering the low vaporpressures of nitronaphthalenes (3.2 × 10-2 Pa at 25 °C), onewould expect filtration efficiencies comparable to that offluoranthene (1.2 × 10-3 Pa). This is not the case. On the

FIGURE 2. Effects of DPFs and catalysts on PAH emissions (µg/L fuel). ref: reference fuel; Fe: iron-doped fuel (4.5 µg/g); Cu:copper/iron-doped fuel (9.0 ( 0.5/7.5 ( 0.7µg/g); without DPF(black); F: with DPF (grey). Mean, minimum, and maximumvalues are given. Compounds 3–8 are carcinogenic according toWHO (20).

3776 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 10, 2008

opposite, a substantial proportion of the emitted nitronaph-thalenes must have formed de novo in both DPFs. Anysampling artifact can be excluded because none of the given13C6-labeled naphthalene was converted to 13C6-nitronaph-thalenes.

Whereas about 33% of 3-nitrophenanthrene (11) wasretained in both DPFs, 9-nitrophenanthrene (12) was foundonly after DPFs. Again, a sampling artifact can be excludedin case of 9-nitrophenanthrene because <1% of the given13C6-labeled phenanthrene was converted to 13C6-3-nitro-phenanthrene only. Similarly, 9-nitroanthracene (13) was

not detected in engine-out exhaust, but released from bothDPFs. We conclude that a selective nitration of phenanthreneand anthracene must occur in both DPFs. Figure 4 displaysthe stereoselective nitration of phenanthrene and anthracene,which preferentially proceeds at hydrogen atoms in peri-positions. Both, phenanthrene and anthracene are annellatedPAHs with increased reactivities at position 9. The observednitration products are in accordance with a preferredsubstitution at these benzylic carbon atoms.

Emissions of 4-ring nitro-PAHs 14-16 are reduced by40–60% when using DPFs. This is of toxicological relevancebecause 3-nitrofluoranthene (14) and 1-nitropyrene (15) areboth mutagenic (27). Once more, nitration during samplingmust be small, about 2–6% of the 13C3-labeled pyrene wasconverted to 13C3-1-nitropyrene, whereas no 13C3-4-nitro-pyrene was detected.

DPF-supported nitration of PAHs was observed in fourout of eight cases. The resulting nitro-PAH profiles, withrelevant proportions of 1- and 2-nitronaphthalene, 9-nitro-phenanthrene, 9-nitroanthracene, 3-nitrofluoranthene, and1-nitropyrene still resemble those of combustion processes(28), albeit their relative proportions have changed. Noindications were found that DPFs enhance nitration reactionsvia the two-step mechanism, as observed in the atmosphere(17, 18, 27). Initial activation of a PAH via radical addition,followed by addition of the nitrating species, favors theformation of other nitro-PAH isomers. The occurrence of2-nitrofluoranthene and 2-nitropyrene is related to this typeof chemistry (17, 18, 27). Nitration of PAHs during samplingcan occur and precautions such as sample cooling or dilutionshould be foreseen to minimize such artifacts. However, theuse of 13C-labeled precursor PAHs is recommended to verifythe extent of nitration during sampling and cleanup.

DPF-Effects on Total Aryl Hydrocarbon Emissions. BothDPFs reduced emissions of all investigated PAHs includingthe carcinogenic ones as well as some critical precursor PAHs,which are transformed to mutagenic nitro-PAHs whenreleased to the atmosphere. Effects on individual nitro-PAHsvaried to a larger extent. Certain nitro-PAHs were retained,others were formed in DPFs. Nitration was found to bestereoselective, preferentially occurring at hydrogen atomsin peri-positions. PAH and nitro-PAH profiles of dieselexhaust changed considerably and a wider use of DPFtechnology is assumed to influence PAH and nitro-PAHpatterns at traffic-exposed sites. This should be consideredwhen using PAH and nitro-PAH data for future sourceapportionments (29–32).

FIGURE 3. Effects of DPFs and catalysts on nitro-PAH emissions(µg/L fuel). ref: reference fuel; Fe: iron-doped fuel (4.5 µg/g); Cu:copper/iron-doped fuel (9.0 ( 0.5/7.5 ( 0.7µg/g); without DPF(black); F: with DPF (grey). Mean, minimum, and maximum valuesare given. Compounds 14 and 15 are mutagenic according to WHO(12).

FIGURE 4. Scheme of the stereoselective nitration of phenan-threne and anthracene to 9-nitrophenanthrene (12) and 9-nitroan-thracene (13) as observed for both DPFs.

VOL. 42, NO. 10, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 3777

An alternative approach was used to assess effects of DPFson the large variety of aryl hydrocarbons (AHs) found in dieselexhaust. The same exhaust samples were tested against areporter gene assay, based on the AH-receptor, sensitive tomany aryl hydrocarbons (4). Figure 5 displays the total AHRagonist emissions and the contribution of six individual PAHs(Table S2). Both, the iron- and the copper-catalyzed DPFsreduced total AHR agonist emissions by 80–90%. The sixanalyzed PAHs (1, 4–8) for which relative potency values areavailable, accounted for <1% of the total AHR-mediatedresponse, indicating that the majority of the AHR agonistspresent in diesel exhaust is not identified. However, com-parable filtration efficiencies were noticed for total AHRagonists and for individual PAHs. We conclude that, ingeneral, aryl hydrocarbons are retained in both DPFs, andthe DPF-induced formation of certain nitro-PAHs does notoutrange these overall effects.

Environmental Perspectives. Both DPFs not only sup-ported the formation of certain nitro-PAHs, they alsoenhanced the degradation of others. The overall contributionsof these competing reactions must depend on the operatingconditions. Reported emission factors represent a mixed full-and partial-engine load operation, including load transitions,with exhaust temperatures of 100–500 °C. Such operationmodes are typical for construction machinery, but othertemperature- and precursor-levels have to be expected foron-road applications, which were not investigated. Othercatalytic DPFs with, e.g., noble metal coated substrates arein use. Such systems increase the NO2 content of the exhaustand may affect PAH and nitro-PAH pattern differently.

Despite these open questions, we conclude that bothtested DPFs efficiently filtered PAHs, including the carci-nogenic ones. Certain nitro-PAHs were formed in DPFs, butothers are reduced and nitration is not a general trend.Furthermore, both DPFs lowered total AHR agonist con-centrations by 80–90%, indicating that not only the analyzedPAHs and nitro-PAHs are retained in both DPFs, but alsomost of the large number of unidentified aryl hydrocarbonsmust be substantially removed. Based on these findings, bothDPF systems would be recommended for VERT approval.However, approval was denied for the copper-catalyzed DPFbecause of its increased PCDD/F formation potential (5). Itshould be noted that other criteria such as filtration efficiency(>95%) and penetration of nanoparticles (<5%) for a sizerange of 10–400 nm have to be fulfilled for VERT approvalboth for a new and an aged DPF (>2000 h of operation) (33).

In general, the given examples indicate that an assessmentof risks and benefits of a new technology should also address

its potentials toward the formation of toxic secondarypollutants. Bioassays such as the applied reporter gene assayor others may support the risk assessment. In case of DPFs,one should include an assessment of the nitration potentialand the PCDD/F formation potential.

The Swiss VERT procedures for DPF approval, which arenow accepted as a national standard (34), have been appliedto more than 20 different DPFs including coated DPFs andsystems relying on fuel-additives (fuel-borne catalyst) (33).The majority of the evaluated filters fulfilled all VERT criteria.This means that approved DPFs not only lower the numberof diesel particles by more than 95%, but also reduceemissions of individual carcinogenic PAHs such as ben-zo(a)pyrene, with low risks for a secondary formation of newpollutants such as nitro-PAHs and PCDD/Fs. In Switzerland,based on these findings, VERT-approved DPFs are nowconsidered the best available technology to reduce sootparticles and genotoxic compounds of diesel exhaust.Consequently, national occupational health authorities(SUVA) responsible for respiratory air quality standards atworkplaces, have decided together with the federal office forthe environment (FOEN) that DPF use is now mandatory forlarger construction engines (>18kW) (33, 35).

AcknowledgmentsThe fuel additives have been developed at the Institute ofPetroleum Processing (ITN), Krakow, Poland, during theDIEXFIL project (Development of Diesel Exhaust GasesFiltration Technology with Application of Fuel AdditivesEnabling Continuous Regeneration of Filters to Minimizethe Particulate Emission of City Busses, EU Project ref. Nr.CRAFT-1999-70982). We cordially thank L. Ziemianski, S.Oleksiak, and Z. Stepien from the ITN for the good col-laboration and for financial support.

Supporting Information AvailableFigures S1 and S2 display the ISO 8178/4 C1 cycle valid forconstruction engines, pre and post-DPF exhaust tempera-tures, and the sampling system used in accordance with EN1948–1 (19). Experimental conditions and emission factorsfor major combustion products are reported in Table S1.Tables S2 and S3 contain data on PAH, nitro-PAH, and AHR-agonist emission factors. This material is available free ofcharge via the Internet at http://pubs.acs.org.

Literature Cited(1) Armstrong, A.; Hutchinson, E.; Unwin, J.; Fletcher, T. Lung cancer

risk after exposure to polycyclic aromatic hydrocarbons: a reviewand meta analysis. Environ. Health Perspect. 2004, 112, 970–978.

(2) Brüske-Hohlfeld, I.; Möhner, M.; Ahrens, W.; Pohlabeln, H.;Heinrich, J.; Kreuzer, M.; Jöckel, K.-H.; Wichmann, H.-E. Lungcancer risk in male workers occupationally exposed to dieselmotor emissions in Germany. Am. J. Ind. Med. 1999, 36, 405–414.

(3) SUVA, Swiss National Accident Insurance Organization, Gren-zwerte am Arbeitsplatz. SUVA/1903 d Luzern, Switzerland, 2000.

(4) Wenger, D.; Gerecke, A. C.; Heeb, N. V.; Zennegg, M.; Kohler,M.; Naegeli, H.; Zenobi, R. Secondary effects of catalytic dieselparticulate filters: reduced aryl hydrocarbon receptor- mediatedactivity of the exhaust Environ. Sci. Technol. [Online early access].DOI: 10.1021/es071827x. Published Online: March 5, 2008.

(5) Heeb, N. V.; Zennegg, M.; Gujer, E.; Honegger, P.; Zeyer, K.;Gfeller, U.; Wichser, A.; Kohler, M.; Schmid, P.; Emmenegger,L.; Ulrich, A.; Wenger, D.; Petermann, J.-L.; Czerwinski, J.;Mosimann, T.; Kasper, M.; Mayer, A. Secondary effects ofcatalytic diesel particulate filters: Copper-induced de novoformation of PCDD/Fs. Environ. Sci. Technol. 2007, 41, 5789–5794.

(6) Schauer, J. J.; Kleeman, M. J.; Cass, G. R.; Simoneit, B. R. T.Measurement of emissions from air pollution sources. 2. C1

through C30 organic compounds from medium duty dieseltrucks. Environ. Sci. Technol. 1999, 33, 1578–1587.

FIGURE 5. Effects of DPFs and catalysts on total AH-receptoragonist emissions and contribution of six individual PAHs (1, 4–8) (µg TCCD-CEQ/L fuel). Agonist emissions are expressed as2,3,7,8-tetrachlorodibenzodioxin CALUX equivalents (TCDD-CEQ).ref: reference fuel; Fe: iron-doped fuel (4.5 µg/g); Cu: copper/iron-doped fuel (9.0 ( 0.5/7.5 ( 0.7µg/g); without DPF (black); F:with DPF (grey). Mean, minimum, and maximum values aregiven.

3778 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 10, 2008

(7) Marr, L. C.; Kirchstetter, T. W.; Harley, R. A.; Miguel, A. H.;Hering, S. V.; Hammond, S. K. Characterization of polycyclicaromatic hydrocarbons in motor vehicle fuels and exhaustemissions. Environ. Sci. Technol. 1999, 33, 3091–3099.

(8) Schauer, J. J.; Kleeman, M. J.; Cass, G. R.; Simoneit, B. R. T.Measurement of emissions from air pollution sources. 5. C1

through C32 organic compounds from gasoline-powered motorvehicles. Environ. Sci. Technol. 2002, 36, 1169–1180.

(9) Zielinska, B.; Sagebiel, J.; Arnott, W. P.; Rogers, C. F.; Kelly, K. E.;Wagner, D. A.; Lighty, J. S.; Sarofim, A. F.; Palmer, G. Phase andsize distribution of polycyclic aromatic hydrocarbons in dieseland gasoline vehicles. Environ. Sci. Technol. 2004, 38, 2557–2567.

(10) Miguel, A. H.; Kirchstetter, E. W.; Harley, R. A.; Hering, S. V.On-road emissions of particulate polycyclic aromatic hydro-carbons and black carbon from gasoline and diesel vehicles.Environ. Sci. Technol. 1998, 32, 450–455.

(11) Staehelin, J.; Keller, C.; Stahel, W.; Schläpfer, K.; Wunderli, S.Emission factors from road traffic from a tunnel study (Gubristtunnel, Switzerland). Part III: Results of organic compounds,SO2 and specification of organic exhaust emission. Atmos.Environ. 1998, 32, 999–1009.

(12) Environmental Health Criteria 229, Selected Nitro- And Nitro-Oxy-Polycyclic Aromatic Hydrocarbons; World Health Organiza-tion: Geneva, Switzerland, 2003; http://www.who.int/ipcs/publications/ehc/.

(13) Durant, J.L.Jr.; Lafleur, A. L.; Penman, B. W.; Crespi, C. L. Humancell mutagenicity of oxygenated, nitrated, and unsubstitutedpolycyclic aromatic hydrocarbons associated with urban aero-sols. Mutat. Res. 1996, 371, 123–157.

(14) Hannigan, M. P.; Cass, G. R.; Penman, B. W.; Crespi, C. L.; Lafleur,A. L.; Busby, W.F.Jr.; Thilly, W. G. Human cell mutagens in LosAngeles air. Environ. Sci. Technol. 1997, 31, 438–447.

(15) Hannigan, M. P.; Cass, G. R.; Penman, B. W.; Crespi, C. L.; Lafleur,A. L.; Busby, W. F. Jr.; Thilly, W. G.; Simoneit, B. R. Bioassay-directed chemical analysis of Los Angeles airborne particulatematter using a human cell mutagenicity assay. Environ. Sci.Technol. 1998, 32, 3502–3514.

(16) Ramdahl, T.; Zielinska, B.; Arey, J.; Atkinson, R.; Winer, A. M.;Pitt, J.N. Jr. Ubiquitous occurrence of 2-nitrofluoranthene and2-nitropyrene in air. Nature 1986, 321, 425–427.

(17) Finlayson-Pitts, B. J.; Pitts, J. N. Tropospheric air pollution: ozone,airborne toxics, polycyclic aromatic hydrocarbons, and particles.Science. 1997, 276, 1045–1052.

(18) Atkinson, R.; Arey, J. Atmospheric chemistry of gas-phasepolycyclic aromatic hydrocarbons: formation of atmosphericmutagens. Environ. Health Perspect. 1994, 102, 117–126.

(19) EN 1948–1, Stationary Source Emissionssdetermination of theMass Concentration of PCDDs/PCDFs and Dioxin-Like PCBssPart 1: Sampling of PCDDs/PCDFs; CEN, European Committeefor Standardization: Brussels, Belgium, 1996; http://www.cen.eu/cenorm/.

(20) Environmental Health Criteria 202, Selected Non-HeterocyclicPolycyclic Aromatic Hydrocarbons; World Health Organiza-tion: Geneva, Switzerland, 1998; http://www.who.int/ipcs/publications/ehc/.

(21) Heeb, N. V.; Forss, A.-M.; Bach, C.; Mattrel, P. Velocity-dependentemission factors of benzene, toluene and C2-benzenes of apassenger car equipped with and without a regulated 3-waycatalyst. Atmos. Environ. 2000, 34, 1123–1137.

(22) Heeb, N. V.; Forss, A.-M.; Weilenmann, M. Pre- and post-catalyst-, fuel-, velocity- and acceleration-dependent benzeneemission data of gasoline-driven EURO-2 passenger cars andlight duty vehicles. Atmos. Environ. 2002, 36, 4745–4756.

(23) Bruehlmann, S.; Forss, A.-M.; Steffen, D.; Heeb, N. V. Benzene:a secondary pollutant formed in the three-way catalyst. Environ.Sci. Technol. 2005, 39, 331–338.

(24) Bruehlmann, S.; Novak, P.; Lienemann, P.; Trottman, M.; Gfeller,U.; Zwicky, C. N.; Bommer, B.; Huber, H.; Wolfensberger, M.;Heeb, N. V. Three-way-catalyst induced benzene formation: Aprecursor study. Appl. Catal. B: Environ. 2007, 70, 276–283.

(25) Heeb, N. V.; Forss, A.-M.; Bach, C.; Reimann, S.; Herzog, A.;Jackle, H. W. A comparison of benzene, toluene and C2-benzenesmixing ratios in automotive exhaust and in the suburbanatmosphere during the introduction of catalytic convertertechnology to the Swiss car fleet. Atmos. Environ. 2000, 34, 3103–3116.

(26) Harley, R. A.; Hooper, D. S.; Kean, A. J.; Kirchstetter, T. W.; Hesson,J. M.; Balberan, N. T.; Stevenson, E. D.; Kendall, G. R. Effects ofreformulated gasoline and motor vehicle fleet turnover onemissions and ambient concentrations of benzene. Environ.Sci. Technol. 2006, 40, 5084–5088.

(27) Finlayson-Pitts, B. J. Jr. Chemistry and mutagenic activity ofairborne polycyclic aromatic hydrocarbons and their derivatives.In Atmospheric Chemistry: Fundamentals and ExperimentalTechniques; John Wiley & Sons: New York, 1986; pp 870–958.

(28) Bezabeh, D. Z.; Bamford, H. A.; Schantz, M. M.; Wise, S. A.Determination of nitrated polycyclic aromatic hydrocarbons indiesel particulate-related standard reference materials by usinggas chromatography/mass spectrometry with negative ionchemical ionization. Anal. Bioanal. Chem. 2003, 375, 381–388.

(29) Harrison, R. M.; Smith, D. J. T.; Luhana, L. Source apportionmentof atmospheric polycyclic aromatic hydrocarbons collected froman urban location in Birmingham, UK. Environ. Sci. Technol.1996, 30, 825–832.

(30) Schauer, J. J.; Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.;Cass, G. R. Source apportionment of airborne particulate matterusing organic compounds as tracers. Atmos. Environ. 1996, 30,3837–3855.

(31) Larsen, R. K.; Baker, J. E. Source apportionment of polycyclicaromatic hydrocarbons in the urban atmosphere: A comparisonof three methods. Environ. Sci. Technol. 2003, 37, 1873–1881.

(32) Robinson, A. L.; Subramanian, R.; Donahue, N. M.; Bernardo-Bricker, A.; Rogge, W. F. Source apportionment of molecularmarkers and organic aerosol-1. Polycyclic aromatic hydrocar-bons and methodology for data visualization. Environ. Sci.Technol. 2006, 40, 7803–7810.

(33) FOEN/Suva Filter List, Tested and Approved Particle Filter Systemsfor Retrofitting Diesel EnginessStatus December 2006; FOEN,Federal Office for the Environment: Bern, Switzerland, 2006;http://www.bafu.admin.ch/publikationen/.

(34) SNR 277205 Testing Particle Filter Systems for Internal Combus-tion Engines; SNV, Swiss Normative Organization: Winterthur,Switzerland, 2007; http://www.snv.ch.

(35) Guideline air Pollution Control at Construction Sites, Con-struction Guideline Air; FOEN, Federal Office for the Environ-ment: Bern, Switzerland, 2004; http://www.bafu.admin.ch/publikationen/.

ES7026949

VOL. 42, NO. 10, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 3779