Metal Particle Emissions in the Exhaust Stream of Diesel Engines: An Electron Microscope Study

7
Metal Particle Emissions in the Exhaust Stream of Diesel Engines: An Electron Microscope Study Anthi Liati,* ,Daniel Schreiber, Panayotis Dimopoulos Eggenschwiler, and Yadira Arroyo Rojas Dasilva Empa Material Science and Technology, Laboratory of Internal Combustion Engines, Ueberlandstrasse 129, CH-8600, Dü bendorf, Switzerland Empa, Electron Microscopy Center, Ueberlandstrasse 129, CH-8600, Dü bendorf, Switzerland ABSTRACT: Scanning electron microscopy and transmission electron micros- copy were applied to investigate the morphology, mode of occurrence and chemical composition of metal particles (diesel ash) in the exhaust stream of a small truck outtted with a typical after-treatment system (a diesel oxidation catalyst (DOC) and a downstream diesel particulate lter (DPF)). Ash consists of Ca-Zn-P-Mg-S-Na-Al-K-phases (lube-oil related), Fe, Cr, Ni, Sn, Pb, Sn (engine wear), and Pd (DOC coating). Soot agglomerates of variable sizes (<0.55 μm) are abundant upstream of the DPF and are ash-free or contain notably little attached ash. Post-DPF soot agglomerates are very few, typically large (>15 μm, exceptionally 13 μm), rarely <0.5 μm, and contain abundant ash carried mostly from inside the DPF. The ash that reaches the atmosphere also occurs as separate aggregates ca. 0.22 μm in size consisting of sintered primary phases, ca. 20400 nm large. Insoluble particles of these sizes may harm the respiratory and cardiovascular systems. The DPF probably promotes breakout of large soot agglomerates (mostly ash- bearing) by favoring sintering. Noble metals detached from the DOC coating may reach the ambient air. Finally, very few agglomerates of Feoxide nanoparticles form newly from engine wear and escape into the atmosphere. 1. INTRODUCTION Diesel particulate matter (PM) is a source of pollution with potential impacts on the environment, including human health. During recent decades, continuous concern has been raised with respect to PM emitted by diesel engines, as this material has detrimental eects on the environment. 1 The World Health Organization (WHO) reported in 2012 that diesel engine exhaust is classied as carcinogenic to humans. 2 Diesel PM consists predominantly of soot but also includes a minor fraction of noncarbonaceous, nonvolatile, inorganic components referred to as metal PM or ash PM. Ash PM is of submicrometer size and is thus inhalable. It includes (a) species formed chemically from combustion of lubricating oil metal additives and, to a lesser degree, trace metals in the fuel. Depending on the saturation levels, the nonvolatile species (metals and salts) can either form separate solid particles or deposit on soot. 3,4 Ash species consist of oxides, sulfate and phosphate compounds of mainly Ca, Mg, and Zn. 57 (b) Metals derived from engine wear and corrosion (e.g., Fe, Cr, Ni, Cu, Sn). To reduce diesel PM emissions, stringent regulations were introduced that led to considerable improvements in both optimization of engine performance and sophisticated after- treatment systems. An important step toward drastic diesel PM reduction was the introduction of diesel particulate lters (DPF), which are able to eliminate diesel PM by up to ca. 99%. 810 While soot accumulated in the DPF can be oxidized and removed during the so-called passive and active regeneration processes, 11 ash PM cannot be chemically eliminated and thus irreversibly and gradually plugs the DPF. It also forms deposits on other components of the engine and after-treatment system. Ash sintering and development of ash layers on the DPF walls occur over a time scale of thousands of hours of lter operation. 12 As a result of continuous ash accumulation, the eective DPF volume is reduced, the pressure drop sensitivity of the DPF may be altered, 13 the lifetime of the DPF is decreased, fuel consumption is aected, and a minor portion of the ash PM may escape into the atmosphere. 6,14,15 Moreover, ash metals such as P and S have been connected to chemical degradation of catalysts. 16 Although fewer ash emissions can be expected by lowering the concentration of ash-producing elements in the lubricating oil, such a reduction would not be benecial for the protection of engine components. Nevertheless, there are recently promising developments of low-ash lube oils, especially for P- and Zn-bearing compounds, 17 but these oils still must be tested. In contrast to soot, which has received much attention among researchers, ash has been inadequately studied 6,13,15,18,19 despite its high importance with respect to its possible connection with adverse health eects 3,20 and deterioration of engine eciency. 6 Measurements of diesel PM in the exhaust stream (e.g., particle mass, number or size distribution) are commonly Received: July 15, 2013 Revised: November 19, 2013 Accepted: November 25, 2013 Published: November 25, 2013 Article pubs.acs.org/est © 2013 American Chemical Society 14495 dx.doi.org/10.1021/es403121y | Environ. Sci. Technol. 2013, 47, 1449514501

Transcript of Metal Particle Emissions in the Exhaust Stream of Diesel Engines: An Electron Microscope Study

Page 1: Metal Particle Emissions in the Exhaust Stream of Diesel Engines: An Electron Microscope Study

Metal Particle Emissions in the Exhaust Stream of Diesel Engines: AnElectron Microscope StudyAnthi Liati,*,† Daniel Schreiber,† Panayotis Dimopoulos Eggenschwiler,† and Yadira Arroyo Rojas Dasilva‡

†Empa Material Science and Technology, Laboratory of Internal Combustion Engines, Ueberlandstrasse 129, CH-8600, Dubendorf,Switzerland‡Empa, Electron Microscopy Center, Ueberlandstrasse 129, CH-8600, Dubendorf, Switzerland

ABSTRACT: Scanning electron microscopy and transmission electron micros-copy were applied to investigate the morphology, mode of occurrence andchemical composition of metal particles (diesel ash) in the exhaust stream of asmall truck outfitted with a typical after-treatment system (a diesel oxidationcatalyst (DOC) and a downstream diesel particulate filter (DPF)). Ash consists ofCa-Zn-P-Mg-S-Na-Al-K-phases (lube-oil related), Fe, Cr, Ni, Sn, Pb, Sn (enginewear), and Pd (DOC coating). Soot agglomerates of variable sizes (<0.5−5 μm)are abundant upstream of the DPF and are ash-free or contain notably littleattached ash. Post-DPF soot agglomerates are very few, typically large (>1−5 μm,exceptionally 13 μm), rarely <0.5 μm, and contain abundant ash carried mostlyfrom inside the DPF. The ash that reaches the atmosphere also occurs as separateaggregates ca. 0.2−2 μm in size consisting of sintered primary phases, ca. 20−400 nm large. Insoluble particles of these sizes mayharm the respiratory and cardiovascular systems. The DPF probably promotes breakout of large soot agglomerates (mostly ash-bearing) by favoring sintering. Noble metals detached from the DOC coating may reach the ambient air. Finally, very fewagglomerates of Fe−oxide nanoparticles form newly from engine wear and escape into the atmosphere.

1. INTRODUCTION

Diesel particulate matter (PM) is a source of pollution withpotential impacts on the environment, including human health.During recent decades, continuous concern has been raisedwith respect to PM emitted by diesel engines, as this materialhas detrimental effects on the environment.1 The World HealthOrganization (WHO) reported in 2012 that diesel engineexhaust is classified as carcinogenic to humans.2

Diesel PM consists predominantly of soot but also includes aminor fraction of noncarbonaceous, nonvolatile, inorganiccomponents referred to as metal PM or ash PM. Ash PM isof submicrometer size and is thus inhalable. It includes (a)species formed chemically from combustion of lubricating oilmetal additives and, to a lesser degree, trace metals in the fuel.Depending on the saturation levels, the nonvolatile species(metals and salts) can either form separate solid particles ordeposit on soot.3,4 Ash species consist of oxides, sulfate andphosphate compounds of mainly Ca, Mg, and Zn.5−7 (b)Metals derived from engine wear and corrosion (e.g., Fe, Cr,Ni, Cu, Sn).To reduce diesel PM emissions, stringent regulations were

introduced that led to considerable improvements in bothoptimization of engine performance and sophisticated after-treatment systems. An important step toward drastic diesel PMreduction was the introduction of diesel particulate filters(DPF), which are able to eliminate diesel PM by up to ca.99%.8−10 While soot accumulated in the DPF can be oxidizedand removed during the so-called passive and activeregeneration processes,11 ash PM cannot be chemically

eliminated and thus irreversibly and gradually plugs the DPF.It also forms deposits on other components of the engine andafter-treatment system. Ash sintering and development of ashlayers on the DPF walls occur over a time scale of thousands ofhours of filter operation.12 As a result of continuous ashaccumulation, the effective DPF volume is reduced, thepressure drop sensitivity of the DPF may be altered,13 thelifetime of the DPF is decreased, fuel consumption is affected,and a minor portion of the ash PM may escape into theatmosphere.6,14,15 Moreover, ash metals such as P and S havebeen connected to chemical degradation of catalysts.16

Although fewer ash emissions can be expected by loweringthe concentration of ash-producing elements in the lubricatingoil, such a reduction would not be beneficial for the protectionof engine components. Nevertheless, there are recentlypromising developments of low-ash lube oils, especially for P-and Zn-bearing compounds,17 but these oils still must be tested.In contrast to soot, which has received much attention

among researchers, ash has been inadequately studied6,13,15,18,19

despite its high importance with respect to its possibleconnection with adverse health effects3,20 and deterioration ofengine efficiency.6

Measurements of diesel PM in the exhaust stream (e.g.,particle mass, number or size distribution) are commonly

Received: July 15, 2013Revised: November 19, 2013Accepted: November 25, 2013Published: November 25, 2013

Article

pubs.acs.org/est

© 2013 American Chemical Society 14495 dx.doi.org/10.1021/es403121y | Environ. Sci. Technol. 2013, 47, 14495−14501

Page 2: Metal Particle Emissions in the Exhaust Stream of Diesel Engines: An Electron Microscope Study

carried out with methods using particle counters (scanningmobility particle sizers (SMPS)) or electrical low-pressureimpactors (ELPI).10,21 Such techniques had a significantcontribution to abatement measures against diesel PMpollution as well as to testing the efficiency of DPFs. Thesedevices are able to detect and measure agglomerates in theexhaust stream with rigorous statistics. However, they are notable to discriminate between soot and ash agglomerates andcannot distinguish any individual constituents of the agglom-erates. A recently developed device, the soot particle aerosolmass spectrometer (SP-AMS), can provide information on theelemental composition of engine-out emissions (Ca, Zn, Mg,S).22

Within the framework of the present study, we applyscanning electron microscope (SEM) and transmission electronmicroscope (TEM) techniques to investigate the morphology,mode of occurrence, chemical composition, and variations inthe relative amount of ash PM at different sites of the exhauststream of a small truck on a chassis dynamometer. Thesemicroscopy methods apply high magnifications and aim atrevealing important details of the materials down to themicrometer and nanometer scales. Consequently, the numberof examined samples cannot be very high. Studies of ash insidethe DPF are not included in the present work because this wasthe topic of previous research.14,23 The aim of the present workis to highlight the differences in the characteristics and relativeamount of metal-bearing particles (ash) in the exhaust streambefore entering into and after exiting the DPF, with possibleimpacts on metal particles that may escape in the ambient air.The study of metal particles in this paper is bound mainly toatmospheric pollution but is also expected to aid in minimizingthe detrimental effects of metal PM on exhaust after-treatmentsystems.

2. EXPERIMENTAL SECTION2.1. Experimental Setup and Sampling Procedure.

The vehicle used for the experiments was an Iveco Daily (2.3-L4-cylinder (F1A) common rail diesel engine with turbocharger;MY 2003; Euro IV emission limits). Sampling was performedduring steady state operation at 2000 rpm engine speed and 13kW output. Commercial fuel (S content <10 ppm) andcommercial lubricating oil were used. The exhaust after-treatment system consisted of a diesel oxidation catalyst(DOC) and a downstream DPF (Figure 1A). Note that theDOC promotes passive regeneration in the DPF via NO2-assisted soot oxidation. The assembly of serial productioncomponents was positioned following the turbine exit,approximately 1.5 m further downstream, to allow easyaccessibility.The DPF used was new and is one of the most common

systems currently on the road with a total volume of 6.3 L. It isconstructed of SiC (without catalytic coating), has a so-called“honeycomb” structure and consists of 25 square segments.The segments consist of numerous square channels, 1.5 mm insize, separated by ca. 0.35 mm thick porous walls. The oppositeends of each channel are alternately plugged such that the PM-bearing exhaust stream can exit only after passing through theporous walls to the adjacent outlet channels and thus reach theambient air nearly free of PM (Figure 1B; see ref 23 foradditional details on the DPF).Before sampling, the DPF was degreened for several hours at

exhaust temperatures of ∼400 °C, subsequently loaded withsoot and experienced a limited number of active regenerations

during which temperatures reached 800−900 °C (downstreamof the DPF). Sampling was performed using a pump at 0.5L/min under normal operating conditions at nearly half the timebetween two active regeneration phases, at which point thebuild-up of a soot cake can be assumed with certainty. The PMwas collected directly on 3 mm copper-supported carbon-coated lacy TEM grids using an electrostatic particle samplerconnected with a short sampling line at different sites of theexhaust pipe. The sampling site near the TEM grids was heatedto prevent condensation.Sampling was performed as follows (Figure 1A): (a)

immediately upstream of the DOC (sampling time of 1 min),(b) immediately downstream of the DOC (sampling time of 1min), and (c) immediately downstream of the DPF (samplingtime of 2−5 min). Upstream and downstream of the DOC, thesampled exhaust gas was diluted at a 1:10 ratio with clean air,whereas downstream of the DPF, it was sampled undiluted dueto the low number of soot agglomerates that exit the DPF. Thetemperature of the exhaust gas upstream of the DOC,downstream of the DOC and downstream of the DPF wasapproximately 275, 295, and 255 °C, respectively. Thesetemperatures are normal for the chosen engine operating point.

2.2. Analytical Techniques. The PM-loaded TEM gridswere used for both SEM and TEM studies. A Hitachi S-4800instrument combined with an energy dispersive system (EDX)for qualitative chemical analysis was used for SEM imaging. AJEOL 2200FS microscope with an Omega filter, a Schottky fieldemission gun at 200 kV and a point-to-point resolution of 0.23nm equipped with an EDX detector for elemental analysis wasused for TEM imaging. Images were collected both in bright-field (BF) and dark-field (DF) STEM mode as well as in high-resolution (HRTEM) mode.

2.3. Terminology. Before proceeding to the results of thispaper, it is important to clarify the following terms used in thiswork:

1. The terms “ash/metal components”, “ash PM”, and“metal PM” applied in the literature are used here in thesame sense.

2. The term “agglomerate” is generally used to describe aconglomerate of particles held together loosely, whereasthe term “aggregate” is applied for particles that arefirmly attached to each other.

Figure 1. (A) Sampling assembly used for the experiments; (B)Schematic illustration of neighboring DPF channels plugged atopposite ends; black spots represent diesel PM (not to scale).

Environmental Science & Technology Article

dx.doi.org/10.1021/es403121y | Environ. Sci. Technol. 2013, 47, 14495−1450114496

Page 3: Metal Particle Emissions in the Exhaust Stream of Diesel Engines: An Electron Microscope Study

3. The terms “agglomerate” and “aggregate” as used in thepresent paper are equivalent to the term “particle”reported in the majority of articles dealing withmeasurements of the size/mass distribution of sootagglomerates and ash aggregates in the exhaust stream bymeans of, for example, scanning mobility particle sizers(SMPS) or electrical low-pressure impactors (ELPI).10,21

The term “particle” as used in the present paper refers tothe individual primary particle constituents of theagglomerates/aggregates. The size of agglomerates/aggregates and their primary particle constituents, asreported in this paper, refer to those physically measuredon the TEM grids.

3. RESULTS OF SEM IMAGING3.1. Pre-DPF Samples. Pre-DPF samples collected up-

stream and downstream of the DOC show no differences in theash nature, relative amount and morphological characteristicsand are further considered as a single group. The studiedsamples contain numerous pieces of ash. In one 80 × 80 μmlarge square of the TEM grid, at least 5−6 pieces of ash ≥1 μmin size and several smaller pieces were found in addition toabundant soot agglomerates (Figure 2). The size of the ashaggregates/fragments ranges generally between 0.2 and 2 μm(rarely up to 3 and 3.5 μm), and their shape is irregular andusually rounded.

The EDX analyses of several ash aggregates and/orfragments yielded the following elements: Ca, Zn, P, Mg, S,Na, Al, K (attributed to additives in the lube oil), Fe, Cr, Ni, Sn,Pb, Al (engine wear and/or corrosion of engine and after-treatment material).Soot agglomerates are abundant, mostly a few hundreds of

nanometers in size, while very large ones (ca. 1.5−3 μm and,more rarely, even larger) are common. The SEM-EDX spectraof the soot agglomerates yielded no ash elements implying thatash is not attached to soot at this site of the exhaust stream orthat it occurs in quantities that are not detectable within thesensitivity of the SEM-EDX system.

3.2. Post-DPF Samples. The samples downstream of theDPF contain notably low amounts of diesel PM (Figure 2).One to two ash pieces (ca. 1−2 μm) and a few smaller ones (afew hundreds of nm) as well as soot agglomerates usually ca.1−5 μm in size and rarely larger were found in numerous of the80 × 80 μm large squares of the TEM grids (Figure 2). In onecase, a 13 × 3 μm soot agglomerate was identified (Figure 2F).Several TEM grid squares were completely free of diesel PM.High magnification images (180−300 kx) of isolated ash

aggregates reveal that they consist of several individual phases(ca. 20−400 nm) with rounded outlines sintered together(Figure 3A−C). In addition to separate ash aggregates, ash wasalso detected via EDX analyses as attached to soot

Figure 2. SEM images of samples with soot and ash PM upstream (A−C) and downstream of the DPF (D−F). The post-DPF samples show adramatic decrease in ash and soot PM compared to the pre-DPF ones. Large and very large soot agglomerates (E, F) are common downstream of theDPF. Rectangles mark sites with ash pieces. Circles in (D) mark sites with soot.

Environmental Science & Technology Article

dx.doi.org/10.1021/es403121y | Environ. Sci. Technol. 2013, 47, 14495−1450114497

Page 4: Metal Particle Emissions in the Exhaust Stream of Diesel Engines: An Electron Microscope Study

agglomerates. It is reminded that soot agglomerates upstreamof the DPF were free of ash or contained very low ash amounts.Interestingly, small soot agglomerates (<0.5−1 μm), which

are abundant upstream of the DPF, are very rarely found inpost-DPF samples. An important issue of interest raisedtherefore here is that the soot agglomerates that escapethrough the DPF to the ambient air are rather large in size (>1μm and even up to 13 μm), whereas very small ones (<ca. 0.5μm) are scarce. Even more importantly, these large sootagglomerates carry ash components, which may be liberated inthe atmosphere either mechanically or via oxidation of the sootcarrier. It would be of interest to follow up on the mechanism,by which large soot agglomerates are pushed through the DPFpores rather than small ones. One hypothesis could be thatduring normal operation, soot agglomerates (possibly carryingash from inside the DPF) intrude in the filter wall and aretrapped in the pores, where they grow larger from continuouslyincoming soot and ultimately block a portion of the wall pores.Blocking of the pores causes a pressure difference between theinlet and outlet channels, which leads to an increasing force onthe soot agglomerates from the inflowing stream. Thus, aportion of the agglomerates may finally escape to the ambientair. Although the DPF is quite promising and has provedefficient for retaining diesel PM, it cannot prevent (or may evenpromote) large soot agglomerates (>1 μm) from escaping intothe ambient air. Quantification of the ash escaping filtration isbeyond the purpose of the present study.

4. RESULTS OF TEM IMAGING

4.1. Pre-DPF Samples. While SEM imaging provides adetailed analysis of topographic structures, that is, surfacefeatures, (HR)TEM can provide detailed data on the innerstructure of material. Soot agglomerates on the carbon-coatedfilm of the TEM grids are visible at a higher focal planecompared with that of the flat carbon film. For optimumimaging, soot agglomerates trapped on holes of the lacy film arechosen.The TEM imaging of pre-DPF samples allows the

recognition of several ash aggregates and further reveals aclear distinction of individual phases within the aggregates,which are shown to consist chemically of either the same ordifferent substances. Figure 3D shows such an aggregateconsisting of distinguishable Ca-bearing (lighter gray) and Sn-bearing (darker) phases. It is noted that Figures 3D−F arebright-field STEM images, which have the potential to displaydifferent contrasts for the various chemical elements accordingto their atomic number; the lighter elements exhibit brightercontrast than heavier elements (in Figure 3D, the lighter Ca-phase appears brighter than the heavier Sn-phase). Theindividual ash phases observed in the aggregates display a sizerange of 45−160 nm. The most frequently found element issulfur either combined with Ca (in the form of CaSO4) or not,in agreement with previous findings.6 Additionally, Fe-, Cr-,Ni-, K-, Na-, Cl-, Sn-, Sb-, Al-, Si-, and Mg-bearing compoundswere sporadically detected.The soot agglomerates are abundantly dispersed over the

sampling surface of the pre-DPF samples (Figure 3D, G). Inmost cases, either no ash elements or very low amounts of S,

Figure 3. (A−C): SEM images of ash aggregates downstream of the DPF. Primary ash phases with mostly round outlines can be distinguished in (B)and (C). (D−I): BF-STEM images; (D, G) upstream of the DPF showing thin soot agglomerates with no or very little ash as well as a multiphaseash aggregate in (D); (E, F, H, I): downstream of the DPF showing multiphase ash aggregates (E, F) and ash-bearing thick soot agglomerates (H, I).The rectangles mark sites of EDX analyses. Elements in parentheses occur in minor amounts.

Environmental Science & Technology Article

dx.doi.org/10.1021/es403121y | Environ. Sci. Technol. 2013, 47, 14495−1450114498

Page 5: Metal Particle Emissions in the Exhaust Stream of Diesel Engines: An Electron Microscope Study

Cr, Mn, and Fe were detected on the soot agglomerates (Figure3G).4.2. Post-DPF Samples. The post-DPF samples exhibit

notably low amounts of diesel PM because the highest fractionis retained by the DPF. Ash occurs as separate aggregatesconsisting of primary phases with the same or differentchemical composition. Examples of ash aggregates consistingof Ca-bearing phases as well as S, Zn, Mg, and minor amountsof Cr, Fe, Mn, and Ni are shown in Figure 3E, F. Primary ashparticles 40−400 nm in size are distinguished in the aggregates.The most frequently observed ash element is sulfur. Moreover,Zn-, Mg-, Fe-, Mn-, Cr-, and Ni-bearing compounds are alsoidentified.The post-DPF samples show abundant ash attached on soot

(Figure 3H, I). EDX analyses of several soot agglomeratesclearly reveal the presence of such ash elements as Ca, P, Zn,Mg, S, sometimes also Al and K (related to lubricating oiladditives) as well as Fe, Mn, Cr, Ni, and occasionally Ti, whichare connected to engine wear (see also above; Section 3.2.).To investigate possible variations in the relative amount of

ash elements, element mapping was carried out on large sootagglomerates. Element mapping on a region of a large sootagglomerate that escaped filtration (Figure 4) reveals that themost widespread elements are S, Mg, Mn (homogeneouslydistributed) and partly Si (locally enriched). Zinc is lessabundant, and Fe and Ca are locally enriched.

It is worth noting that a clear Pd peak was identified on twoof the analyzed soot agglomerates. This observation isexplained by assuming detachment of Pd particles from thecoating material of the DOC, subsequent transport by theexhaust stream into the DPF downstream and further transportto the ambient air. Detachment of noble metals from the DOC

(Pt with and without its substrate) and deposition in the DPFhas been reported previously.14,24

In addition to entire soot agglomerates, primary sootparticles also were analyzed by EDX to search for ash metalspossibly present inside such particles.25 The primary sootparticles selected for analysis were chosen based on STEMimages at the edge (thinnest part) of the agglomerates becausein the BF STEM images, ash elements will appear darker due totheir higher atomic weight compared with that of carbon. Suchan example is illustrated in Figure 5A. The EDX analyses clearly

yielded S, P, and sometimes minor Fe, Cr, and Mn peaks.However, no clear metal phases could be distinguished on theHRTEM images of the soot particles (Figure 5B) implying thatthe metal-bearing particles are very small and thus difficult todistinguish with certainty at the actual resolution of the image.Comparing the pre-DPF and post-DPF samples, it becomes

evident that in contrast to the soot agglomerates analyzedupstream of the DPF in which metallic components were notidentified or were present in minor amounts, the ones thatescape in the ambient air carry both ash (s.s.) and wear metals.This observation can be explained as follows: Ash PM upstreamof the DPF is rather scarcely dispersed within the exhauststream and is thus more unlikely to attach onto soot, unless thisprocess occurs already in the cylinder. In contrast, ash thataccumulates inside the DPF during engine operation can beeasily attached onto the continuously incoming soot agglom-erates, which could potentially carry it through the inlet/outletchannel walls and expose it to the atmosphere (see above;Section 3.2.) Thus, the DPF promotes the formation of largesoot agglomerates, which are likely to also carry ash into theambient air if they escape filtration.

4.2.1. Particular Nanosized Iron Oxide Agglomerates. Oneof the samples downstream of the DPF contains twoagglomerates (2.1 × 2.4 μm and 1 × 0.6 μm large) consistingof nanosized primary particles with a crystalline structure.Based on EDX analyses, the particles of both agglomerates areFe-oxides with low amounts of Mn and Cr and with or withoutSb. The primary particle constituents of these agglomerates areboth spherical as well as square with rounded edges (Figure 6A,B). The spherical particles are 20−100 nm in size, (most ofthem are around 50−60 nm) while the square ones are 5−30nm. In the second agglomerate (Figure 6C, D), the individualparticle constituents show irregular shapes with roundedoutlines and are 10−15 nm in size. Primary soot particles ofsignificantly larger size are rarely found attached to the Fe-oxideparticles (Figure 6C).

Figure 4. DF-STEM images (A, B) of a soot agglomerate downstreamof the DPF and EDX element mapping of a region of this agglomerate.Grayscale bar indicates minimum (black) and maximum (white)element abundances (black is equal to zero). The arrow in the Ca-mapindicates a Ca-rich site.

Figure 5. (A): BF-STEM image of a soot agglomerate showing at theedge primary soot particle(s) (in rectangle) potentially containing ashphases. (B) HRTEM image of particles in (A) analyzed by EDX. Nodistinctive ash phases can be recognized in (B) (see text).

Environmental Science & Technology Article

dx.doi.org/10.1021/es403121y | Environ. Sci. Technol. 2013, 47, 14495−1450114499

Page 6: Metal Particle Emissions in the Exhaust Stream of Diesel Engines: An Electron Microscope Study

The shape, size, and mode of occurrence of the Fe-oxidenanoparticles imply that they are newly formed at hightemperatures, likely from a liquid or partly gas phase. Theprocess of Fe nanoparticle formation and oxidation is expectedto take place in the cylinder, where high temperatures prevail.Because neither Fe-borne additives (e.g., ferrocene3,26) norbiofuels were used during the experiment, the nanosized Fe-oxide particles may have formed via combustion of extremelysmall quantities of wear materials split off by mechanicalabrasion, most likely from the injection system or from otherFe-bearing components of the engine prior to the cylinder. Thecomposition of the particles (Fe with some Mn and Cr), whichcorresponds to that of an alloy commonly used in enginecomponents, is in line with this assumption. Alternatively, usedoil containing Fe-bearing engine wear materials may haveentered the combustion chamber and provided the Fe for theformation of the Fe-oxide nanoparticles. Once formed, the iron-oxide agglomerates can enter the exhaust system and eventuallyreach the ambient air.

5. DISCUSSIONThe advantage of the electron microscopic methods used in thepresent paper is that they physically sample, examine andcharacterize the nature and properties of emitted dieselparticles and agglomerates. Although a larger sample size andmore rigorous statistical analysis would be desirable for drawingspecific conclusions, this cannot be the case here becauseelectron microscope techniques aim at focusing on micro- andnanometer scales to reveal important details.The results of this paper highlight several important points

with respect to the nature, morphology, size, and differences inthe relative amounts of particles and agglomerates/aggregatesthat participate in the composition of diesel PM emissionsbefore and after exhaust after-treatment:

1. Ash aggregates consisting of sintered primary particlesescape into the ambient air either attached onto large

soot agglomerates or in the form of isolated individualentities not bound to soot. The size of the ash aggregatesgenerally ranges between 0.2 and 2 μm, and their shape isirregular and usually rounded. The primary constituentsof the aggregates are ca. 20−400 nm in size. Micrometer-to nanometer-size particles may be harmful to therespiratory and cardiovascular systems, especially if theyare insoluble, as is the case with the ash components.Taking into account that metal emissions are produced

not only by diesel but by any internal combustion engine,the findings of this research are gaining broadsignificance. Gasoline cars, for instance, currently haveno after-treatment system for PM. Detailed studiesconcerning metal emissions from gasoline cars arelacking.

2. Ash aggregates upstream and downstream of the DPFconsist of Ca, Zn, P, Mg, S, Na, Al, K (from the lube oil),Fe, Cr, Ni, Sn, Al, Sb (from engine wear and/orcorrosion) as well as Pd (from the DOC coating).

3. Soot agglomerates upstream of the DPF are metal-free orcontain notably low amounts of ash. In contrast, the fewsoot agglomerates that escape in the ambient aircommonly carry lubricating oil- and wear-relatedsubstances from ash accumulations inside the DPF.

4. The majority of soot agglomerates that escape to theambient air through the DPF are of quite large size (>1−5 μm and even up to 13 μm). Very small sootagglomerates (<0.5 μm) are rather scarce in post-DPFsamples.Based on findings (3) and (4), it is further inferred

that: (a) the DPF can promote breakout of large sootagglomerates in the atmosphere by causing sintering ofthe trapped particles. Soot agglomerates that escapefiltration are of significantly larger dimensions thanhitherto believed. (b) The ash accumulated duringengine operation inside the DPF can be carried awayas separate aggregates but also as attached onto escapingsoot agglomerates. Because the escaping soot agglomer-ates are of rather large size, they are likely to carryrelatively high ash quantities. Ash may be subsequentlyliberated from the soot agglomerate carriers, eithermechanically or after eventual soot oxidation. Thus, theamount of ash escaping to the ambient air is increased byattachment onto soot and should be added to theseparate ash aggregates that escape filtration. Thissituation could cause additional environmental concernsand requires further investigation.

5. Noble metals (here Pd) used in the coating material ofthe DOC can be detached and carried away by theexhaust stream through the DPF to the atmosphere.

6. Trace amounts of Fe-bearing wear may participate in thecombustion process, thereby producing agglomerates ofFe-oxide nanoparticles, which enter the exhaust streamand may reach the ambient air. The source of Fe couldbe Fe pieces split off from components of the engineprior to the injection system or from the injection systemitself.

■ AUTHOR INFORMATIONCorresponding Author*(A.L.) Phone: +41 58 7654190; fax: +41 58 7654041; e-mail:[email protected].

Figure 6. TEM (A, C) and HRTEM images (B, D) of Fe-oxideagglomerates found in samples downstream of the DPF. Therectangles denote sites of EDX analyses.

Environmental Science & Technology Article

dx.doi.org/10.1021/es403121y | Environ. Sci. Technol. 2013, 47, 14495−1450114500

Page 7: Metal Particle Emissions in the Exhaust Stream of Diesel Engines: An Electron Microscope Study

NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSWe gratefully acknowledge financial support from theCompetence Center for Energy and Mobility (CCEM).

■ REFERENCES(1) Bond, T. C.; Streets, D. G.; Yarber, K. F.; Nelson, S. M.; Woo, J.-H.; Klimont, Z. A technology-based global inventory of black andorganic carbon emissions from combustion. J. Geophys. Res. 2004, 109(D14), D14203.(2) World Health Organization. International Agency for Researchon Cancer, Press release No. 213, 2012.(3) Lee, D.; Miller, A.; Kittelson, D.; Zachariah, M. R. Character-ization of metal-bearing diesel nanoparticles using single-particle massspectrometry. J. Aerosol Sci. 2006, 37 (1), 88−110.(4) Kittelson, D. B.; Watts, W. F.; Johnson, J. P. On-road andlaboratory evaluation of combustion aerosolsPart 1: Summary ofdiesel engine results. J. Aerosol Sci. 2006, 37 (8), 913−930.(5) Bardasz, E. A.; Cowling, S.; Panesar, A.; Durham, J.; Tadrous, T.N. Effects of lubricant derived chemistries on performance of thecatalyzed diesel particulate filters. SAE Int. 2005, DOI: 10.4271/2005-01-2168.(6) Sappok, A. G.; Wong, V. W. Detailed chemical and physicalcharacterization of ash species in diesel exhaust entering aftertreatmentsystems. SAE Int. 2007, DOI: 10.4271/2007-01-0318.(7) Vaaraslahti, K.; Keskinen, J.; Giechaskiel, B.; Solla, A.; Murtonen,T.; Vesala, H. Effect of lubricant on the formation of heavy-duty dieselexhaust nanoparticles. Environ. Sci. Technol. 2005, 39 (21), 8497−8504.(8) Fino, D. Diesel emission control: Catalytic filters for particulateremoval. Sci. Technol. Adv. Mater. 2007, 8 (1−2), 93.(9) Mohr, M.; Forss, A.-M.; Lehmann, U. Particle emissions fromdiesel passenger cars equipped with a particle trap in comparison toother technologies. Environ. Sci. Technol. 2006, 40 (7), 2375−2383.(10) Schreiber, D.; Forss, A. M.; Mohr, M.; Dimopoulos, P. Particlecharacterisation of modern CNG, gasoline and diesel passenger cars.SAE Int. 2007, DOI: 10.4271/2007-24-0123.(11) Twigg, M. V. Progress and future challenges in controllingautomotive exhaust gas emissions. Appl. Catal., B 2007, 70 (1−4), 2−15.(12) Sappok, A.; Kamp, C.; Wong, V. Sensitivity analysis of ashpacking and distribution in diesel particulate filters to transientchanges in exhaust conditions. SAE Int. J. Fuels Lubr. 2012, 5 (2),733−750.(13) Sappok, A.; Santiago, M.; Vianna, T.; Wong, V. W.Characteristics and effects of ash accumulation on diesel particulatefilter performance: rapidly aged and field aged results. SAE Int. 2009,DOI: 10.4271/2009-01-1086.(14) Liati, A.; Dimopoulos Eggenschwiler, P.; Muller Gubler, E.;Schreiber, D.; Aguirre, M. Investigation of diesel ash particulatematter: A scanning electron microscope and transmission electronmicroscope study. Atmos. Environ. 2012, 49 (0), 391−402.(15) Miller, A.; Ahlstrand, G.; Kittelson, D.; Zachariah, M. The fateof metal (Fe) during diesel combustion: Morphology, chemistry, andformation pathways of nanoparticles. Combust. Flame 2007, 149 (1−2), 129−143.(16) Bodek, K. M.; Wong, V. V. The effects of sulfated ash,phosphorus and sulfur on diesel aftertreatment systemsA review.SAE Int. 2007, DOI: 10.4271/2007-01-1922.(17) McGeehan, J. A.; Van Dam, W.; Narasaki, K.; Boffa, A. B.;Rutherford, J. A.; Carabell, K.; Stokes, C.; Nelson, K.; Yeh, S. W.Extending the boundaries of diesel particulate filter maintenance withultra-low ash - zero-phosphorus oil. SAE Int. J. Fuels Lubr. 2012, 5 (3),1240−1263.(18) Kimura, K.; Lynskey, M.; Corrigan, E. R.; Hickman, D. L.;Wang, J.; Fang, H. L.; Chatterjee, S. Real world study of diesel

particulate filter ash accumulation in heavy-duty diesel trucks. SAE Int.2006, DOI: 10.4271/2006-01-3257.(19) Watson, S. A. G., Lubricant-derived ash - In engine sources andopportunities for reduction. PhD Dissertation, Massachussets Instituteof Technology 2010.(20) Costa, D. L.; Dreher, K. L. Bioavailable transition metals inparticulate matter mediate cardiopulmonary injury in healthy andcompromised animal models. Environ. Health Perspect. 1997, 105 (5),1053−1060.(21) Zervas, E.; Dorlhene, P. Comparison of exhaust particle numbermeasured by EEPS, CPC, and ELPI. Aerosol Sci. Technol. 2006, 40(11), 977−984.(22) Cross, E. S.; Sappok, A.; Fortner, E. C.; Hunter, J. F.; Jayne, J.T.; Wong, V. W.; Trimborn, A.; Worsnop, D. R.; Kroll, J. H.; Brooks,W. A.; Onasch, T. B. Real-time measurements of engine-out traceelements: application of a novel soot particle aerosol massspectrometer for emissions characterization. J. Eng. Gas TurbinesPower 2012, 134 (7), 072801−072801.(23) Liati, A.; Dimopoulos Eggenschwiler, P. Characterization ofparticulate matter deposited in diesel particulate filters: Visual andanalytical approach in macro-, micro- and nano-scales. Combust. Flame2010, 157 (9), 1658−1670.(24) Liati, A.; Spiteri, A.; Dimopoulos Eggenschwiler, P.; Vogel-Schauble, N. Microscopic investigation of soot and ash particulatematter derived from biofuel and diesel: Implications for the reactivityof soot. J. Nanopart. Res. 2012, 14 (11), 1−18.(25) Kirchner, U.; Scheer, V.; Vogt, R.; Kagi, R. TEM study onvolatility and potential presence of solid cores in nucleation modeparticles from diesel powered passenger cars. J. Aerosol Sci. 2009, 40(1), 55−64.(26) Schulz, G. A. S.; Tamborim, S.; Cardoso, G.; Santos, T.; Lissner,E.; Cataluna, R. Carbon oxidation generated in diesel engines usingiron-doped fuel. Green Chem. 2012, 14 (2), 514−518.

Environmental Science & Technology Article

dx.doi.org/10.1021/es403121y | Environ. Sci. Technol. 2013, 47, 14495−1450114501