characterization of chemical and particulate emissions from aircraft engines

ARTICLE IN PRESS

1352-2310/$ - se

doi:10.1016/j.at

�CorrespondEnvironmental

Hall A321, Riv

Tel.: +1 951 78

E-mail addr1Now at Joh

Atmospheric Environment 42 (2008) 4380–4392

www.elsevier.com/locate/atmosenv

Characterization of chemical and particulate emissions fromaircraft engines

Harshit Agrawala,b, Aniket A. Sawanta,b,1, Karel Jansena,b, J. Wayne Millera,b,David R. Cocker IIIa,b,�

aDepartment of Chemical and Environmental Engineering, University of California, Bourns Hall A321,

UC Riverside, Riverside, CA 92521, USAbCollege of Engineering, Center for Environmental Research and Technology (CE-CERT),

1084 Columbia Avenue, Riverside, CA 92507, USA

Received 7 August 2007; received in revised form 4 January 2008; accepted 5 January 2008

Abstract

This paper presents a series of measurements from four on-wing, commercial aircraft engines, including two newer

CFM56-7 engines and two earlier CFM56-3 engines. Samples were collected from each engine using a probe positioned

behind the exhaust nozzle of the aircraft, chocked on a concrete testing pad. The emission factors for particulate matter

mass, elemental and organic carbon, carbonyls, polycyclic aromatic hydrocarbons, n-alkanes, dioxins, metals and ions are

reported for four different engine power setting modes. The emissions indices of particulate matter, elemental and organic

carbon are highly power dependent for these engines. Particulate matter emission indices (g kg�1 fuel) are found to increase

from 1.1E�02 to 2.05E�01 with increase in power from idle to 85%. The elemental carbon to organic carbon varies from

0.5 to 3.8 with change in power from idle to 85%. The carbonyl emissions are dominated by formaldehyde. The emission

index of formaldehyde ranges from 2.3E�01 to 4.8E�01 g kg�1 fuel. The distribution of metals depends on the difference

in the various engines. The dioxin emissions from the aircraft engines are observed to be below detection limit.

r 2008 Elsevier Ltd. All rights reserved.

Keywords: Jet turbine emissions; PAHs; Carbonyls; Particulate matter; Aircraft

1. Introduction

The growth of commercial air traffic over the lastdecade has led to an increased contribution to the localinventory of gaseous and particle emissions from theoperations of vehicles, ground-equipment and aircraft

e front matter r 2008 Elsevier Ltd. All rights reserved

mosenv.2008.01.069

ing author at: Department of Chemical and

Engineering, University of California, Bourns

erside, Riverside, CA 92521, USA.

1 5695; fax: +1 951 781 5790.

ess: [email protected] (D.R. Cocker III).

nson-Matthey.

associated with airports. Some studies in past haveexamined the emissions of nitrogen oxides, blackcarbon (BC) and organic composition of exhaust fromturbine engines; however, still there is a need forinformation on the emission factors of speciatedparticulate matter (PM), carbonyls, polycyclic aromatichydrocarbons (PAH), metals and speciated hydrocar-bon (HC) as a function of engine load. These data arenecessary for detailed chemical emissions inventorydevelopment, and hazardous risk assessment.

Spicer et al. (1984) first reported oxides ofnitrogen (NOx), carbon monoxide (CO), PM, and

.

www.elsevier.com/locate/atmosenv

dx.doi.org/10.1016/j.atmosenv.2008.01.069

mailto:[email protected]

ARTICLE IN PRESS

Table 1

List of engines tested and associated airframes

Date Airframe Engine Plane no.

23 August 2005 B737-700 CFM56-7B22 Plane 1




H. Agrawal et al. / Atmospheric Environment 42 (2008) 4380–4392 4381

speciated hydrocarbon (HC) emission rates from jetaircraft turbine engines. Further studies by Spiceret al. (1992, 1994) provided additional informationon the detailed organic composition of turbineengine emissions. During engine idle, they foundthat emissions were dominated by products fromfuel cracking, unburned fuel and products ofincomplete combustion, with ethene, propene,acetylene and formaldehyde comprising 30–40%of the total HC emissions. Slemr et al. (2001) reportsimilar findings for a more modern commercial highbypass turbofan (CFM56-2C1) and an older tech-nology engine (Rolls Royce M45H Mk501).Furthermore, Slemr et al. (2001) noted that theemission indices (EIs) for these engines are highlypower dependent and dominated by alkenes andalkynes related to fuel cracking and aromaticcompounds arising from unburned fuel.

Petzold et al. (1999) reports the characterizationon BC aerosol from the jet exhaust as a function ofvarious engine thrust settings. Popp et al. (1999)measured variety of aircrafts in a mix of idle, taxi-out, and takeoff modes via remote sensing atLondon’s Heathrow Airport. Herndon et al.(2006) presents speciated HC emissions from theplumes of in-use commercial aircraft during airportoperations. Anderson et al. (2006) have reportedselected carbon species EIs for a Rolls RoyceRB211-535-E4 turbofan engine. Recently, Wey etal. (2007) investigated the gas-phase and particleemissions from a CFM56-2C1 engine.

The present study presents comprehensive mea-surements and analyses of the PM mass, metals,elemental and organic carbon, carbonyls and PAHsfrom two Boeing 737-700 s equipped with CFM56-7engines and two Boeing 737-300s equipped withCFM56-3 engines as a function of engine load usinga modified EPA landing takeoff (LTO) cycle.

2. Experimental methods

2.1. Engine description and fuel

Sampling was conducted on two Boeing 737-300aircraft equipped with CFM56-3 turbofans and twoBoeing 737-700 aircraft with CFM56-7 turbofanengines made by CFM international. There arecurrently more CFM turbofans in service than anyother turbofan (CFM, 2006). Both the CFM56-3and CFM56-7 were running on Jet A fuel for thistest. Engine specifications are provided in Table 1.

2.2. Test cycle

The EPA’s LTO cycle uses four thrust settings:7%, 30%, 85%, and 100% of full power (US EPA,September 1985); however, for this research, wemodified the EPA LTO cycle to include additionalthrust settings at 4%, 40%, and 65% of full powerand eliminated the 100% power, as this point wasnot feasible during ground operation. To ensuresufficient sample loading for subsequent chemicalanalysis, the 4% and 7% load points were groupedtogether (called Mode 1) and the 30% and 40%load points were grouped together (Mode 2), whilethe 65% (Mode 3) and 85% (Mode 4) load pointswere tested separately. Sampling was initialized 15 safter stable operating conditions were achieved andstopped 15 s prior to switching to the next loadpoint. During transitions between modes, all flowlines were directed to the bypass valves.

2.3. Exhaust sampling and sampling lines

The aircraft engine emissions were tested on-wingat the Ground Runup Enclosure (GRE) at OaklandInternational Airport. Custom-designed probes andextensive support equipment were used to sample jetexhaust in the on-wing position (Wey, 2006). Threesets of sampling lines were used in this research.Raw exhaust was extracted from the port andstarboard combustor/engine exhaust flow throughprobes positioned about 1m from the exhaustnozzle exit plane as this distance is used duringcollection of the engine certification data in theICAO database.

On the starboard side of the aircraft a 50-ft by 3/8-in stainless steel insulated sampling line heated to300 1F transferred the raw exhaust to the primarysampling system used to measure the mass ofspeciated PM and speciated HC. A second 75-ftby 3/8-in insulated stainless line was used to carrysample of raw exhaust extracted from the portsideof the aircraft. In the case of the second line, once

ARTICLE IN PRESSH. Agrawal et al. / Atmospheric Environment 42 (2008) 4380–43924382

the extracted sample reached the measurementbench, it was divided with a ‘‘tee’’ into two streams.The portion that flowed directly through wasdedicated for the analysis of the hexavalent chro-mium and the sample portion at right angles to thisoriginal flow was used for the dioxin analysis. Athird sample line of 3/8-in uninsulated stainless lineof 25 feet was used to deliver diluted exhaust samplefrom a community-mixing box. This sample wascollected with PM sampler that added nitrogendilution right at the sample tip to suppressparticle–particle interactions during transport inthe sample lines and limit condensation. Thissample line was the feed for our realtime particleanalyzer. Typical dilution ratios were in the range10–40.

2.4. Sampling system

Based on prior research experience and theplanned operating cycle, a special complex systemwas designed for sampling multiple modes of a jetexhaust with rapid automatic switching between thespecified engine operating modes (Fig. 1). Thesystem made use of computer controlled switchingvalves and high-volume flow capabilities to collect

Fig. 1. Schematic representat

sufficient molecules onto the substrate in the limitedtime allotted for each load point. Samples weretaken from the plume of the turbofan exhaustthrough a probe, and directed through heated linesto the sampling system developed in house foraircraft sampling. The four heated 3/8-in stainless-steel sampling lines directed sample exhaust into amixing column and through a PM2.5 particleimpactor. The sampling system allowed collectionof both gas and particle phase components onvarious sampling media. Critical orifices wereused to control flow rates through all systems andwere continuously monitored for inlet temperature,and inlet and out pressure. With the exception ofthe SUMMAs Canisters, all flows were operatedunder choked conditions (outlet pressure 50.52�inlet pressure). SUMMAs Canisters were filledthrough a small orifice to lower the samplingrate, but absolute flow was not controlled sinceSUMMAs Canisters measurements are made on avolume concentration basis. In addition, the designincluded a laminar flow element that continuouslymonitored flow into the low flow carbonyl-samplingmanifold. Flow rates for each engine load pointwere corrected for pressure and temperature. Theparticle transmission losses are not accounted for inthis study.

ion of sampling system.

ARTICLE IN PRESSH. Agrawal et al. / Atmospheric Environment 42 (2008) 4380–4392 4383

A final overarching check was built into thesampling system by placing SUMMAs Canistersat one of the extremities of the system where theflow rate was the lowest and most sensitive tosampling errors. These canisters were analyzedby an external laboratory for carbon dioxide(CO2). The CO2 values were especially important,as they would indicate if ambient air was introducedto the system and to verify CO2 measurementswith CO2 measurements in the raw exhaust for EIcalculations.

2.5. Measurement of PM mass, metals, and ions

PM2.5 mass, metals, and ions were collected on47mm diameter 2 mm pore Teflo filters (Pall Gel-man, Ann Arbor, MI). The filters were measured forPM2.5 using a Cahn C-35 (Madison, WI) micro-balance following Code of Federal Regulations-40(CFR40). Before and after collection, the filterswere conditioned for 24 h in an environmentallycontrolled room (RH ¼ 40%, T ¼ 25 1C) andweighed daily until two consecutive weight mea-surements were within 3 mg. The Teflo filters weresubsequently analyzed for metals using the X-rayfluorescence (XRF) method as per EPA IO-3 at theSouth Coast Air Quality Management District.Finally, the filters were extracted with high perfor-mance liquid chromatography (HPLC) water andisopropyl alcohol and analyzed for ions (sulfate,chloride, and nitrate) using a Dionex DX-120 ionchromatograph.

2.6. Measurement of elemental and organic carbon

(EC-OC)

EC-OC analysis was performed on PM samplescollected on 2500 QAT-UP Tissuquartz Pall (AnnArbor, MI) 47mm filters that were preconditionedat 600 1C for 5 h. A 1.5-cm2 punch was taken fromthe quartz filter and analyzed with a SunsetLaboratory (Forest Grove, OR) Thermal/OpticalCarbon Aerosol Analyzer according to the NIOSH5040 reference method (NIOSH, 1996).

2.7. Carbonyls

Traditional air monitoring methods for very-volatile and volatile organic compounds (VVOC/VOC) are not sufficiently sensitive enough tomeasure the low levels of most compounds foundin an exhaust from a lean burn engine, like a

turbine. Accordingly, most of the sampling in thiswork makes use of selective adsorbents for con-centrating the molecules of interest after the exhaustgas was passed through the Teflon filter.

Carbonyls were collected on 2,4-dinitrophenylhy-drazine (DNPH) coated silica cartridges (WatersCorp., Milford, MA) after the Teflo filter. DNPHcartridges were extracted using 5mL of acetonitrileand injected into an Agilent 1100 series HPLCequipped with a 5 mm Deltabond AK Resolution(200 cm� 4.6mm ID) with upstream guard columnand a Diode array detector. The HPLC sampleinjection and operating conditions followed theSAE 930142HP protocol (Siegl et al., 1993).

2.8. C10 to C30 hydrocarbons, including naphthalene

and PAHs

Raw jet exhaust flows were collected through aquartz filter and into a column packed withpolyurethane foam (PUF)/XAD-4 resin. A portionof the quartz filter was used to analyze for theelemental and organic carbon. Both the PUF/XAD-4 cartridge and the remainder of quartz filter wasextracted with methylene chloride and analyzedusing a modified method EPA TO13A protocol(GC–MS analysis) to determine total emission ratesfor n-alkanes and PAHs. Details on the analysismethod are found in Shah et al. (2004a, b).

3. Results and discussion

3.1. Sampler QA/QC

A custom-built high-volume sampler was de-signed and constructed for this test. The systemdesign included a number of internal consistencychecks to verify the integrity of the system andthe data set. These included: redundant filtercollection for carbon analysis (elemental andorganic carbon), redundant flow measurement andsensing devices, and redundant internally measuredCO2 concentrations to account for dilution. Allflows were verified at the point of sample collectionto be consistent with manufacturer specifications fororifices and pressure drops. The sampler wasdesigned to provide mass concentrations in thesample stream for subsequent EI calculations usingCO2 data.

For example, direct comparison of two parallelsamples made between the total carbon collected onthe two branches of the sampling system is

ARTICLE IN PRESS

0500

100015002000250030003500400045005000

0

TC (QZ) ug/m3

TC (Q

Z/PU

F) u

g/m

3

1000 2000 3000 4000 5000

Fig. 2. Correlation of total carbon concentration measured

between parallel sample trains. Equation of the line:

y ¼ 0.978x+131.31; coefficient of determination ¼ 0.9936.

y = 1.02x + 15R2 = 0.997

0

1000

2000

3000

4000

5000

6000

0

TC Quartz (ug m-3)

PM T

eflo

(ug

m-3

)

1000 2000 3000 4000 5000

Fig. 3. Comparison of PM mass concentration (measured on

Teflo) vs. TC mass concentration measured on high flow quartz.

Table 2

EI (g kg�1 fuel)

Mode 1 Mode 2 Mode 3 Mode 4

CFM56-7B22

PM 1.25E�02 3.18E�02 6.20E�02 1.07E�01

EC 3.86E�03 2.60E�03 3.63E�02 6.31E�02

OC 1.14E�02 4.64E�03 9.98E�03 2.56E�02

CFM56-3B1

PM 1.41E�02 8.80E�03 5.97E�02 2.78E�01

EC 6.74E�03 5.28E�03 1.56E�02 2.35E�01

OC 9.84E�03 5.27E�03 2.65E�02 3.37E�02

CFM56-3B2

PM 1.56E�02 2.98E�02 1.23E�01 3.29E�01

EC 8.52E�03 2.06E�02 9.07E�02 2.60E�01

OC 1.23E�02 9.77E�03 2.39E�02 6.19E�02

CFM56-7B22

PM 2.60E�03 3.34E�03 6.02E�02 1.07E�01

EC 1.22E�03 2.41E�03 5.15E�02 7.24E�02

OC 4.92E�03 4.25E�03 9.05E�03 3.71E�02

H. Agrawal et al. / Atmospheric Environment 42 (2008) 4380–43924384

displayed (mgm�3 basis) in Fig. 2. One train wasdesigned for high flow through the quartz to ensuresufficient sample for EC-OC analyses while theother train is the quartz-PUF combination forsemi-volatile analysis. The agreement between theparallel samples with a slope (0.98) and regressionvalue obtained (R2

¼ 0.994) provides confidencein the quality of samples on these high-volumelines.

Next, the total PM mass concentration measuredon the Teflo filters are compared with the totalcarbon concentration obtained on the high flowquartz leg. This comparison provides anotherconsistency check between measurements of PM.Excellent agreement was noted between these legs(Fig. 3) with a slope of 1.02 and an R2 of 0.997noted for the test. This provides confidence in themeasured values on these three legs of the samplingsystem.

3.2. Particulate matter, EC-OC

The EI is defined as the mass of pollutant emittedper unit mass of fuel burned. Table 2 summarizesthe emission indices measured for PM and EC-OC.A 10–40-fold increase in the PM EI is noted as thepower increased from idle (Mode 1) to 85% power(Fig. 4). Furthermore, the PM shifts from OC richfor idle mode (Mode 1) to EC rich with increasingpower (Fig. 4). Total carbon and BC EIs data byPetzold and Schroeder (1998) also support thisobservation. Recently, Wey et al. (2007) have alsoreported that particle mass and BC EIs areminimum at low powers and increased with power,reaching values more than 0.3 g kg�1 fuel at powerlevels higher than 85%, which is congruent with theresults observed in this study. The OC positiveartifact in this study was assumed to offsethydrogen and oxygen content of the organic mass,based on previous research (Shah et al., 2004a, b).The good correlation between TC and PM supportsthis correction.

3.3. Carbonyls

The major three contributors to the carbonylemissions are formaldehyde, acetaldehyde, andacetone. The carbonyl HPLC-DAD analysisshowed a large, co-eluting peak (identity unknown)with acrolein, which prevented us from accurately

ARTICLE IN PRESS

0.35

0.3

0.25

0.2

0.15

0.1

0.05

0

g/kg

fuel

Mod

e1

Mod

e2

Mod

e3

Mod

e4

Mod

e1

Mod

e2

Mod

e3

Mod

e4

Mod

e1

Mod

e2

Mod

e3

Mod

e4

Mod

e1

Mod

e2

Mod

e3

Mod

e4

CFM56-7B22 CFM56-3B1 CFM56-3B2 CFM56-7B22

PM EC OC

Fig. 4. Emissions Indices of PM, EC and OC for different planes as a function of modes.

mg/

kg fu

el

Mod

e1

Mod

e2

Mod

e3

Mod

e4

Mod

e1

Mod

e2

Mod

e3

Mod

e4

Mod

e1

Mod

e2

Mod

e3

Mod

e4

Mod

e1

Mod

e2

Mod

e3

Mod

e4

CFM56-7B22 CFM56-3B1 CFM56-3B2 CFM56-7B22

1000

100

10

1

Formaldehyde Acetaldehyde Acetone Other∗

Fig. 5. Relative contribution of carbonyls for different planes as a function of mode. *Other does not include acrolein.


quantifying acrolein in the present study. Theunknown compound was unstable in solution uponrefrigeration for extended periods of time. The EIsof carbonyls are given in Fig. 5 for different planesas a function of engine power. Formaldehyde andacetaldehyde are most dominant carbonyl species inthe aircraft exhaust emissions. The formaldehydeemission indices in this study are low as comparedto that observed by Spicer et al. (1992). The largevariability in the formaldehyde measurements is

also explained by Knighton et al. (2007) to becaused by changes in ambient temperature.

3.4. Polycyclic aromatic hydrocarbons (PAH) and

n-alkanes

PAHs and n-alkanes (C12–C30) were collected onquartz followed by a PUF-XAD-4-PUF sandwich.The EIs, calculated from the analysis of XAD-4 andquartz filter, for PAHs and n-alkanes is shown in

ARTIC

LEIN

PRES

S

Table 3

Emission indices of different PAHs and alkanes as a function of engine load

Unit Plane 1 Plane 2 Plane 3 Plane 4

Mode 1 Mode 2 Mode 3 Mode 4 Mode 1 Mode 2 Mode 3 Mode 4 Mode 1 Mode 2 Mode 3 Mode 4 Mode 1 Mode 2 Mode 3 Mode 4

PAH

Naphthalene g kg�1

fuel

1.64E�02 7.15E�01 5.30E�01 7.74E�02 9.69E�03 1.26E+00 5.52E+00 6.68E�02 2.16E�02 2.18E�01 2.25E+00 1.06E+00 7.53E�03 1.90E+00 6.26E�01 6.85E�01

1-Methylnaphthalene g kg�1

fuel

1.94E�03 5.25E�03 1.23E�03 1.25E�03 6.75E�04 2.28E�03 7.69E�03 7.72E�04 9.01E�04 8.03E�04 6.58E�03 1.33E�03 4.01E�04 3.23E�03 1.19E�03 6.56E�04

2-Methylnaphthalene g kg�1

fuel


Acenaphthylene g kg�1

fuel


Acenaphthene g kg�1

fuel


Fluorene g kg�1

fuel


Phenanthrene g kg�1

fuel


Anthracene g kg�1

fuel


Fluoranthene g kg�1

fuel


Pyrene g kg�1

fuel


Benz(a)anthracene g kg�1

fuel


Chrysene g kg�1

fuel


Benzo(b)fluoranthene g kg�1

fuel

1.35E�05 6.32E�06 8.42E�06 o9.0E�09 o8.7E�10 o9.3E�10 1.69E�05 o5.7E�09 9.85E�06 8.62E�06 o2.2E�09 o7.6E�09 4.78E�06 5.57E�06 o1.3E�09 o5.2E�09

Benzo(k)fluoranthene g kg�1

fuel

3.75E�05 1.37E�05 7.23E�06 o9.0E�09 o8.7E�10 o9.3E�10 3.73E�05 6.35E�05 2.05E�05 1.42E�05 9.85E�06 5.83E�05 9.97E�06 9.15E�06 5.64E�06 3.99E�05

Benzo(a)pyrene g kg�1

fuel

9.88E�06 o1.7E�09 o2.0E�09 o9.0E�09 o8.7E�10 o9.3E�10 1.45E�05 2.39E�05 2.26E�05 3.35E�06 5.34E�06 2.43E�05 1.10E�05 2.16E�06 3.06E�06 1.66E�05

Indeno[1,2,3-

cd]pyrene

g kg�1

fuel

o2.3E�09 o1.7E�09 o2.0E�09 o9.0E�09 o8.7E�10 o9.3E�10 1.16E�05 o5.7E�09 4.99E�06 o1.5E�09 o2.2E�09 1.96E�05 2.42E�06 o9.7E�10 o1.3E�09 1.34E�05

H.

Ag

raw

al

eta

l./

Atm

osp

heric

En

viron

men

t4

2(

20

08

)4

38

0–

43

92

4386

ARTIC

LEIN

PRES

SDibenz[a,h]anthracene g kg�1

fuel

o2.3E�09 o1.7E�09 o2.0E�09 o9.0E�09 o8.7E�10 o9.3E�10 1.93E�05 o5.7E�09 o1.4E�09 o1.5E�09 o2.2E�09 1.65E�05 o6.6E�10 o9.7E�10 o1.3E�09 1.13E�05

Benzo[ghi]perylene g kg�1

fuel

o2.3E�09 o1.7E�09 o2.0E�09 o9.0E�09 o8.7E�10 o9.3E�10 o1.4E�09 o5.7E�09 5.24E�06 o1.5E�09 o2.2E�09 o7.6E�09 2.55E�06 o9.7E�10 o1.3E�09 o5.2E�09

Alkanes

Dodecane g kg�1

fuel


Tetradecane g kg�1

fuel


Hexadecane g kg�1

fuel


Octadecane g kg�1

fuel


Nonadecane g kg�1

fuel


Eicosane g kg�1

fuel


Docosane g kg�1

fuel


Tetracosane g kg�1

fuel


Hexacosane g kg�1

fuel


Octacosane g kg�1

fuel


Triacontane g kg�1

fuel


H.

Ag

raw

al

eta

l./

Atm

osp

heric

En

viron

men

t4

2(

20

08

)4

38

0–

43

92

4387

ARTICLE IN PRESS

Table 4

Detection limit for various dioxins analyzed

Analyte DL

(pgm�3)

2,3,7,8-TCDD 2,3,7,8-Tetrachlorodibenzo-p-

Dioxin

o0.78

1,2,3,7,8-PeCDD 1,2,3,7,8-Pentachlorodibenzo-p-

dioxin

o1.06

1,2,3,4,7,8-

HxCDD

1,2,3,4,7,8-Hexacholorodibenzo-p-

dioxin

o1.68

1,2,3,6,7,8-

HxCDD

1,2,3,6,7,8-

Hexacholorodibenzofuran

o1.48

1,2,3,7,8,9-

HxCDD

1,2,3,7,8,9-

hexachlorodibenzofuran

o1.58

1,2,3,4,6,7,8-

HpCDD

1,2,3,4,6,7,8-Heptachlorodibenzo-

p-dioxin

o1.04

OCDD Octachlorodibenzofuran o2.43

2,3,7,8-TCDF 2,3,7,8-Pentachlorodibenzofuran o0.78

1,2,3,7,8-PeCDF 1,2,3,7,8-Pentachlorodibenzofuran o1.16

2,3,4,7,8-PeCDF 2,3,4,7,8-Pentachlorodibenzofuran o1.09

1,2,3,4,7,8-

HxCDF

1,2,3,4,7,8-


o0.38

1,2,3,6,7,8-

HxCDF

1,2,3,6,7,8-


o0.31

2,3,4,6,7,8-

HxCDF

2,3,4,6,7,8-


o0.35

1,2,3,7,8,9-

HxCDF

1,2,3,7,8,9-


o0.45

1,2,3,4,6,7,8-

HpCDF

1,2,3,4,6,7,8-

Heptachlorodibenzofuran

o0.69

1,2,3,4,7,8,9-

HpCDF

1,2,3,4,7,8,9-

Heptachlorodibenzofuran

o0.9

OCDF Octachlorodibenzo-p-dioxin o1.94

g/kg

fuel

Mod

e1

Mod

e2

Mod

e3

Mod

e4

Mod

e1

Mod

e2

Mod

e3

CFM56-7B22 CFM56-3B1

2.E-03

2.E-03

2.E-03

1.E-03

1.E-03

1.E-03

8.E-04

6.E-04

4.E-04

2.E-04

0.E+00

AlSMnOtherSiFeCrMg

Fig. 6. Distributio

H. Agrawal et al. / Atmospheric Environment 42 (2008) 4380–43924388

Table 3 for each aircraft. In this work, ‘‘naphtha-lenic’’ PAHs include naphthalene and its 1-methyland 2-methyl derivatives; while ‘‘non-naphthalenic’’PAHs range from acenaphthylene to dibenz[a,h]an-thracene in increasing molecular weight. Thenaphthalenic PAHs are the overwhelmingly domi-nant PAH species measured. The relative distribu-tions of the substituted naphthalenes to non-substituted naphthalenes for the idle modes are ingeneral agreement with the work from Spicer(Spicer et al., 1992, 1994). The naphthalene EIincreased as power increased from idle mode fallingoff as the engine operated at the highest mode for allengines tested.

3.5. Dioxins

In humans, the highly chlorinated dioxins arestored in fatty tissues and are neither readilymetabolized nor excreted. The estimated elimina-tion half-life for highly chlorinated dioxins (4-8chlorine atoms) ranges from 7.8 to 132 years (Geyeret al., 2002). Since the health impacts of even lowlevels of dioxins are significant, dioxin measure-ments were performed for this study. Resultsshowed the concentration of dioxins was belowmethod detection limits for all the aircraft engines.The outside laboratories, which carried out theanalyses, showed excellent recovery of the spiked

Mod

e4

Mod

e1

Mod

e2

Mod

e3

Mod

e4

Mod

e1

Mod

e2

Mod

e3

Mod

e4


n of Metals.

ARTIC

LEIN

PRES

S

Table 5

Emission indices of various metals for four planes as a function of engine load

Unit Plane 1 Plane 2 Plane 3 Plane 4

Mode 1 Mode 2 Mode 3 Mode 4 Mode 1 Mode 2 Mode 3 Mode 4 Mode 1 Mode 2 Mode 3 Mode 4 Mode 1 Mode 2 Mode 3 Mode 4

Metals

Mn gkg�1 fuel 6.69E�06 4.53E�06 2.35E�05 2.04E�05 4.19E�07 2.09E�06 2.28E�06 1.72E�05 2.58E�06 8.46E�06 2.38E�05 4.04E�05 o 2.3E�07 o 3.3E�07 o 5.1E�07 9.74E�06

S g kg�1 fuel o2.3E�05 5.43E�06 2.10E�05 o8.7E�05 2.47E�05 2.29E�05 1.48E�04 5.58E�04 7.30E�05 1.89E�04 4.89E�04 8.01E�04 o5.2E�06 o7.3E�06 1.56E�04 9.01E�05

Al g kg�1 fuel o1.1E�04 o7.5E�05 o1.0E�04 o4.2E�04 6.32E�05 o3.5E�05 1.18E�04 3.40E�04 o5.4E�05 1.35E�04 2.79E�04 o3.1E�04 3.45E�05 8.05E�05 8.26E�05 o2.0E�04

Mg gkg�1 fuel o1.3E�04 1.42E�04 o1.2E�04 o5.1E�04 4.69E�05 o4.0E�05 8.97E�05 o2.9E�04 5.56E�05 1.07E�04 o1.3E�04 6.18E�04 5.20E�05 o4.0E�05 1.11E�04 4.38E�04

Cr g kg�1 fuel 5.22E�05 3.62E�05 1.56E�04 2.09E�04 8.37E�07 1.00E�05 5.32E�06 3.14E�05 5.81E�06 3.24E�05 1.42E�04 1.43E�04 9.17E�07 o6.1E�07 o9.7E�07 7.30E�06

Fe g kg�1 fuel 2.94E�05 1.54E�05 1.73E�05 7.64E�05 2.09E�06 2.29E�05 o4.4E�07 3.43E�05 8.40E�06 3.39E�05 3.39E�05 2.57E�05 9.17E�06 3.87E�06 o4.0E�07 4.87E�06

Si g kg�1 fuel o1.3E�04 o9.0E�05 o1.2E�04 o5.1E�04 7.45E�05 8.30E�05 o8.0E�05 o2.9E�04 o6.0E�05 1.94E�04 o1.3E�04 o3.7E�04 o3.0E�05 o4.0E�05 o7.0E�05 o2.4E�04

P g kg�1 fuel o2.3E�05 o1.5E�05 o2.1E�05 o8.7E�05 o7.1E�06 o7.1E�06 o1.3E�05 o4.8E�05 o1.1E�05 o1.2E�05 o2.1E�05 o6.2E�05 o5.2E�06 o7.3E�06 o1.2E�05 o4.1E�05

Cl g kg�1 fuel o7.6E�06 o5.1E�06 3.58E�05 o2.9E�06 o2.3E�06 o2.3E�06 7.68E�05 o1.6E�05 o3.6E�06 o4.0E�06 o7.1E�06 o2.1E�05 o1.7E�06 o2.4E�06 o3.8E�06 o1.4E�05

K gkg�1 fuel o1.9E�06 o1.3E�06 o1.7E�06 o7.2E�06 o5.9E�07 o5.9E�07 o1.1E�06 o4.1E�06 o9.1E�07 o9.9E�07 o1.8E�06 o5.2E�06 o4.3E�07 o6.1E�07 o9.7E�07 o3.4E�06

Ca g kg�1 fuel 6.69E�06 9.96E�06 6.18E�06 o7.2E�06 4.19E�07 3.75E�06 2.28E�06 5.15E�05 2.58E�06 1.41E�05 2.51E�05 4.41E�05 1.53E�06 o6.1E�07 o9.7E�07 o3.4E�06

Ti g kg�1 fuel 5.35E�06 9.05E�07 4.94E�06 o2.6E�06 2.09E�06 2.50E�06 4.56E�06 8.58E�06 o2.7E�06 4.94E�06 o5.3E�06 o1.5E�05 1.53E�06 1.29E�06 o2.9E�06 o1.0E�05

V gkg�1 fuel o2.2E�06 9.05E�07 o2.1E�06 5.09E�06 o6.9E�07 o6.9E�07 7.61E�07 o4.7E�06 6.46E�07 o1.2E�06 1.25E�06 o6.1E�06 o5.1E�07 o7.1E�07 o1.1E�06 o4.1E�06

Co g kg�1 fuel 2.68E�06 o5.2E�07 o7.2E�07 o2.9E�06 o2.4E�07 o2.4E�07 1.52E�06 8.58E�06 o3.7E�07 7.05E�07 8.78E�06 3.68E�06 o1.8E�07 8.61E�07 6.83E�07 2.43E�06

Ni g kg�1 fuel 5.35E�06 4.53E�06 3.71E�06 1.02E�05 o1.1E�07 1.67E�06 o1.9E�07 1.14E�05 6.46E�07 2.82E�06 1.76E�05 2.21E�05 3.06E�07 8.61E�07 o1.7E�07 2.43E�06

Cu g kg�1 fuel o3.3E�07 o2.3E�07 o3.1E�07 o1.3E�06 o1.1E�07 4.17E�07 1.52E�06 2.86E�06 o1.6E�07 2.12E�06 o3.1E�07 o9.2E�07 o7.6E�08 4.31E�07 o1.7E�07 o6.1E�07

Zn g kg�1 fuel o3.3E�07 3.62E�06 o3.1E�07 o1.3E�06 4.19E�06 4.59E�06 6.08E�06 2.57E�05 6.46E�07 2.47E�05 1.25E�06 4.41E�05 9.17E�07 4.31E�07 6.83E�07 2.43E�06

Ga g kg�1 fuel 1.74E�05 4.53E�06 6.18E�06 7.64E�05 o1.5E�05 o1.5E�05 1.52E�06 o1.0E�04 1.23E�05 4.94E�06 o4.4E�05 2.57E�05 o1.1E�05 o1.5E�05 o2.4E�05 4.87E�06

Ge g kg�1 fuel o1.9E�06 1.81E�06 o1.7E�06 o7.2E�06 8.37E�07 1.25E�06 o1.1E�06 3.14E�05 6.46E�07 4.23E�06 6.27E�06 1.84E�05 o4.3E�07 1.72E�06 o9.7E�07 1.22E�05

As g kg�1 fuel 1.34E�06 o2.2E�07 o3.1E�07 o1.3E�06 o1.1E�07 o1.0E�07 o1.9E�07 o7.1E�07 o1.6E�07 o1.8E�07 o3.1E�07 3.68E�06 o7.6E�08 4.31E�07 6.83E�07 4.87E�06

Se g kg�1 fuel o1.9E�06 o1.3E�06 o1.7E�06 o7.2E�06 o5.9E�07 o5.9E�07 o1.1E�06 o4.1E�06 o9.1E�07 o9.9E�07 o1.8E�06 o5.2E�06 o4.3E�07 o6.1E�07 o9.7E�07 o3.4E�06

Rb g kg�1 fuel o7.8E�07 o5.2E�07 o7.2E�07 o2.9E�06 o2.4E�07 o2.4E�07 o4.4E�07 o1.7E�06 o3.7E�07 7.05E�07 o7.3E�07 o2.1E�06 o1.8E�07 o2.5E�07 o3.9E�07 o1.4E�06

Sr g kg�1 fuel o1.9E�06 o1.3E�06 o1.7E�06 o7.2E�06 o5.9E�07 o5.9E�07 o1.1E�06 o4.1E�06 o9.1E�07 o9.9E�07 o1.8E�06 o5.2E�06 o4.3E�07 o6.1E�07 o9.7E�07 o3.4E�06

Y gkg�1 fuel o3.3E�07 o2.2E�07 1.24E�06 o1.3E�06 o1.1E�07 4.17E�07 1.52E�06 5.72E�06 o1.6E�07 o1.8E�07 o3.1E�07 3.68E�06 o7.6E�08 8.61E�07 6.83E�07 o6.1E�07

Mo gkg�1 fuel 8.03E�06 3.62E�06 3.71E�06 3.05E�05 1.26E�06 4.17E�07 1.52E�06 2.86E�06 1.29E�06 2.12E�06 6.27E�06 1.47E�05 o4.3E�07 8.61E�07 6.83E�07 7.30E�06

Pd g kg�1 fuel o2.2E�06 9.05E�07 4.94E�06 o8.5E�06 o6.9E�07 o6.9E�07 6.08E�06 o4.7E�06 o1.1E�06 7.05E�07 2.51E�06 1.10E�05 3.06E�07 1.29E�06 o1.1E�06 7.30E�06

Cd g kg�1 fuel o1.9E�06 5.43E�06 6.18E�06 5.09E�06 3.35E�06 o5.9E�07 2.28E�06 5.72E�06 3.23E�06 o9.9E�07 3.76E�06 1.10E�05 1.53E�06 o6.1E�07 2.05E�06 1.46E�05

In g kg�1 fuel 1.07E�05 7.24E�06 6.18E�06 6.11E�05 1.67E�06 2.09E�06 9.13E�06 1.43E�05 5.17E�06 9.17E�06 1.51E�05 2.94E�05 2.44E�06 2.15E�06 6.83E�07 1.70E�05

Sn g kg�1 fuel 1.34E�06 1.81E�06 o2.1E�06 o8.5E�06 1.67E�06 o6.9E�07 o1.3E�06 1.43E�05 1.94E�06 o1.2E�06 1.00E�05 3.68E�06 o5.1E�07 1.29E�06 o1.1E�06 9.74E�06

Sb g kg�1 fuel o2.2E�06 o1.5E�06 o2.1E�06 o8.5E�06 o6.9E�07 o6.9E�07 o1.3E�06 o4.7E�06 o1.1E�06 o1.2E�06 o2.1E�06 o6.1E�06 o5.1E�07 o7.1E�07 o1.1E�06 o4.1E�06

Ba g kg�1 fuel o1.9E�05 2.99E�05 6.18E�06 2.55E�05 5.02E�06 2.50E�06 3.04E�06 o4.1E�05 7.11E�06 1.13E�05 2.26E�05 3.68E�06 o4.3E�06 o6.1E�06 1.09E�05 o3.4E�05

La g kg�1 fuel o4.7E�05 o3.2E�05 o4.3E�05 o1.8E�04 o1.5E�05 o1.5E�05 o2.7E�05 o1.0E�04 o2.3E�05 o2.5E�05 o4.4E�05 o1.3E�04 o1.1E�05 o1.5E�05 o2.4E�05 o8.6E�05

Pb g kg�1 fuel o5.7E�06 o3.8E�06 o5.2E�06 o2.6E�06 o1.8E�06 o1.7E�06 o3.2E�06 o1.2E�05 o2.7E�06 o3.0E�06 o5.3E�06 o1.5E�05 o1.3E�06 o1.8E�06 o2.9E�06 o1.0E�05

Tl g kg�1 fuel o9.5E�05 o6.4E�05 o8.7E�05 o3.6E�04 o2.9E�05 o2.9E�05 o5.4E�05 o2.0E�04 o4.6E�05 o5.0E�05 o8.9E�05 o2.6E�04 o2.2E�05 o3.1E�05 o4.8E�05 o1.7E�04

H.

Ag

raw

al

eta

l./

Atm

osp

heric

En

viron

men

t4

2(

20

08

)4

38

0–

43

92

4389


conjugers in their supplied cartridges and no dioxinor additional conjugers were found. The detectionlimits for all the dioxins analyzed is presented inTable 4.

3.6. Metals and ions

The distributions of metals for each enginesampled using XRF are provided in Fig. 6. It isnoted that plane 1 has a significant Cr reading and

Table 7

Sulfate (as sulfur) emission indices

Plane Units Mode 1

CFM56-7B22 EI (g kg�1) 1.86E�03

CFM56-3B1 EI (g kg�1) 3.22E�04

CFM56-3B2 EI (g kg�1) 6.88E�04

CFM56-7B22 EI (g kg�1) 2.94E�04

6.00E+03

5.00E+03

4.00E+03

3.00E+03

2.00E+03

1.00E+03

0.00E+00

μμg/m

3

Mod

e1

Mod

e2

Mod

e3

Mod

e4

Mod

e1

Mod

e2

Mod

e3°


DMM(integrated da

Fig. 7. Comparison of PM mass concentration (mgm�3) between

Table 6

Hexavalent chromium analysis

Sample Time (h) Mean (ppt) EI (g kg�1)

Plane1 4 118 1.65E�06

Plane2 4 47 o5.0E�07

Plane3 4 565 o4.18E�08

Plane4 4 7504 5.96E�05

Ambient-1 8 81 1.1E�06

Ambient-2 1 43 o3.2E�08

that plane 4 had extremely low Cr readings. Thevariability in the metal distribution was muchgreater between engines than between engine loads.The XRF analyses cannot distinguish the oxidationstate of the chromium. The EIs of different metalsare presented in Tables 5 and 6.

The Teflo filter was also analyzed for ionicspecies. It is noted that the mass of the ionscollected on the Teflo filter were so low that onlysulfate ions were above the detection limits of theinstrument. In case of the sulfate, the extracted ions(o1 ppm in water) were very close to the lowerdetection limit of the instrument. Table 7 sum-marizes the sulfate (as sulfur) EIs measured. Themeasurements of sulfur from XRF correlate wellwith the sulfur as sulfate analysis by IC confirmingthat most of the sulfur on the filter exists assulfate. The sulfur in the fuel used for the studywas 300 ppm. The sulfate EIs compare very wellwith the data presented by Petzold and Schroeder(1998), assuming same sulfur from fuel to sulfateconversion.

Mode 2 Mode 3 Mode 4

1.04E�03 1.29E�03 4.40E�03

3.81E�04 1.18E�03 5.00E�03

1.10E�03 2.02E�03 5.57E�03

3.16E�04 4.59E�04 7.12E�03

Mod

e4

Mod

e1

Mod

e2

Mod

e3

Mod

e4

Mod

e1

Mod

e2

Mod

e3

Mod

e4


ta, ug/m3) PM teflon filter data(ug/m3)

DMM and Teflon filter data. 1PM mass data not available.

ARTICLE IN PRESS

4.50E+02

4.00E+02

3.50E+02

3.00E+02

2.50E+02

2.00E+02

1.50E+02

1.00E+02

5.00E+01

0.00E+00

12:5

1:36

AM

1:26

:24

AM

1:55

:12

AM

2:24

:00

AM

2:52

:48

AM

3:21

:36

AM

3:50

:24

AM

4:19

:12

AM

µ µg/m

3

Engine Power up for 1 min

Engine on

Enginewarm up

65% load

85% load

85% load

Fig. 8. Realtime DMM mass concentrations in diluted exhaust for CFM56-7B22.


Results of the chromium analysis are shown inTable 6. The values for ambient air after 1 h ofsampling were below the method lower detectionlimit and those over an 8 h period were about theexpected ambient value. With respect to Cr(VI),results for the aircraft were very low as expected,except for one aircraft which showed a largeresponse for Cr(VI). This large response of Cr(VI)measured by SCAQMD method 205.1 for thissample was confirmed with subsequent analysiswith ICP-MS analysis. The high hexavalent chro-mium emission from this one plane could beassociated either with engine wear or samplecontamination.

3.7. PM comparison between DMM and Teflo filter

PM mass was also measured by Dektai massmonitor (DMM). DMM is a realtime instrumentdesigned for combustion particulate mass emissionmeasurement. The DMM data integrated over eachmode correlates with the PM mass measured onTeflon. A best fit comparison between the twomethods reveals a slope of 1.02 with an R2 of 0.70(Fig. 7). Realtime emission rates for PM (Fig. 8) fora typical aircraft engine reveals a dramatic PM

spikes during change in power settings and followsstepwise the power settings of the test. These trendsare consistent with the integrated filter massesproviding further confidence in the PM emissionrates reported.

4. Summary and conclusion

The first on-wing investigation for characteriza-tion of commercial aircraft emissions was success-fully conducted. The data acquired can be useful forsource profiling for chemical mass balance modelingor for hazardous risk assessment. The analysisshowed the emission rates for aircraft turbineengines are highly dependent on power. Hydro-carbon emissions are most significant at idlingconditions whereas PM emissions are significant atthe higher power levels associated with takeoff andlanding. The data demonstrated that the elementalcarbon portion of PM increases with increase inpower. It is noted that the emissions of dioxins werebelow method detection limits. Emissions of metalsdo not show any significant trends with changingload conditions, which implies that the metalsdistribution is dependent on the difference in thevarious engines.


Acknowledgments

We acknowledge California Air Resource Board(CARB) for funding the project. We thank ourpartners: Mr. Robert Howard and others at ArnoldEngineering Development Center (AEDC), Univer-sity of Missouri-Rolla (UMR), Aerodyne Research,Inc. (ARI), National Aeronautics and Space Admin-istration (NASA), Oakland International Airportand Southwest Airlines and South Coast Air QualityManagement District (SCAQMD). We acknowledgeMs. Kathalena Cocker, Anthony Turgman andVaralakshmi Jayaram for analyses support.

References

Anderson, B.E., Chen, G., Blake, D.R., 2006. Hydrocarbon

emissions from a modern commercial airliner. Atmospheric

Environment 40, 3601–3612.

CFM, 2006. CFM Fleet Statistics, October 1, 2005. Obtained from

the CFM website /http://www.cfm56.com/engines/stats.htmlS.Geyer, H.J., Schramm, K.W., Feicht, E.A., Behechti, A.,

Steinberg, C., Bruggemann, R., Poiger, H., Henkelmann, B.,

Kettrup, A., 2002. Half-lives of tetra-, penta-, hexa-, hepta-,

and octachlorodibenzo-p-dioxin in rats, monkeys, and

humans–a critical review. Chemosphere 48, 631–644.

Herndon, S.C., Todd Rogers, N., Edward J Dunlea, N., John T

Jayne, N., Richard Miake-Lye, N., Berk Knighton., N., 2006.

Hydrocarbon emissions from in-use commercial aircraft

during airport operations. Environmental Science and Tech-

nology 40, 4406–4413.

Knighton, W.B., Rogers, T.M., Anderson, B.E., Herndon, S.C.,

Yelvington, P.E., Miake-Lye, R.C., 2007. Quantification of

aircraft engine hydrocarbon emissions using proton transfer

reaction mass spectrometry. Journal of Propulsion and Power

23 (5), 949–958.

NIOSH (National Institute of Occupational Safety and Health),

1996. NIOSH Manual of Analytical Methods.

Petzold, A., Schroeder, F.P., 1998. Jet engine exhaust aerosol

characterization. Aerosol Science and Technology 28, 62–76.

Petzold, A., Strom, J., Schroeder, F.P., Karcher, B., 1999.

Carbonaceous aerosol in jet engine exhaust: emission

characteristics and implications for heterogeneous chemical

reactions. Atmospheric Environment 33, 2689–2698.

Popp, P.J., Bishop, G.A., Stedman, D.H., 1999. Method for

commercial aircraft nitric oxide emission measure-

ments. Environmental Science and Technology 33,

1542–1544.

Shah, S.D., Cocker, D.R., Miller, J.W., Norbeck, J.M., 2004a.

Emission rates of particulate matter and elemental and

organic carbon from in-use diesel engines. Environmental

Science and Technology 38 (9), 2544–2550.

Shah, S.D., Temitope, A., Ogunyoku, J., Wayne Miller, Cocker

III, David R., 2004b. On-road emission rates of PAH and n-

Alkane compounds from heavy-duty diesel vehicles. Environ-

mental Science and Technology 39 (14), 5276–5284.

Siegl, W.O., Richert, J.F.O., Jensen, T.E., Schuetzle, D., Swarin,

S.J., Loo, J.F., Prostak, A., Nagy, D., Schlenker, A.M., 1993.

Improved emissions speciation methodology for phase II of

the auto/oil air quality improvement research programs

hydrocarbons and oxygenates. SAE Technical Paper Series

(930142) Special Publication SP-1000.

Slemr, F., Giehl, H., Habram, M., Slemr, J., Schlager, H.,

Schulte, P., Haschberger, P., Lindermeir, E., Dopelheuer, A.,

Plohr, M., 2001. In-flight measurements of aircraft CO and

nonmethane hydrocarbon emission indices. Journal of Geo-

physical Research 106 (7), 485–7494.

Spicer, C.W., et al., 1984. Composition and Photochemical

Reactivity of Turbine Engine Exhaust, prepared by Battelle

Laboratories, prepared for the Air Force Engineering &

Services Center, March 1984.

Spicer, C.W., Holdren, M.W., Smith, D.L., Hughes, D.P., Smith,

M.D., 1992. Chemical composition of exhaust from aircraft

turbine engines. Journal of Engineering: Gas Turbines Power

114, 111–115.

Spicer, C.W., Holdren, M.W., Riggin, R.M., Lyon, T.F., 1994.

Chemical composition and photochemical reactivity of

exhaust from aircraft turbine engines. Annals of Geophysics

12, 12,944–12,955.

US EPA, September 1985. Compilation of Air Pollutant

Emission Factors, Volume II, Mobile Sources. AP-42 Aircraft

Data from February 1980, Ann Arbor, MI.

Wey, C.C., 2006. Aircraft particle emissions experiment (APEX).

NASA TM, 2006-214382.

Wey, C.C., Anderson, B.E., Wey, C., Miake-Lye, R.C., White-

field, P., Howard, R., 2007. Overview on the aircraft particle

emissions experiment. Journal of Propulsion and Power 23

(5), 898–905.

http://www.cfm56.com/engines/stats.html

characterization of chemical and particulate emissions from aircraft engines

Documents