12
Meteorologische Zeitschrift, Vol. 14, No. 4, 465-476 (August 2005) c by Gebrüder Borntraeger 2005 Article Particle emissions from aircraft engines – a survey of the European project PartEmis ANDREAS PETZOLD 1 , MARKUS FIEBIG 1 , LUTZ FRITZSCHE 1 , CLAUDIA STEIN 1 , ULRICH SCHUMANN 1 ,CHRIS W. WILSON 2 ,CHRIS D. HURLEY 2 ,FRANK ARNOLD 3 ,ELENI KATRAGKOU 3 , URS BALTENSPERGER 4 ,MARTIN GYSEL 4 ,STEPHAN NYEKI 4 ,REGINA HITZENBERGER 5 ,HEIN- RICH GIEBL 5 ,KEVIN J. HUGHES 6 ,RALF KURTENBACH 7 ,PETER WIESEN 7 ,PAUL MADDEN 8 ,HANS PUXBAUM 9 ,SUSI VRCHOTICKY 9 and CLAUS WAHL 10 1 DLR, Institute of Atmospheric Physics, Oberpfaffenhofen, Germany 2 QinetiQ, Farnborough, UK 3 Max Planck Institute for Nuclear Physics, Heidelberg, Germany 4 Paul Scherrer Institute, Villigen, Switzerland 5 University of Vienna, Vienna, Austria 6 University of Leeds, Leeds, UK 7 University of Wuppertal, Wuppertal, Germany 8 Rolls Royce plc, UK 9 Vienna University of Technology, Vienna, Austria 10 DLR, Institute for Combustion Technology, Cologne, Germany (Manuscript received January 20, 2004; in revised form September 3, 2004; accepted November 4, 2004) Abstract An overview of the goals and achievements of the European PartEmis project (Measurement and prediction of emissions of aerosols and gaseous precursors from gas turbine engines) is presented. PartEmis was focussed on the characterisation and quantication of exhaust emissions from a gas turbine engine. The engine was composed of a combustor and a unit to simulate a 3-shaft turbine section (so-called Hot End Simulator; HES). A comprehensive suite of aerosol, gas and chemi-ion measurements were conducted under different, i) com- bustor and HES operating conditions, ii) fuel sulphur concentrations. Measured aerosol properties were mass and number concentration, size distribution, mixing state, thermal stability of internally mixed particles, hy- groscopicity, cloud condensation nuclei (CCN) activation potential, and chemical composition. Furthermore, chemi-ions, non-methane volatile organic compounds (NMVOCs) and OH were monitored. The combustor operation conditions corresponded to modern and older engine gas path temperatures at cruise altitude, with fuel sulphur contents (FSC) of 0.05, 0.41, and 1.270 g kg 1 . The combustor behaved like a typical aircraft engine combustor with respect to thermodynamic data and main emissions, which suggests that the PartEmis database may be applicable to contemporary aircraft engines. The conclusions drawn from the PartEmis ex- periment are discussed separately for combustion particles, ultrane particles, sulphate-containing species and chemi-ions, particle hygrioscopic growth and CCN activation, gaseous organic fraction, and emission properties. Zusammenfassung Die in dem europäischen Projekt PartEmis (Messung und Modellierung der Emissionen von Partikeln und gasförmigen Vorläufern aus Gasturbinen) erzielten Ergebnisse werden vorgestellt. PartEmis konzentrierte sich auf die Charakterisierung und Quantizierung von Abgasemissionen aus einer Gasturbine. Das Triebwerk bestand aus einer Brennkammer und einer Einheit zur Simulierung einer 3-stugen Turbinensektion, dem so genannten Hot End Simuator. Ein umfassendes System zur Messung von Aerosolen, Gasen und Ionen wurde i) unter verschiedenen Betriebsbedingungen der Brennkammer und des Hot End Simulators, und ii) bei ver- schiedenen Schwefelgehalten im verwendeten Treibstoff eingesetzt. Die gemessenen Aerosoleigenschaften umfassten Massen- und Anzahlkonzentration, Größenverteilung, Mischungszustand, thermische Stabilität in- tern gemischter Partikel, Hygroskopizität, Aktivierbarkeit zur Bildung von Wolkentropfen, und chemische Zusammensetzung. Weiterhin wurden Chemi-Ionen, Nicht-Methan Kohlenwasserstoffe (NMVOC) und OH gemessen. Die Betriebsbedingungen der Brennkammer entsprachen den Gaseintrittstemperaturen alter und moderner Triebwerke in Reiseughöhe. Der Schwefelgehalt im Treibstoff variierte von 0,05 über 0,41 bis 1,270 g kg 1 . Die Brennkammer verhielt sich im Bezug auf thermodynamische Daten und Emissionsdaten wie die typische Brennkammer eines Flugzeugtriebwerks. Davon ausgehend lassen sich die erreichten Ergeb- nisse auf moderne Flugzeugtriebwerke übertragen. Die Schlussfolgerungen aus dem PartEmis Experiment werden getrennt für Verbrennungspartikel, ultrafeine Partikel, sulfathaltige Komponenten und Chemi-Ionen, Feuchtewachstum der Partikel und Aktivierbarkeit zu Wolkentropfen, gasförmige organische Verbindungen und Emissionseigenschaften der Brennkammer diskutiert. Corresponding author: Andreas Petzold, DLR Institute of Atmospheric Physics, Oberpfaffenhofen, 82234 Wessling, Germany, e-mail: [email protected] DOI: 10.1127/0941-2948/2005/0054 0941-2948/2005/0054 $ 5.40 c Gebrüder Borntraeger, Berlin, Stuttgart 2005

Particle emissions from aircraft engines a survey of the ... · 466 A. Petzold et al.: Particle emissions from aircraft engines Meteorol. Z., 14 , 2005 1 Introduction The role of

  • Upload
    others

  • View
    1

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Particle emissions from aircraft engines a survey of the ... · 466 A. Petzold et al.: Particle emissions from aircraft engines Meteorol. Z., 14 , 2005 1 Introduction The role of

Meteorologische Zeitschrift Vol 14 No 4 465-476 (August 2005)ccopy by Gebruumlder Borntraeger 2005 Article

Particle emissions from aircraft engines ndash a survey of theEuropean project PartEmis

ANDREAS PETZOLDlowast1 MARKUS FIEBIG1 LUTZ FRITZSCHE1 CLAUDIA STEIN1 ULRICHSCHUMANN1 CHRIS W WILSON2 CHRIS D HURLEY2 FRANK ARNOLD3 ELENI KATRAGKOU3URS BALTENSPERGER4 MARTIN GYSEL4 STEPHAN NYEKI4 REGINA HITZENBERGER5 HEIN-RICH GIEBL5 KEVIN J HUGHES6 RALF KURTENBACH7 PETER WIESEN7 PAUL MADDEN8 HANSPUXBAUM9 SUSI VRCHOTICKY9 and CLAUS WAHL10

1DLR Institute of Atmospheric Physics Oberpfaffenhofen Germany2QinetiQ Farnborough UK3Max Planck Institute for Nuclear Physics Heidelberg Germany4Paul Scherrer Institute Villigen Switzerland5University of Vienna Vienna Austria6University of Leeds Leeds UK7University of Wuppertal Wuppertal Germany8Rolls Royce plc UK9Vienna University of Technology Vienna Austria10DLR Institute for Combustion Technology Cologne Germany

(Manuscript received January 20 2004 in revised form September 3 2004 accepted November 4 2004)

AbstractAn overview of the goals and achievements of the European PartEmis project (Measurement and prediction ofemissions of aerosols and gaseous precursors from gas turbine engines) is presented PartEmis was focussedon the characterisation and quantification of exhaust emissions from a gas turbine engine The engine wascomposed of a combustor and a unit to simulate a 3-shaft turbine section (so-called Hot End Simulator HES)A comprehensive suite of aerosol gas and chemi-ion measurements were conducted under different i) com-bustor and HES operating conditions ii) fuel sulphur concentrations Measured aerosol properties were massand number concentration size distribution mixing state thermal stability of internally mixed particles hy-groscopicity cloud condensation nuclei (CCN) activation potential and chemical composition Furthermorechemi-ions non-methane volatile organic compounds (NMVOCs) and OH were monitored The combustoroperation conditions corresponded to modern and older engine gas path temperatures at cruise altitude withfuel sulphur contents (FSC) of 005 041 and 1270 g kgminus1 The combustor behaved like a typical aircraftengine combustor with respect to thermodynamic data and main emissions which suggests that the PartEmisdatabase may be applicable to contemporary aircraft engines The conclusions drawn from the PartEmis ex-periment are discussed separately for combustion particles ultrafine particles sulphate-containing speciesand chemi-ions particle hygrioscopic growth and CCN activation gaseous organic fraction and emissionproperties

ZusammenfassungDie in dem europaumlischen Projekt PartEmis (Messung und Modellierung der Emissionen von Partikeln undgasfoumlrmigen Vorlaumlufern aus Gasturbinen) erzielten Ergebnisse werden vorgestellt PartEmis konzentrierte sichauf die Charakterisierung und Quantifizierung von Abgasemissionen aus einer Gasturbine Das Triebwerkbestand aus einer Brennkammer und einer Einheit zur Simulierung einer 3-stufigen Turbinensektion dem sogenannten Hot End Simuator Ein umfassendes System zur Messung von Aerosolen Gasen und Ionen wurdei) unter verschiedenen Betriebsbedingungen der Brennkammer und des Hot End Simulators und ii) bei ver-schiedenen Schwefelgehalten im verwendeten Treibstoff eingesetzt Die gemessenen Aerosoleigenschaftenumfassten Massen- und Anzahlkonzentration Groumlszligenverteilung Mischungszustand thermische Stabilitaumlt in-tern gemischter Partikel Hygroskopizitaumlt Aktivierbarkeit zur Bildung von Wolkentropfen und chemischeZusammensetzung Weiterhin wurden Chemi-Ionen Nicht-Methan Kohlenwasserstoffe (NMVOC) und OHgemessen Die Betriebsbedingungen der Brennkammer entsprachen den Gaseintrittstemperaturen alter undmoderner Triebwerke in Reiseflughoumlhe Der Schwefelgehalt im Treibstoff variierte von 005 uumlber 041 bis1270 g kgminus1 Die Brennkammer verhielt sich im Bezug auf thermodynamische Daten und Emissionsdatenwie die typische Brennkammer eines Flugzeugtriebwerks Davon ausgehend lassen sich die erreichten Ergeb-nisse auf moderne Flugzeugtriebwerke uumlbertragen Die Schlussfolgerungen aus dem PartEmis Experimentwerden getrennt fuumlr Verbrennungspartikel ultrafeine Partikel sulfathaltige Komponenten und Chemi-IonenFeuchtewachstum der Partikel und Aktivierbarkeit zu Wolkentropfen gasfoumlrmige organische Verbindungenund Emissionseigenschaften der Brennkammer diskutiert

lowastCorresponding author Andreas Petzold DLR Institute of Atmospheric Physics Oberpfaffenhofen 82234Wessling Germany e-mail andreaspetzolddlrde

DOI 1011270941-294820050054

0941-294820050054 $ 540

ccopy Gebruumlder Borntraeger Berlin Stuttgart 2005

466 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

1 Introduction

The role of aviation-related particle emissions in the up-per troposphere and lowermost stratosphere is a matterof concern particularly with respect to a possible effecton the composition of upper tropospheric aerosol andthe life cycle of cirrus clouds (BOUCHER 1999 SCHU-MANN and STROumlM 2001) Despite recent considerableprogress in the characterisation of the aerosol particlesemitted from aircraft engines (eg ANDERSON et al1998 BROCK et al 2000 HAGEN et al 2001 PET-ZOLD and SCHROumlDER 1998 PETZOLD et al 1999SCHUMANN et al 2002) microphysical and chemicalproperties as well as the influence of engine operatingconditions still remain poorly characterised with respectto particulate emissions However this kind of data setforms a basic pre-requisite for modelling approacheswhich investigate the role of aviation aerosols in thephysics and chemistry of the upper free troposphere andlowermost stratosphere and their effect on earthrsquos cli-mate

In the framework of the European project PartEmis(ldquoMeasurement and prediction of emissions of aerosolsand gaseous precursors from gas turbine enginesrdquo) theinfluence of operation conditions and fuel sulphur con-tent (FSC) on the microphysical and chemical proper-ties of particles emitted from an aircraft engine sim-ulator was investigated For these purposes an enginesimulator which met the ICAO (International Civil Avi-ation Organisation) engine emissions smoke standardwas operated on a test rig at the QinetiQ (formerly De-fence Evaluation and Research Agency DERA) test siteat Pyestock UK This engine simulator consisted of areal aircraft engine combustor and a so-called Hot EndSimulator (HES) which simulates the pressure and tem-perature profiles found in an aircraft engine turbine sec-tion In a first experiment in February 2001 the emis-sion properties of the combustor were studied Thesedata were analysed and used as boundary conditions

Table 1 Combustor operating conditions during the 2001 Combus-tor and 2002 HES campaigns

Combustor Combustor Campaign HES CampaignParameters

Old Modern Old ModernT combustor 566 766 566 760inlet KT combustor 1125 1448 1125 1360outlet KP combustor 705 82 705 82inlet 105 PaAir mass flow 212 212 190 205kg sminus1Fuel flow 0032 0042 0030 0035kg sminus1Air fuel ratio 66 503 66 58

for a second experiment in MarchApril 2002 whichfocused on the emission properties of the combustor-HES combination Fig 1 shows a schematic of thecombustor-HES combination including the sample ex-traction points used during the combustor experimentsin 2001 and the HES experiments in 2002

The aerosol microphysical and chemical proper-ties investigated in this study covered number sizeand mass concentration of combustion aerosol particlesand of volatile condensation particles Furthermore thevolatile fraction of internally mixed combustion parti-cles particle hygroscopicity and cloud condensation nu-clei (CCN) activation potential were studied In addi-tion the emission of chemi-ions non-methane volatileorganic compounds (NMVOCs) and OHwas monitoredA detailed description of the PartEmis experiments isgiven by WILSON et al (2004)

2 Methods

A tubo-annular transpiration cooled combustor waschosen for tests on QinetiQrsquos High Pressure CombustionRig The combustor did not represent a particular type ofaircraft engine technology Although the combustor isbased on a 30-year old design it has been substantiallymodified with new cooling and fuel injection systemsaccording to current legislation and technological ad-vances such that it meets the emissions requirements ofmodern civil combustors In the actual engine itself tenof these combustors would be circumferentially spacedaround the engine core The exit nozzle used was a 36ldquosmilerdquo shaped discharge nozzle as employed in theengine Air was delivered from a four-stage compres-sor and heater unit For all investigated compounds andaerosols the concentration levels in the inlet air weredetermined and subtracted from the measured resultsif applicable A back-pressurising valve located down-stream of the combustor maintained the correct pressurein the system After extraction of sample exhaust at thecombustor exit the remainder of the high-temperatureexhaust gases were cooled using water injection via aset of spray-bars The standard test-rig instrumentationcovered control parameters such as temperature pres-sure fuel and air flow rates as well as CO CO2 totalhydrocarbons (HC) NOX and smoke number For moredetails on the technical set-up of the test rig and on op-eration conditions see WILSON et al (2004)

Two engine operating conditions were simulatedto represent the thermodynamic conditions found inthe turbine stages of legacy referred to here as ldquooldrdquoand modern aircraft cruise conditions at 35000 feet(sim 11700 m) This will ensure that the conversion offuel sulphur to oxidised sulphur species is replicatedThese old and modern combustor operating conditionswere also chosen to meet the ICAO smoke emission

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 467

Figure 1 Schematic of the PartEmis combustor and Hot End Simulator HES the HES is composed of three separate heat exchanger

stages referred to as high (hp) intermediate (ip) and low pressure (lp) stages

Table 2 Available equipment and parameters

Parameter TechniqueVolatile condensation particles Condensation Particle Size Analyser (CPSA)(number size) Differential Mobility Analyser (DMA) connected to a ThermodenuderNon-volatile carbonaceous particles Condensation Particle Counter (CPC)(number size) Scanning Mobility Particle Sizer (SMPS)

DMA connected to a ThermodenuderVolatility Tandem-DMA

Surface density Diffusion ChargerAerosol absorption coefficient babs Aerosol Absorption PhotometerSmoke Number SN Smoke number methodWater uptake at RH le 100 Hygroscopicity Tandem-DMACloud condensation nuclei activation Thermal gradient CCN counterChemical composition Berner low pressure impactor (BLPI) samples

filter stack samplesanalysis viagravimetry (total mass)multi-step combustion method (organicelementalcarbon)evolved gas analysis (volatility of carbonaceouscompounds)gas chromatography (aliphatic constituents)ion chromatography (major ions)

CO CO2 HC NOx O2 H2 Standard test rig instrumentationHONO HNO3 Chemical ionisation mass spectrometrySO2 SO3 H2SO4 Paul Ion Trap Mass Spectrometry (PITMAS)Chemi-ionsSO2Gaseous hydrocarbon species Quasi-on-line gas chromatography (GC)

Sampling and off-line analysisPartially oxidised hydrocarbons DNPH methodOH Laser-induced fluorescence (LIF)

level Fuel sulphur content (FSC) levels covered therange from low sulphur fuel (005 g kgminus1) to the con-temporary average (041 g kgminus1) and to a maximum ofabout three times this average (1270 g kgminus1) definedhere as Low Mid and High (L M H) FSC respectivelyThe sulphur content was varied by adding known quanti-ties of benzenethiol (C6H5SH) to the low FSC fuel Theoperation conditions are summarised in Table 1

The employed aerosol measurement methods con-sisted of various size-selective methods aerosol ab-sorption photometers thermodenuder methods a CloudCondensation Nuclei (CCN) Counter (HITZENBERGER

et al 2003a) an extensive set of chemical analyti-cal methods (PETZOLD et al 2003a PUXBAUM et al2003) mass spectrometry methods for the detection ofsulphur-containing species and laser-induced fluores-

cence for OH measurements A summary of deployedmethods is given in Table 2

The black carbon (BC) mass concentration and theaerosol absorption coefficient in the diluted sample gaswere measured using a Particle Soot Absorption Pho-tometer (PSAP Radiance Research Inc Seattle USA)The specific absorption cross section at λ = 055micromσabs = 67 m2 gminus1 was used for conversion of lightabsorption to BC mass The particle number concen-trations were measured in the size intervals D = 4minus7nm 7ndash9 nm 9ndash20 nm and 20ndash250 nm with parallel-operated condensation particle counters (CondensationParticle Size Analyser CPSA) which were set to differ-ent lower instrument cut-off diameters Particles of sizeD gt 9 nm which are detected in CPSA channels 3 and4 are in the following referred to as N10 Particle size

468 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

distributions were measured with a commercial Scan-ning Mobility Particle Sizer (SMPS TSI Model 3071)and a Volatility-Tandem Differential Mobility Analyser(V-TDMA) in the size range Dgt 13 nm

The aerosol mixing state and the volatile fractionof the combustion aerosol were determined using ther-mal denuder methods which differentiate particles bynumber according to their thermal stability As wasinvestigated in previous calibration experiments ther-mal denuder temperatures of 390 K and 573ndash625 Krespectively distinguish volatile (vaporises at 390 K)sulphuric-acid like material from semivolatile (vaporisesbetween 390 and 573 K) and refractory black-carbonlike material The thermal treatment of almost monodis-perse particle size fractions which were pre-selected bya DMA provides information whether the aerosol isinternally or externally mixed The ratio of diametersafter and before thermal treatment yields the particleshrinkage factor D32D

30 This factor describes the par-

ticle volume reduction by the thermal treatment ie thevolatile volume fraction of internally mixed particles is1minusD3D3

0The hygroscopic particle growth factor under sub-

saturated conditions was measured with a Hygroscop-icity Tandem DMA (GYSEL et al 2002 2003a) Dur-ing the measurements the instrument was kept at a con-stant temperature of T sim= 298 K The obtained hygro-scopic growth factor describes the particle growth bywater uptake from the gas phase at a selected relativehumidity Measurements were performed in the rela-tive humidity range from 70 to approximately 95 As was demonstrated by GYSEL at al (2002 2003a)from KOEHLER theory the hygroscopic growth factorof a particle is a function of dry particle size relativehumidity and particle coating with a water-soluble sub-stance For the PartEmis experimental data the coatingsubstance was sulphuric acid The magnitude of the par-ticle growth at a given relative humidity and for a givensize of the dry particle can be directly linked to volumefraction of water-soluble matter contained in the inves-tigated particles During the PartEmis experiments thehygroscopic growth factors measured for fixed dry par-ticle sizes but variable fuel sulphur contents were usedfor the determination of the coating of the combustionaerosol by sulphuric acid The number of particles whichbecome finally activated as cloud condensation nuclei inthe case of water-supersaturation were measured with astatic thermal diffusion chamber CCN counter (GIEBL

et al 2002) The instrument was operated at a super-saturation ratio of 06ndash07 with respect to liquid waterThe activation ratio or fraction of CCN activated par-ticles respectively was determined as the ratio of thenumber of activated CCN to the number of total particles(HITZENBERGER et al 2003a) The ability of combus-tion particles for taking up water from the gas phase is a

key property in the formation processes for liquid waterclouds However it is also of relevance for jet conden-sation trail formation (SCHUMANN 1996) and it mayas well play a role in the formation of ice nuclei fromcombustion particles (DEMOTT et al 1999)

Aerosol samples were collected with filter stack sam-plers and low pressure Berner impactors for total aerosoland size-resolved chemical mass analysis (PUXBAUM etal 2003) The chemical composition of the carbona-ceous fraction of the exhaust particles was determinedusing multi-step combustion methods Total carbon (TC)was determined by a combustion technique involvingcombustion in a vertical furnace and detection of CO2with a commercial NDIR (non dispersive Infrared spec-trometer) unit Aliquots of a collected filter sample wereburnt in the ceramic tube of a combustion furnace withan excess of oxygen at a temperature of 1000C Ele-mental carbon (EC) was determined by the ldquoCachierrdquomethod (CACHIER et al 1989) ie another aliquot ofthe filter sample was treated for 2 h at 320C in an oxy-gen atmosphere to remove organic carbon compoundsand then analysed with the same TC analysis methodThe remaining carbon was attributed to elemental car-bon (EC) Thermograms of the collected carbonaceousmaterial were measured with a thermo-optical evolvedgas analysis method (SCHMID et al 2001) To deter-mine anions eg SO2minus

4 as well as cations two Dionexisocratic systems with electrochemical suppression wereused

In addition the aerosol was sampled on quartz tubescontaining glass wool The aerosol loaded quartz tubeswere analysed by 5 min thermodesorption in a He-lium atmosphere at 130C and 350C respectively Thedesorbed C2ndashC12 non-methane hydrocarbons (NMHC)were analysed with a gas chromatograph (GC HewlettPackard) NMHCs were monitored using a compact GCinstrument (Airmovoc 2010) with a cryotrap as an en-richment system and a flame ionisation detector C2ndashC10 aliphatic and aromatic NMHCswere measured witha time resolution of 10 min and with detection limits inthe pptv range Selected partially oxidised hydrocarbons(OHC) such as carbonyl compounds and organic acidswere collected by using commercial sampling cartridgescoated with 24-Dinitrophenylhydrazine This chemicalcompound forms with the carbonyl molecules a stablecompound called hydrazone This hydrazone is stablefor weeks and can be analysed later in the LaboratoryTo analyse the carbonyl compounds the hydrazones arerinsed out of the cartridge and separated in a High Per-formance Liquid Chromatograph HPLC The OHCweremeasured with a time resolution of 4 min and with de-tection limits in the ppbv range All hydrocarbon com-pound measurements were corrected for line losses andbackground values

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 469

Figure 2 Left size distribution of combustion aerosol with (diamonds) and without condensation particle formation (triangles) RightExternal aerosol mixture composed of volatile condensation particles (grey) and coated nonvolatile combustion particles (black)

Gaseous sulphuric acid was measured using a CIMS(Chemical Ionisation Mass Spectrometer) apparatusequipped with an Ion Trap Mass Spectrometer (ITMS)The CIMS apparatus comprises a stainless steel flow re-actor (inner diameter 4 cm total length 164 cm) heatedup to 80C a high frequency glow discharge ion sourcefor reagent ion production and a Quadrupole Ion TrapMass Spectrometer for ion detection The ITMS allowsCID (Collision-Induced-Dissociation) investigations ofmass selected ions to be performed which in turn allowsmuch better species identification than can be offeredby a conventional mass spectrometer Reactant ions ofthe type NO3 (HNO3)n (n= 012) are produced in aseparate chamber where a high frequency gas dischargeburns in a buffer gas of O2 and NO2 N2 and injectedinto the flow reactor upstream of the ion sampling orificeof the Mass Spectrometer After injection ions are car-ried by the gas stream to the mass spectrometer Whengaseous sulphuric acid is present in the flow reactorsome fraction of the reagent ions reacts via a gas phaseion molecule reaction with sulphuric acid By measur-ing the abundance ratio of product and reagent ions thegaseous sulphuric acid concentration in the flow reactoris determined For more details see KATRAGKOU et al(2004) and references given there

OH laser induced fluorescence (LIF) measurementswere performed in the exhaust duct of the HES The ex-periment comprised a NdYAG pumped dye laser set upto generate light at λ = 281913 nm in order to excite theQ1(1) transition of OH A2Σ(ν prime = 1)larr X2Π(ν primeprime = 0)chosen as it is one of the strongest transitions at the tem-perature expected at the observation point Fluorescencewas detected perpendicular to the path of the laser fromthe ν prime = 0rarr vprimeprime = 0 transition around 309 nm via a lensand a customised interference filter In order to max-imise the signal to noise ratio by discriminating againstRaman scatter expected from N2 and H2O an interfer-ence filter was centred at 30875 nm with a band passof 496 nm A gated CCD camera cooled to ndash30C was

used to collect the signal allowing signal accumulationover long times in order to enhance sensitivity

Since the full set of applied instruments was con-nected to separate sampling lines according to specificinstrument requirements the precise determination ofthe dilution factors was a crucial task After exit fromthe traversing probe the sample was cooled to sim 150Cusing a water-jacket and delivered to the undiluted anddiluted sample lines The undiluted line was insulatedalong its entire length in order to avoid cooling andtherefore wall losses of less volatile combustion prod-ucts A detailed description of the extracting probe andthe sample transportation system is given in WILSON

et al (2004) Dilution with filtered air was necessaryfor the majority of instruments to reduce concentrationsdown to a measurable level The sample dilution factorwas determined from CO2 mixing ratios measured in thesupplied dilution air diluted sample and undiluted sam-ple as well as from temperature and pressure sensorsthroughout the sample delivery system The average di-lution factor was 65 (PETZOLD et al 2003a)

Although stable gaseous species such as CO2 andwater vapour are unaffected by intrusive samplingaerosols may be lost due to diffusion and sedimenta-tion to the surfaces of the sampling system Estimatesof expected line losses were obtained from experiments(HURLEY 1996) performed with aerosol generated bya spark discharge generator (PALAS GfG 1000) Theaerosol was fed into sample lines of different cross sec-tion and length and the sample flow through the lineswas varied From these studies expected line losses arelt 20 by number of particles for the conditions dur-ing the PartEmis experiments Since line losses affectpredominantly small particles lt 20 nm in diameter theaerosol mass is assumed to be almost unaffected by linelosses However the values given for emission indiceshave to be taken as upper values Size distributions re-ported for combustion particles cover a size range of 10ndash250 nm The size spectra used for the investigation of the

470 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

impact of operation conditions and fuels sulphur con-tent were not corrected for line losses because correc-tions are similar for all test points Thus relative trendsconcerning the influence of fuel sulphur and operationconditions on the microphysical properties of the emit-ted particles are not affected by line losses Becauseof the observed size range of exhaust particles countmedian diameters and geometric standard deviations ofmeasured size distributions are also expected to be onlymarginally influenced by line losses

3 Results

31 Microphysical properties

The aerosol emitted from the Combustor is composedof primary particles forming inside the combustor andof volatile condensation particles nucleating in the cool-ing exhaust gas from gaseous precursors The aerosolsize distribution in the diluted exhaust gas of the com-bustor is shown in Fig 2 for conditions with and with-out condensation particle formation While the mode ofcombustion aerosol particles dominates the size rangeD gt 15 nm with a maximum at 40 nm another modeof nanoparticles is present in all cases in the size rangeD= 7minus20 nm Condensation particle formation is oc-curring only at high FSC conditions and contributes tothe size range D= 4minus7 nm

The emission properties with respect to the primaryaerosol are shown in terms of BC mass (EIBC ) TCmass (EITC) and number (EIN10) emissions per kg ofconsumed fuel in Fig 3 The number emission indexEIN10 refers to particles with diameters D ge 10 nm InFig 3 also the count median diameter (CMD) of theexhaust aerosol size distribution is shown Plotted arecorresponding values for the combustor emissions andemissions from the HES low pressure stage In the firstplace it has to be mentioned that all the results are con-sistent At old cruise conditions an increasing numberemission at constant black carbon mass emission is bal-anced by a decreasing particle size with respect to theparticle electrical mobility diameter It should be notedthat these results originate from different measurementtechniques

When operated at old engine conditions with lowcombustor inlet temperature TC EIBC varies from 0051to 0081 g kgminus1 Operated at modern engine condi-tions with the higher inlet temperature EIBC ranges from0028 to 0044 g kgminus1 which is a factor of two lowerSimilar relations hold for the other properties How-ever the amount of emitted particulate matter is withinthe range permitted by the ICAO rules in all cases Allparticle emission indices referring to number and massare in close agreement with values obtained from in-flight measurements which span a range from 05ndash075

Figure 3 Emission indices in terms of number (N10 ) total car-

bon mass (TC) black carbon mass (BC) and count median diameter

(CMD) of the combustion aerosol for the Combustor exit (triangles)

and the Hot End Simulator low pressure stage (diamonds) values are

given for old and modern operation conditions as a function of the

fuel sulphur content error bars indicate signal variability in terms of

plusmn1 stdev

g kgminus1 for a Boeing B707 to 001 g kgminus1 for AirbusA 340 and Boeing B737 equipped with CFM 56 se-ries engines (SCHUMANN et al 2002 PETZOLD et al2003a 2003b) The fleet-averaged mass emission indexof 0038 g kgminus1 (PETZOLD et al 1999) lies within therange of values determined for modern operation condi-tions

Mean CMD values are 37ndash41 nm for old engine and40ndash44 nm for modern engine conditions The geomet-ric standard deviation of the size distribution is 167(16ndash18) for both operation conditions (PETZOLD etal 2003a 2003b) The combustor operating conditionshave only a weak influence on the size number and massof emitted particles Size distribution peaking at similardiameters were observed during several in-flight stud-ies HAGEN et al (1996 1998) reported peak diame-ters of 30ndash60 nm and size distribution widths of approx16 During the SULFUR studies (PETZOLD et al 1999SCHUMANN et al 2002) peak diameters of 35 nm wereobserved The PartEmis size distribution parameters arein close agreement with the reported in-flight observa-tions

The influence of the different HES pressure stages isdemonstrated in Fig 3 and with more detail in Fig 4As is shown in Fig 3 the trends observed in particleemission properties in the combustor experiments withvarying fuel sulphur content and operation conditionsagain observed during the HES experiments Except fora step-like increase from the combustor experiments tothe HES experiments which is most likely due to slightlydifferent operating conditions the HES has no influenceon the emission properties of the Combustor-HES com-bination with respect to combustion aerosol particles

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 471

0

5

10

15

20

25

30

000

005

010

015

020

high intermed low pressure

Combustor Hot End Simulator section

EIBC

EIN

EI

BC

g

kg

-1

EI

N

10

14kg

-1

Figure 4 Effect of the Hot End Simulator on the emission propertiesof the Combustor-HES combination operated at modern conditions

error bars indicate signal variability in terms of plusmn1 stdev

see Fig 4 Increasing the FSC slightly increases the to-tal number of emitted particles

Particles smaller than 10 nm contribute less than10 to the total exhaust aerosol in the low and mediumFSC cases At high FSC levels the formation of volatilecondensation particles with diameters Dlt 10 nm is ob-served (PETZOLD et al 2003c) A kinetic model wasused to investigate the nucleation efficiency of the H2O- H2SO4 binary mixture in the exhaust gas and the par-ticle coagulation processes during the transport throughthe sampling system to the measurement instrumenta-tion (VANCASSEL et al 2004) The model confirmedthat the condensation particles form at the dilutor of thesampling system During combustion the sulphur con-tained in the fuel is oxidised to SO2 which is then par-tially converted to S(VI) in the hot exhaust gas The ef-ficiency of conversion from S(IV) to S(VI) is describedby the conversion efficiency calculated as ε =S(VI)STwhere ST is the total fuel sulphur In the combustor ex-haust the S(VI) increases with increasing FSC In thekinetic model study a value for the efficiency was indi-rectly deduced from comparisons between model resultsand measurements In the PartEmis cases ε ranges be-tween roughly 25 and 6 depending on the com-bustor settings and on the microphysical approach usedFor these values which are close to results from di-rect measurements of gaseous sulphuric acid the kineticmodel reproduces the observation of volatile particles inthe size range 4ndash7 nm for the high FSC case while novolatile particles are found by the model for the low andmedium FSC cases

Particles of size D ge 30 nm are almost entirely in-ternally mixed Smaller particles (D lt 20 nm) exhibitvolume fractions of 10ndash15 for volatile and 4ndash10 for semi-volatile material while the volatile fraction de-creases to lt 5 for D ge 50 nm particles (PETZOLD

et al 2003a NYEKI et al 2003) The decrease of thevolatile particle fraction with increasing particle size isalso reproduced by the kinetic model for the H2O ndashH2SO4 binary mixture (VANCASSEL et al 2004) Thesevalues are valid for the conditions met during PartEmisIf the dilution air is more humid and a higher dilutionforces stronger formation of condensation particles thecondensation particles may grow to larger sizes How-ever as was observed in a young aircraft plume at cruiseconditions condensation particles were also observedmainly in the size range Dlt 10 nm (SCHROumlDER et al2000) similar to the PartEmis results

32 Chemical properties

Primary particles emitted from the combustor arepredominantly composed of carbonaceous material(gt95 ) and of water-soluble inorganic componentsThe carbonaceous material is composed of organic com-pounds and of elemental or black carbon The organiccompounds evaporate and oxidise at different tempera-tures and become classified according to this criterionCompounds which evaporate at temperatures as low as400 K are classified as semi-volatie OC compoundsevaporating compounds evaporating up to 800 K aretermed nonvolatile OC and the remainder is elementalcarbon Fig 5 shows the thermal stability of an exhaustaerosol sampled during operation with low FSC fuel atmodern conditions In the applied version of evolved gasanalysis (SCHMID et al 2001) the blackness of the filtersample is recorded parallel to the heating of the sampleAs is demonstrated in Fig 5 the filter starts to becomeldquowhiterdquo again as soon as the elemental or black carbonstarts to oxidise The evaporation and subsequent oxi-dation of OC compounds leaves the filter black Thusthe onset of increasing filter transmittance is an addi-tional measure for separating organic and elemental car-bon The application of this methods yields an averagecontribution of black carbon to total carbon of 40ndash80 depending on the combustion conditions (PUXBAUM etal 2003) From an aerosol loaded glass wool samplemore than 180 different NMHCs were desorbed underan Helium atmosphere Only 04 wt of the NMHCswere desorbed at T le 400 K whereas 996 wt weredesorbed at T = 400ndash620 K This is in good agreementwith the result of the thermogram in Fig 5 Elemental(EC) or black carbon (BC) respectively is chemicallyalmost inert at low and moderate temperatures and ther-mally stable up to temperatures of about 700 K

As determined from direct measurements of gaseousH2SO4 the efficiency of conversion in the combus-tor exhaust ranges between 09 plusmn 05 and 13 plusmn07 (KATRAGKOU et al 2004) In the HES ex-haust S(VI) increases with increasing FSC fuel flowand from the high-pressure to the low-pressure HESstage For FSC=1270 g kgminus1 the conversion efficiency

472 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

is 23plusmn 12 at the low pressure stage for the mod-ern cruise condition and 14plusmn 07 for the old cruiseThe higher ε observed for the modern cruise conditionssuggests that modern engines have a larger ε comparedto old engines Conversion efficiencies determined fromin-flight data at cruise altitude gave values 04ndash45 (SCHUMANN et al 2002) which is in good agreementwith the PartEmis ground test values

Chemical composition information was obtainedfrom Berner low pressure impactor samples which weretaken at a sample line temperature of 150C (PUXBAUMet al 2003) At these conditions most of the sulphuricacid remains in the gas phase while only a small fractionis more strongly bound to or chemi-sorbed at the par-ticle surface and will be observed in particulate mattersamples The chemical analysis results indicate that themass distribution of the carbonaceous matter matchesthe volume distribution of the aerosol The mass dis-tribution of chemisorbed sulphate however follows thesurface distribution of the aerosol which was calculatedfrom aerosol size spectra assuming a spherical particleshape (PUXBAUM et al 2003) This observation indi-cates that the sulphate is chemisorbed at the particle sur-face The chemisorbed fraction of sulphuric acid corre-sponds to a coverage of max 01 monolayer in the highFSC case This fraction of converted S(VI) is less than01 of total S(IV) present in the fuel The overall par-titioning of sulphur-containing species at 150C is 97 SO2le 27 gaseous H2SO4lt 03 chemisorbedAfter cooling to environmental temperatures the gasphase sulphuric acid can condense and cover the com-bustion particles with about one monolayer of sulphuricacid on average (HITZENBRGER et al 2003a)

33 Particle hygroscopic properties

Particle hygroscopic properties were measured at ambi-ent conditions using a Hygroscopicity Tandem Differ-ential Mobility Analyser (H-TDMA) system (GYSEL etal 2003a 2003b) The instrument measures the hygro-scopic growth factor (g) of particles with initial dry sizesDo = 30 50 and 100 nm over the range RHsim 70minus95 where g(RH) = D(RH)Do Growth factors were found toincrease with increasing FSC at fixed particle size (egDo = 50 nm g (95 ) = 101 and 116 for FSC = lowand high respectively) and to decrease with increasingparticle size at fixed FSC (eg mid FSC g (95 ) =114 110 and 103 for Do = 30 50 and 100 nm respec-tively) The results suggest an increasing amount of sul-phuric acid adsorbed on combustion particles from thegas phase with increasing FSC

Cloud condensation nuclei (CCN) were measured inthe diluted line using a static thermal diffusion typeCCN counter (GIEBL et al 2002) Particles are exposedto controlled supersaturations of water vapour whichenables some of the particles to become activated and

Figure 5 Evolving CO2 as a function of the sample heating temper-ature for a Combustor aerosol sample semi-volatile OC oxidises at

450 K non-volatile OC at 600 K and elemental carbon above 700 K

grow to large (gt 10 microm) droplets During PartEmis thesupersaturation was set tosim 07 The fraction of CCNin the aerosol (relative to the concentration of particlesof Dgt 10 nm) increases with increasing FSC but evenfor low FSC the fraction of CCN (at 0755 supersat-uration) is higher than the fraction of particles with di-ameters larger than 280 nm ie the size at which wet-table insoluble particles are activated according to theKelvin effect At high FSC the fraction of CCN in-creases by about a factor 55 for old and 12 for mod-ern conditions However pure sulphuric acid particlesof the same size which become activated according toKOumlHLER theory would give an activation ratio which isstill larger by a factorsim 15 compared to the observationseven at high FSC Thus the KOumlHLER model cannot re-produce measured activation properties This is in agree-ment with the fact that large combustion particles are notat all completely water soluble The Kelvin model pro-vides good predictions for the low FSC level but acti-vation diameters (ratios) are increasingly overestimated(underestimated) at medium and high FSC This indi-cates that the soluble fraction of the combustion parti-cles increases from low values at low FSC level to rel-atively large values at high FSC An empirical coatedsphere model which equates the measured volatile andsemivolatile fraction with soluble sulphuric acid under-estimates the activation diameter by 45 to 55 underall conditions A small amount of semivolatile organ-ics with low or absent water-solubility adsorbed on thecombustion particles might be one reason for this dis-crepancy Further details are given by HITZENBERGER

et al (2003a 2003b)

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 473

34 Gaseous hydrocarbon emission

More than 100 different non-methane volatile organiccompounds (NMVOCs) such as non-methane hydrocar-bons (NMHCs) and selected partially oxidised hydro-carbons (OHCs) emitted from the aircraft engine com-bustor and the hot end simulator (HES) were identifiedand quantified These species accounted for up to 91wt of the total hydrocarbons (THCs) The NMHCemission indices of 5ndash15 microg kgminus1 in total are compa-rable with the AEROTRACE (Measurement of TraceSpecies in the Exhaust from Aero Engines EC projectAERA-CT94-0003 1994ndash1997 Co-ordinated by Qine-tiQ) study Both studies show up to 1000 times loweremission indices in comparison with emissions from ve-hicular traffic The influence of the operation conditionson NMHC emissions of the combustor and HES was in-vestigated for the identified NMHCs From the six con-ditions which were investigated (2 operation conditions3 HES pressure stages) most compounds show a de-crease of NMHC emissions with increasing combustiontemperature at the same pressure stage This is in agree-ment with results from the combustor exit measurementsand with the results of the AEROTRACE study Emis-sion indices at HES pressure stages high (HP)- inter-mediate (IP)- and low pressure (LP) for fixed FSC andcombustor operation conditions were also investigatedMost compounds show a decrease of NMHC emissionswith decreasing pressure in the HES at similar combus-tor operation conditions and fixed FSC The decrease ofthe pressure in the HES corresponds to an increase ofthe residence time in a turbine and indicates a decreaseof NMHC emissions along the different turbine sections

35 Emission of ions and charged sootparticles

The ions measured during the HES campaign(HAVERKAMP et al 2004 WILHELM et al 2004) areproduct ions formed from primary ions via ion moleculereactions involving either few reaction-steps withmassive neutral molecules or numerous reaction-stepswith relatively small neutral molecules Their lifetimeswith respect to ion-ion recombination are around 9 msafter the high pressure turbine section The ion emissionindex Ei is up to 45 times 1016 ions per kg fuel burntAs is discussed by HAVERKAMP et al (2004) neitherion-ion recombination nor ion removal by attachmentto combustion particles give sufficient explanations forthe observed ion emission index Therefore detailedmodel calculations of the ion chemistry taking place inthe combustor with particular emphasis on the sulphurspecies effect is required (HAVERKAMP et al 2004)It seems that ion-induced nucleation can explain atleast partly the observed volatile aerosols Electricallycharged small positive and negative soot particles

(CSP) have also been detected during both PartEmiscampaigns Charged soot particles are found to havediameters around 6 nm and a total concentration of upto 45times107 cmminus3 (positive and negative) correspondingto a total emission index of ECSP = 23times 1015 CSP perkg fuel burnt (WILHELM et al 2004)

A model which considers the formation and evolu-tion of combustion ions in a combustor of an aircraftengine in dependence on the electron detachment ef-ficiency from negative ions was used to consider theeffect of the transformation of primary negative ionsto more stable secondary negative ions (SOROKIN etal 2003) The formed stable negative ions most prob-ably are sulphur-bearing ions This effect slows downthe charged particle neutralisation rate leading to an in-crease of the concentration of positive and negative ionsat the combustor exit The results of the simulation andtheir comparison with the ground-based experimentaldata obtained within PartEmis support the hypothesisthat the increase of the fuel sulphur content leads to anincrease of the ion concentration at the combustor exit

36 OH measurements

The concentration of OH which is the dominant oxidantin the exhaust gas was detected by laser induced fluo-rescence (LIF) to obtain an absolute OH concentrationmeasurement in order to aid in the modelling of SO3production from the combustor to the engine exit Whenrunning the laser at low laser energies OH in the exhaustduct was successfully detected Calibration experimentsprovide a method to assign the observed OH LIF sig-nals an absolute concentration and from an estimate ormeasurement of the errors for each parameter an overallestimate of the error in the assigned OH concentrationwas made From these calibration experiments very lowvalues of less than 1 ppbv were calculated for the OHconcentration in the exhaust gas after exiting the HESPrevious experiments using a combined Raman scatter-ingLIF technique to look in the exhaust gases 50 cmbehind an aircraft engine exit failed to detect any OHLIF signal and therefore could only infer an upper limitfor OH of 80 ppbv based on the measured detection limitof their apparatus

4 Summary and conclusions

The progress made during PartEmis is manifold Themain conclusions are therefore grouped into differentsubject areas and presented as brief statements For eachsubject area key references are identified if available

41 Combustion aerosol

Refer to WILSON et al (2004) and PETZOLD et al(2003a)

474 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

(i) The combustor behaves like a typical aircraft en-gine combustor with respect to thermodynamic dataand main emissions It may thus be expected that thePartEmis database can be applied as a first order approx-imation to contemporary aircraft engines

(ii) The BC concentration is nearly independent ofFSC level

(iii) The number size distribution of primary particlesin the diameter range 10 lt Dlt 550 nm is almost unin-fluenced by the FSC level or combustor conditions Athigh FSC levels the formation of volatile condensationparticles with diameters Dlt 10 nm occurs

(iv) The volatile volume fraction of particles gt 50 nmincreases from 2 for low FSC to approx 5 for highFSC The volatile volume fraction of particles lt 30 nmis le 15

(v) Volatility measurements indicate an internallymixed aerosol for particles with D gt 15 nm conden-sation particles are below 15 nm in diameter

42 Ultrafine particles

Refer to PETZOLD et al (2003a) and VANCASSEL et al(2004)

(vi) Formation of condensation particles of size D=4minus 7 nm is observed in the Combustor tests for highFSC at old and modern conditions formation occurs ifthe surface area of the combustion aerosol falls below acertain threshold limit the observed number of formedparticles is confirmed by kinetic model studies

(vii) Formation of condensation particles is observedin the HES tests for high FSCmodern conditionslowpressure section only the ldquoformation strengthrdquo in termsof number of condensation particles per number of com-bustion particle is twice as high as for the Combustorcase

(viii) The model studies indicate that condensationparticles are not emitted from the combustor but form inthe sampling line at the sample dilution point

(ix) Ultrafine carbonaceous particles of size 7 nmlt D lt 20 nm are observed independently of the fuelsulphur content a volatility analysis indicates that theseparticles are nonvolatile and may be composed of car-bonaceous compounds

43 Sulphate-containing species chemi-ionsand OH

Refer to KATRAGKOU et al (2004) and HAVERKAMP etal (2004)

(x) S(VI) was measured in the form of sulphuric acidIncreased FSC results in higher S(VI) concentrationsThe measured conversion efficiency for fuel sulphur toS(IV) at the combustor exit varies from 035ndash14 andis independent of the investigated combustor operationconditions and FSC levels

(xi) 10ndash30 of sulphate is chemisorbed on primaryparticles the coverage of primary particles is below 01monolayers of sulphuric acid at 150C and approxi-mately one monolayer at 25C

(xii) Sulphur partitioning at 150C 97 SO2le 27 gaseous H2SO4 lt 03 chemisorbed

(xiii) Positive and negative gaseous ions were de-tected in the exhaust of a combustor with mass numbersup to 1500 The total ion emission index at the HES hpstage is as high as 54times 1017 positive and negative ionsper kg fuel burnt Large ions with mass numbers exceed-ing 4500 contribute about 10 to the total ion numberconcentration

(xiv) OH concentration in the exhaust gas after exit-ing the HES is lt 1 ppbv

44 Hygroscopic growth and CCN activation

Refer to GYSEL et al (2003a) and HITZENBERGER etal (2003a)

(xv) Small differences in particle hygroscopicity areobserved between old and modern cruise conditions

(xvi) Strong effect of the fuel sulphur content is ob-served particle hygroscopicity increases distinctly withincreasing FSC

(xvii) No significant effect of the HES at low andmid FSC level is observed compared to the combustorexit an increased particle hygroscopicity downstreamthe HES compared to the combustor is observed at highFSC indicating continuing transformation of S(IV) toS(VI) during the transfer through the HES

(xviii) CCN activation of combustion particles in-creases with increasing FSC as a result of larger par-ticle hygroscopicity additional effects of adsorbed or-ganic matter cannot be excluded

45 Gaseous organic fraction

(xix) More than 100 different NMVOCs (aliphatic andaromatic hydrocarbons carbonyls and acids) were iden-tified and quantified they account for 91 wt of thetotal detected compounds

(xx) A decrease in NMHC emissions from old tomodern conditions at fixed HES pressure stage and fuelsulphur content (FSC) is found

(xxi) A decrease of NMHC emissions with decreas-ing pressure in the HES at fixed combustor operationconditions and FSC is observed

(xxii) No clear influence on the NMHC emissionswith changing FSC at the same combustor operationcondition is observed

(xxiii) High EIrsquos for organic acids are found Pos-sible explanations are that organic compounds fixed atthe sampling line walls are oxidised by the hot exhaustgases or that the acids are fixed at the surface of thesoot particles which are also trapped in the samplingcartridges

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 475

46 Emission properties

Refer to WILSON et al (2004) and PETZOLD et al(2003a)

(xxiv) Emission properties regarding combustionaerosol mass number and size are determined by thecombustor the influence of the Hot End Simulator isweak

(xxv) The processing of exhaust gas and parti-cles during the transfer through the Hot End Simula-tor increases the particle hygroscopicity and formationstrength of condensation particles both effects are re-lated to gaseous sulphuric acid

The experimental overview of the EU PartEmisproject presents the first comprehensive aerosol and gasmeasurements to be conducted on an aircraft enginecombustor test-rig The database will be used for fur-ther analyses as well as for plume modelling studies ofthe effect of aviation-related particle emissions on theglobal atmosphere and climate

Acknowledgements

The PartEmis project was funded by the European Com-mission under contract no G4RD-CT-2000-00207 Theauthors are very grateful to the test rig operation crew atQinetiQ for their very strong support during the experi-ments

References

ANDERSON BE WR COFER DR BAGWELL JWBARRICK CH HUDGINS KE BRUNKE 1998 Air-borne observations of aircraft aerosol emissions I Totalnonvolatile particle emission indices ndash Geophys Res Lett25 1689ndash1692

BOUCHER O 1999 Influence of air traffic on cirrus occur-rence Nature ndash 397 30ndash31

BROCK CA F SCHROumlDER B KAumlRCHER A PETZOLDR BUSEN M FIEBIG 2000 Ultrafine particle size distri-butions measured in aircraft exhaust plumes ndash J GeophysRes 105 26555ndash26567

CACHIER H MP BREMOND P BIAT-MEacuteNARD 1989Determination of atmospheric soot carbon with a simplethermal method ndash Tellus 41B 379ndash390

DEMOTT PJ Y CHEN SM KREIDENWEIS DCROGERS DE SHERMAN 1999 Ice formation by blackcarbon particles ndash Geophys Res Lett 26 2429ndash2432

GIEBL H A BERNER G REISCHL H PUXBAUM AKASPER-GIEBL R HITZENBERGER 2002 CCN activa-tion of oxalic and malonic acid test aerosols with the Uni-versity of Vienna cloud condensation nuclei counter ndash JAerosol Sci 33 1623ndash1634

GYSEL M E WEINGARTNER U BALTENSPERGER2002 Hygroscopicity of aerosol particles at low tempera-tures 2 Theoretical and experimental hygroscopic proper-ties of laboratory generated aerosols ndash Environ Sci Tech-nol 36 63ndash68

GYSEL M S NYEKI E WEINGARTNER U BAL-TENSPERGER H GIEBL R HITZENBERGER A PET-ZOLD CW WILSON 2003a Properties of jet engine com-bustor particles during the PartEmis experiment Hygro-scopicity at subsaturated conditions ndash Geophys Res Lett30 1566 DOI1010292003GL016896

GYSEL M S NYEKI E WEINGARTNER U BAL-TENSPERGER H GIEBL R HITZENBERGER A PET-ZOLD CW WILSON 2003b Jet engine combustion par-ticle hygroscopicity under subsaturated conditions duringPartEmis ndash J Aerosol Sci 34 (Suppl 2) Abstracts of theEuropean Aerosol Conference S1415ndashS1416

HAGEN DE PD WHITEFIELD H SCHLAGER 1996Particle emissions in the exhaust plume from commercialjet aircraft under cruise conditions ndash J Geophys Res 10119551ndash19557

HAGEN DE PD WHITEFIELD J PALADINO M TRUE-BLOOD H LILENFELD 1998 Particle sizing and emissionindices for a jet engine exhaust sampled at cruise ndash Geo-phys Res Lett 25 1681ndash1684

HAGEN DE PD WHITEFIELD JD PALADINO OSCHMID 2001 Comparative study of jet engine com-bustion emissions and engine operating conditions ndashIn SCHUMANN U G AMANATIDIS (Eds) Aviationaerosols contrails and cirrus clouds Air pollution researchreport 74 106ndash110 European Commission Brussels

HAVERKAMP H S WILHELM A SOROKIN F ARNOLD2004 Positive and negative ion measurements in jet aircraftengine exhaust concentrations sizes and implications foraerosol formation ndash Atmos Environ 38 2879ndash2884

HITZENBERGER R H GIEBL A PETZOLD M GYSELS NYEKI E WEINGARTNER U BALTENSPERGERCW WILSON 2003a Properties of jet engine combus-tor particles during the PartEmis experiment Hygroscopicproperties at supersaturated conditions ndash Geophys ResLett 30 1779 DOI1010292003GL017294

HITZENBERGER R H GIEBL A PETZOILD M GYSELS NYEKI E WEINGARTNER U BALTENSPERGERCW WILSON 2003b CCN activation of jet engine com-bustor particles during PartEmis ndash J Aerosol Sci 34(Suppl 2) Abstracts of the European Aerosol ConferenceS1409ndashS1410

HURLEY CD 1996 AEROTRACE Measurements of Par-ticulates from an Engine Combustor ndash Proceedings Impactof Aircraft Emissions upon the Atmosphere Vol 1 113 In-ternational Colloquium Paris October 1996

KATRAGKOU E S WILHELM F ARNOLD CW WIL-SON 2004 First gaseous Sulfur (VI) measurements in thesimulated internal flow of an aircraft gas turbine engineduring project PartEmis ndash Geophys Res Lett 31 2117DOI1010292003GL018231

NYEKI S M GYSEL E WEINGARTNER U BAL-TENSPERGER R HITZENBERGER A PETZOLD CWWILSON 2003 Volatile properties of jet engine combus-tor particles during PartEmis ndash J Aerosol Sci 34 (Suppl2) Abstracts of the European Aerosol Conference S1411ndashS1412

PETZOLD A FP SCHROumlDER 1998 Jet engine exhaustaerosol characterization ndash Aerosol Sci Technol 28 62ndash76

PETZOLD A A DOumlPELHEUER CA BROCK FPSCHROumlDER 1999 In situ observations and model calcula-

476 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

tions of black carbon emission by aircraft at cruise altitudendash J Geophys Res 104 22171ndash22181

PETZOLD A C STEIN S NYEKI M GYSEL EWEINGARTNER U BALTENSPERGER H GIEBL RHITZENBERGER A DOumlPELHEUER S VRCHOTICKY HPUXBAUM M JOHNSON CD HURLEY R MARSHCW WILSON 2003a Properties of jet engine combus-tor particles during the PartEmis experiment Microphysi-cal and chemical properties ndash Geophys Res Lett 30 1719DOI1010292003GL017283

PETZOLD A CW WILSON U BALTENSPERGERM FIEBIG L FRITZSCHE H GIEBL M GYSELR HITZENBERGER CD HURLEY S NYEKI HPUXBAUM U SCHUMANN C STEIN S VROCHTICKY2003b Particle emissions from aircraft engines ndash anoverview of the European project PartEmis ndash J AerosolSci 34 (Suppl 2) Abstracts of the European Aerosol Con-ference S1405ndashS1406

PETZOLD A AE VRTALA M FIEBIG F FRITZSCHEPH MIRABEL S NYEKI X VANCASSEL CW WIL-SON 2003c Emission of volatile and non-volatile ultra-fine particles from a combustion source during PartEmisndash J Aerosol Sci 34 (Suppl 2) Abstracts of the EuropeanAerosol Conference S1407ndashS1408

PUXBAUM H S VRCHOTICKY A PETZOLD R HITZEN-BERGER S NYEKI CW WILSON 2003 Chemical com-position of jet engine combustor particles during PartEmisndash J Aerosol Sci 34 (Suppl 2) Abstracts of the EuropeanAerosol Conference S1413ndashS1414

SCHMID H L LASKUS H-J ABRAHAM U BAL-TENSPERGER V LAVANCHY M BIZJAK P BURBA HCACHIER D CROW J CHOW T GNAUK A EVENHM TEN BRINK K-P GIESEN R HITZENBERGER CHUEGLIN W MAENHAUT C PIOA CARVALHO J-PPUTAUD D TOOM-SAUNTRY H PUXBAUM 2001 Re-sults of the ldquocarbon conferencerdquo international aerosol car-bon round robin test stage I ndash Atmos Environ 35 2111ndash2121

SCHROumlDER F C BROCK R BAUMANN A PETZOLD RBUSEN P SCHULTE M FIEBIG 2000 In-situ studies onvolatile jet exhaust particle emissions Impact of fuel sul-fur content and thermodynamic conditions on nuclei-modeaerosols ndash J Geophys Res 105 19941ndash19954

SCHUMANN U 1996 Contrail formation from aircraft ex-hausts ndash Meteorol Z NF 5 4ndash23

SCHUMANN U J STROumlM 2001 Aviation impact on atmo-spheric composition and climate ndash In European research inthe stratosphere 1996ndash2000 European Commission EUR19867 Brussels

SCHUMANN U F ARNOLD R BUSEN J CURTIUS BKAumlRCHER A KIENDLER A PETZOLD H SCHLAGERF SCHROumlDER K-H WOHLFROM 2002 Influence of fu-els sulphur on the composition of aircraft exhaust plumesThe experiments SULFUR 1-7 ndash J Geophys Res 107(1010292001JD000813) AAC 2-1 ndash AAC 2-27

SOROKIN A P MIRABEL F ARNOLD 2003 On the In-fluence of fuel fulfur induced stable negative ion formationon the total concentration of ions emitted by an aircraft gasturbine engine Comparison of model and experiment ndash At-mos Chem Phys 3 6001ndash6018

VANCASSEL X A SOROKIN P MIRABEL A PETZOLDCW WILSON 2004 Volatile particles formation duringPartEmis a modelling study ndash Atmos Chem Phys 4 439ndash447

WILHELM S H HAVERKAMP A SOROKIN F ARNOLD2004 Detection of very large ions in aircraft gas turbineengine combustor exhaustcharged small soot particles ndashAtmos Environ 38 4561ndash4569

WILSON CW A PETZOLD S NYEKI U SCHUMANNR ZELLNER 2004 Measurement and Prediction of Emis-sions of Aerosols and Gaseous Precursors from Gas Tur-bine Engines (PartEmis) An Overview ndash Aerosp SciTechnol 8 131ndash143

Page 2: Particle emissions from aircraft engines a survey of the ... · 466 A. Petzold et al.: Particle emissions from aircraft engines Meteorol. Z., 14 , 2005 1 Introduction The role of

466 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

1 Introduction

The role of aviation-related particle emissions in the up-per troposphere and lowermost stratosphere is a matterof concern particularly with respect to a possible effecton the composition of upper tropospheric aerosol andthe life cycle of cirrus clouds (BOUCHER 1999 SCHU-MANN and STROumlM 2001) Despite recent considerableprogress in the characterisation of the aerosol particlesemitted from aircraft engines (eg ANDERSON et al1998 BROCK et al 2000 HAGEN et al 2001 PET-ZOLD and SCHROumlDER 1998 PETZOLD et al 1999SCHUMANN et al 2002) microphysical and chemicalproperties as well as the influence of engine operatingconditions still remain poorly characterised with respectto particulate emissions However this kind of data setforms a basic pre-requisite for modelling approacheswhich investigate the role of aviation aerosols in thephysics and chemistry of the upper free troposphere andlowermost stratosphere and their effect on earthrsquos cli-mate

In the framework of the European project PartEmis(ldquoMeasurement and prediction of emissions of aerosolsand gaseous precursors from gas turbine enginesrdquo) theinfluence of operation conditions and fuel sulphur con-tent (FSC) on the microphysical and chemical proper-ties of particles emitted from an aircraft engine sim-ulator was investigated For these purposes an enginesimulator which met the ICAO (International Civil Avi-ation Organisation) engine emissions smoke standardwas operated on a test rig at the QinetiQ (formerly De-fence Evaluation and Research Agency DERA) test siteat Pyestock UK This engine simulator consisted of areal aircraft engine combustor and a so-called Hot EndSimulator (HES) which simulates the pressure and tem-perature profiles found in an aircraft engine turbine sec-tion In a first experiment in February 2001 the emis-sion properties of the combustor were studied Thesedata were analysed and used as boundary conditions

Table 1 Combustor operating conditions during the 2001 Combus-tor and 2002 HES campaigns

Combustor Combustor Campaign HES CampaignParameters

Old Modern Old ModernT combustor 566 766 566 760inlet KT combustor 1125 1448 1125 1360outlet KP combustor 705 82 705 82inlet 105 PaAir mass flow 212 212 190 205kg sminus1Fuel flow 0032 0042 0030 0035kg sminus1Air fuel ratio 66 503 66 58

for a second experiment in MarchApril 2002 whichfocused on the emission properties of the combustor-HES combination Fig 1 shows a schematic of thecombustor-HES combination including the sample ex-traction points used during the combustor experimentsin 2001 and the HES experiments in 2002

The aerosol microphysical and chemical proper-ties investigated in this study covered number sizeand mass concentration of combustion aerosol particlesand of volatile condensation particles Furthermore thevolatile fraction of internally mixed combustion parti-cles particle hygroscopicity and cloud condensation nu-clei (CCN) activation potential were studied In addi-tion the emission of chemi-ions non-methane volatileorganic compounds (NMVOCs) and OHwas monitoredA detailed description of the PartEmis experiments isgiven by WILSON et al (2004)

2 Methods

A tubo-annular transpiration cooled combustor waschosen for tests on QinetiQrsquos High Pressure CombustionRig The combustor did not represent a particular type ofaircraft engine technology Although the combustor isbased on a 30-year old design it has been substantiallymodified with new cooling and fuel injection systemsaccording to current legislation and technological ad-vances such that it meets the emissions requirements ofmodern civil combustors In the actual engine itself tenof these combustors would be circumferentially spacedaround the engine core The exit nozzle used was a 36ldquosmilerdquo shaped discharge nozzle as employed in theengine Air was delivered from a four-stage compres-sor and heater unit For all investigated compounds andaerosols the concentration levels in the inlet air weredetermined and subtracted from the measured resultsif applicable A back-pressurising valve located down-stream of the combustor maintained the correct pressurein the system After extraction of sample exhaust at thecombustor exit the remainder of the high-temperatureexhaust gases were cooled using water injection via aset of spray-bars The standard test-rig instrumentationcovered control parameters such as temperature pres-sure fuel and air flow rates as well as CO CO2 totalhydrocarbons (HC) NOX and smoke number For moredetails on the technical set-up of the test rig and on op-eration conditions see WILSON et al (2004)

Two engine operating conditions were simulatedto represent the thermodynamic conditions found inthe turbine stages of legacy referred to here as ldquooldrdquoand modern aircraft cruise conditions at 35000 feet(sim 11700 m) This will ensure that the conversion offuel sulphur to oxidised sulphur species is replicatedThese old and modern combustor operating conditionswere also chosen to meet the ICAO smoke emission

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 467

Figure 1 Schematic of the PartEmis combustor and Hot End Simulator HES the HES is composed of three separate heat exchanger

stages referred to as high (hp) intermediate (ip) and low pressure (lp) stages

Table 2 Available equipment and parameters

Parameter TechniqueVolatile condensation particles Condensation Particle Size Analyser (CPSA)(number size) Differential Mobility Analyser (DMA) connected to a ThermodenuderNon-volatile carbonaceous particles Condensation Particle Counter (CPC)(number size) Scanning Mobility Particle Sizer (SMPS)

DMA connected to a ThermodenuderVolatility Tandem-DMA

Surface density Diffusion ChargerAerosol absorption coefficient babs Aerosol Absorption PhotometerSmoke Number SN Smoke number methodWater uptake at RH le 100 Hygroscopicity Tandem-DMACloud condensation nuclei activation Thermal gradient CCN counterChemical composition Berner low pressure impactor (BLPI) samples

filter stack samplesanalysis viagravimetry (total mass)multi-step combustion method (organicelementalcarbon)evolved gas analysis (volatility of carbonaceouscompounds)gas chromatography (aliphatic constituents)ion chromatography (major ions)

CO CO2 HC NOx O2 H2 Standard test rig instrumentationHONO HNO3 Chemical ionisation mass spectrometrySO2 SO3 H2SO4 Paul Ion Trap Mass Spectrometry (PITMAS)Chemi-ionsSO2Gaseous hydrocarbon species Quasi-on-line gas chromatography (GC)

Sampling and off-line analysisPartially oxidised hydrocarbons DNPH methodOH Laser-induced fluorescence (LIF)

level Fuel sulphur content (FSC) levels covered therange from low sulphur fuel (005 g kgminus1) to the con-temporary average (041 g kgminus1) and to a maximum ofabout three times this average (1270 g kgminus1) definedhere as Low Mid and High (L M H) FSC respectivelyThe sulphur content was varied by adding known quanti-ties of benzenethiol (C6H5SH) to the low FSC fuel Theoperation conditions are summarised in Table 1

The employed aerosol measurement methods con-sisted of various size-selective methods aerosol ab-sorption photometers thermodenuder methods a CloudCondensation Nuclei (CCN) Counter (HITZENBERGER

et al 2003a) an extensive set of chemical analyti-cal methods (PETZOLD et al 2003a PUXBAUM et al2003) mass spectrometry methods for the detection ofsulphur-containing species and laser-induced fluores-

cence for OH measurements A summary of deployedmethods is given in Table 2

The black carbon (BC) mass concentration and theaerosol absorption coefficient in the diluted sample gaswere measured using a Particle Soot Absorption Pho-tometer (PSAP Radiance Research Inc Seattle USA)The specific absorption cross section at λ = 055micromσabs = 67 m2 gminus1 was used for conversion of lightabsorption to BC mass The particle number concen-trations were measured in the size intervals D = 4minus7nm 7ndash9 nm 9ndash20 nm and 20ndash250 nm with parallel-operated condensation particle counters (CondensationParticle Size Analyser CPSA) which were set to differ-ent lower instrument cut-off diameters Particles of sizeD gt 9 nm which are detected in CPSA channels 3 and4 are in the following referred to as N10 Particle size

468 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

distributions were measured with a commercial Scan-ning Mobility Particle Sizer (SMPS TSI Model 3071)and a Volatility-Tandem Differential Mobility Analyser(V-TDMA) in the size range Dgt 13 nm

The aerosol mixing state and the volatile fractionof the combustion aerosol were determined using ther-mal denuder methods which differentiate particles bynumber according to their thermal stability As wasinvestigated in previous calibration experiments ther-mal denuder temperatures of 390 K and 573ndash625 Krespectively distinguish volatile (vaporises at 390 K)sulphuric-acid like material from semivolatile (vaporisesbetween 390 and 573 K) and refractory black-carbonlike material The thermal treatment of almost monodis-perse particle size fractions which were pre-selected bya DMA provides information whether the aerosol isinternally or externally mixed The ratio of diametersafter and before thermal treatment yields the particleshrinkage factor D32D

30 This factor describes the par-

ticle volume reduction by the thermal treatment ie thevolatile volume fraction of internally mixed particles is1minusD3D3

0The hygroscopic particle growth factor under sub-

saturated conditions was measured with a Hygroscop-icity Tandem DMA (GYSEL et al 2002 2003a) Dur-ing the measurements the instrument was kept at a con-stant temperature of T sim= 298 K The obtained hygro-scopic growth factor describes the particle growth bywater uptake from the gas phase at a selected relativehumidity Measurements were performed in the rela-tive humidity range from 70 to approximately 95 As was demonstrated by GYSEL at al (2002 2003a)from KOEHLER theory the hygroscopic growth factorof a particle is a function of dry particle size relativehumidity and particle coating with a water-soluble sub-stance For the PartEmis experimental data the coatingsubstance was sulphuric acid The magnitude of the par-ticle growth at a given relative humidity and for a givensize of the dry particle can be directly linked to volumefraction of water-soluble matter contained in the inves-tigated particles During the PartEmis experiments thehygroscopic growth factors measured for fixed dry par-ticle sizes but variable fuel sulphur contents were usedfor the determination of the coating of the combustionaerosol by sulphuric acid The number of particles whichbecome finally activated as cloud condensation nuclei inthe case of water-supersaturation were measured with astatic thermal diffusion chamber CCN counter (GIEBL

et al 2002) The instrument was operated at a super-saturation ratio of 06ndash07 with respect to liquid waterThe activation ratio or fraction of CCN activated par-ticles respectively was determined as the ratio of thenumber of activated CCN to the number of total particles(HITZENBERGER et al 2003a) The ability of combus-tion particles for taking up water from the gas phase is a

key property in the formation processes for liquid waterclouds However it is also of relevance for jet conden-sation trail formation (SCHUMANN 1996) and it mayas well play a role in the formation of ice nuclei fromcombustion particles (DEMOTT et al 1999)

Aerosol samples were collected with filter stack sam-plers and low pressure Berner impactors for total aerosoland size-resolved chemical mass analysis (PUXBAUM etal 2003) The chemical composition of the carbona-ceous fraction of the exhaust particles was determinedusing multi-step combustion methods Total carbon (TC)was determined by a combustion technique involvingcombustion in a vertical furnace and detection of CO2with a commercial NDIR (non dispersive Infrared spec-trometer) unit Aliquots of a collected filter sample wereburnt in the ceramic tube of a combustion furnace withan excess of oxygen at a temperature of 1000C Ele-mental carbon (EC) was determined by the ldquoCachierrdquomethod (CACHIER et al 1989) ie another aliquot ofthe filter sample was treated for 2 h at 320C in an oxy-gen atmosphere to remove organic carbon compoundsand then analysed with the same TC analysis methodThe remaining carbon was attributed to elemental car-bon (EC) Thermograms of the collected carbonaceousmaterial were measured with a thermo-optical evolvedgas analysis method (SCHMID et al 2001) To deter-mine anions eg SO2minus

4 as well as cations two Dionexisocratic systems with electrochemical suppression wereused

In addition the aerosol was sampled on quartz tubescontaining glass wool The aerosol loaded quartz tubeswere analysed by 5 min thermodesorption in a He-lium atmosphere at 130C and 350C respectively Thedesorbed C2ndashC12 non-methane hydrocarbons (NMHC)were analysed with a gas chromatograph (GC HewlettPackard) NMHCs were monitored using a compact GCinstrument (Airmovoc 2010) with a cryotrap as an en-richment system and a flame ionisation detector C2ndashC10 aliphatic and aromatic NMHCswere measured witha time resolution of 10 min and with detection limits inthe pptv range Selected partially oxidised hydrocarbons(OHC) such as carbonyl compounds and organic acidswere collected by using commercial sampling cartridgescoated with 24-Dinitrophenylhydrazine This chemicalcompound forms with the carbonyl molecules a stablecompound called hydrazone This hydrazone is stablefor weeks and can be analysed later in the LaboratoryTo analyse the carbonyl compounds the hydrazones arerinsed out of the cartridge and separated in a High Per-formance Liquid Chromatograph HPLC The OHCweremeasured with a time resolution of 4 min and with de-tection limits in the ppbv range All hydrocarbon com-pound measurements were corrected for line losses andbackground values

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 469

Figure 2 Left size distribution of combustion aerosol with (diamonds) and without condensation particle formation (triangles) RightExternal aerosol mixture composed of volatile condensation particles (grey) and coated nonvolatile combustion particles (black)

Gaseous sulphuric acid was measured using a CIMS(Chemical Ionisation Mass Spectrometer) apparatusequipped with an Ion Trap Mass Spectrometer (ITMS)The CIMS apparatus comprises a stainless steel flow re-actor (inner diameter 4 cm total length 164 cm) heatedup to 80C a high frequency glow discharge ion sourcefor reagent ion production and a Quadrupole Ion TrapMass Spectrometer for ion detection The ITMS allowsCID (Collision-Induced-Dissociation) investigations ofmass selected ions to be performed which in turn allowsmuch better species identification than can be offeredby a conventional mass spectrometer Reactant ions ofthe type NO3 (HNO3)n (n= 012) are produced in aseparate chamber where a high frequency gas dischargeburns in a buffer gas of O2 and NO2 N2 and injectedinto the flow reactor upstream of the ion sampling orificeof the Mass Spectrometer After injection ions are car-ried by the gas stream to the mass spectrometer Whengaseous sulphuric acid is present in the flow reactorsome fraction of the reagent ions reacts via a gas phaseion molecule reaction with sulphuric acid By measur-ing the abundance ratio of product and reagent ions thegaseous sulphuric acid concentration in the flow reactoris determined For more details see KATRAGKOU et al(2004) and references given there

OH laser induced fluorescence (LIF) measurementswere performed in the exhaust duct of the HES The ex-periment comprised a NdYAG pumped dye laser set upto generate light at λ = 281913 nm in order to excite theQ1(1) transition of OH A2Σ(ν prime = 1)larr X2Π(ν primeprime = 0)chosen as it is one of the strongest transitions at the tem-perature expected at the observation point Fluorescencewas detected perpendicular to the path of the laser fromthe ν prime = 0rarr vprimeprime = 0 transition around 309 nm via a lensand a customised interference filter In order to max-imise the signal to noise ratio by discriminating againstRaman scatter expected from N2 and H2O an interfer-ence filter was centred at 30875 nm with a band passof 496 nm A gated CCD camera cooled to ndash30C was

used to collect the signal allowing signal accumulationover long times in order to enhance sensitivity

Since the full set of applied instruments was con-nected to separate sampling lines according to specificinstrument requirements the precise determination ofthe dilution factors was a crucial task After exit fromthe traversing probe the sample was cooled to sim 150Cusing a water-jacket and delivered to the undiluted anddiluted sample lines The undiluted line was insulatedalong its entire length in order to avoid cooling andtherefore wall losses of less volatile combustion prod-ucts A detailed description of the extracting probe andthe sample transportation system is given in WILSON

et al (2004) Dilution with filtered air was necessaryfor the majority of instruments to reduce concentrationsdown to a measurable level The sample dilution factorwas determined from CO2 mixing ratios measured in thesupplied dilution air diluted sample and undiluted sam-ple as well as from temperature and pressure sensorsthroughout the sample delivery system The average di-lution factor was 65 (PETZOLD et al 2003a)

Although stable gaseous species such as CO2 andwater vapour are unaffected by intrusive samplingaerosols may be lost due to diffusion and sedimenta-tion to the surfaces of the sampling system Estimatesof expected line losses were obtained from experiments(HURLEY 1996) performed with aerosol generated bya spark discharge generator (PALAS GfG 1000) Theaerosol was fed into sample lines of different cross sec-tion and length and the sample flow through the lineswas varied From these studies expected line losses arelt 20 by number of particles for the conditions dur-ing the PartEmis experiments Since line losses affectpredominantly small particles lt 20 nm in diameter theaerosol mass is assumed to be almost unaffected by linelosses However the values given for emission indiceshave to be taken as upper values Size distributions re-ported for combustion particles cover a size range of 10ndash250 nm The size spectra used for the investigation of the

470 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

impact of operation conditions and fuels sulphur con-tent were not corrected for line losses because correc-tions are similar for all test points Thus relative trendsconcerning the influence of fuel sulphur and operationconditions on the microphysical properties of the emit-ted particles are not affected by line losses Becauseof the observed size range of exhaust particles countmedian diameters and geometric standard deviations ofmeasured size distributions are also expected to be onlymarginally influenced by line losses

3 Results

31 Microphysical properties

The aerosol emitted from the Combustor is composedof primary particles forming inside the combustor andof volatile condensation particles nucleating in the cool-ing exhaust gas from gaseous precursors The aerosolsize distribution in the diluted exhaust gas of the com-bustor is shown in Fig 2 for conditions with and with-out condensation particle formation While the mode ofcombustion aerosol particles dominates the size rangeD gt 15 nm with a maximum at 40 nm another modeof nanoparticles is present in all cases in the size rangeD= 7minus20 nm Condensation particle formation is oc-curring only at high FSC conditions and contributes tothe size range D= 4minus7 nm

The emission properties with respect to the primaryaerosol are shown in terms of BC mass (EIBC ) TCmass (EITC) and number (EIN10) emissions per kg ofconsumed fuel in Fig 3 The number emission indexEIN10 refers to particles with diameters D ge 10 nm InFig 3 also the count median diameter (CMD) of theexhaust aerosol size distribution is shown Plotted arecorresponding values for the combustor emissions andemissions from the HES low pressure stage In the firstplace it has to be mentioned that all the results are con-sistent At old cruise conditions an increasing numberemission at constant black carbon mass emission is bal-anced by a decreasing particle size with respect to theparticle electrical mobility diameter It should be notedthat these results originate from different measurementtechniques

When operated at old engine conditions with lowcombustor inlet temperature TC EIBC varies from 0051to 0081 g kgminus1 Operated at modern engine condi-tions with the higher inlet temperature EIBC ranges from0028 to 0044 g kgminus1 which is a factor of two lowerSimilar relations hold for the other properties How-ever the amount of emitted particulate matter is withinthe range permitted by the ICAO rules in all cases Allparticle emission indices referring to number and massare in close agreement with values obtained from in-flight measurements which span a range from 05ndash075

Figure 3 Emission indices in terms of number (N10 ) total car-

bon mass (TC) black carbon mass (BC) and count median diameter

(CMD) of the combustion aerosol for the Combustor exit (triangles)

and the Hot End Simulator low pressure stage (diamonds) values are

given for old and modern operation conditions as a function of the

fuel sulphur content error bars indicate signal variability in terms of

plusmn1 stdev

g kgminus1 for a Boeing B707 to 001 g kgminus1 for AirbusA 340 and Boeing B737 equipped with CFM 56 se-ries engines (SCHUMANN et al 2002 PETZOLD et al2003a 2003b) The fleet-averaged mass emission indexof 0038 g kgminus1 (PETZOLD et al 1999) lies within therange of values determined for modern operation condi-tions

Mean CMD values are 37ndash41 nm for old engine and40ndash44 nm for modern engine conditions The geomet-ric standard deviation of the size distribution is 167(16ndash18) for both operation conditions (PETZOLD etal 2003a 2003b) The combustor operating conditionshave only a weak influence on the size number and massof emitted particles Size distribution peaking at similardiameters were observed during several in-flight stud-ies HAGEN et al (1996 1998) reported peak diame-ters of 30ndash60 nm and size distribution widths of approx16 During the SULFUR studies (PETZOLD et al 1999SCHUMANN et al 2002) peak diameters of 35 nm wereobserved The PartEmis size distribution parameters arein close agreement with the reported in-flight observa-tions

The influence of the different HES pressure stages isdemonstrated in Fig 3 and with more detail in Fig 4As is shown in Fig 3 the trends observed in particleemission properties in the combustor experiments withvarying fuel sulphur content and operation conditionsagain observed during the HES experiments Except fora step-like increase from the combustor experiments tothe HES experiments which is most likely due to slightlydifferent operating conditions the HES has no influenceon the emission properties of the Combustor-HES com-bination with respect to combustion aerosol particles

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 471

0

5

10

15

20

25

30

000

005

010

015

020

high intermed low pressure

Combustor Hot End Simulator section

EIBC

EIN

EI

BC

g

kg

-1

EI

N

10

14kg

-1

Figure 4 Effect of the Hot End Simulator on the emission propertiesof the Combustor-HES combination operated at modern conditions

error bars indicate signal variability in terms of plusmn1 stdev

see Fig 4 Increasing the FSC slightly increases the to-tal number of emitted particles

Particles smaller than 10 nm contribute less than10 to the total exhaust aerosol in the low and mediumFSC cases At high FSC levels the formation of volatilecondensation particles with diameters Dlt 10 nm is ob-served (PETZOLD et al 2003c) A kinetic model wasused to investigate the nucleation efficiency of the H2O- H2SO4 binary mixture in the exhaust gas and the par-ticle coagulation processes during the transport throughthe sampling system to the measurement instrumenta-tion (VANCASSEL et al 2004) The model confirmedthat the condensation particles form at the dilutor of thesampling system During combustion the sulphur con-tained in the fuel is oxidised to SO2 which is then par-tially converted to S(VI) in the hot exhaust gas The ef-ficiency of conversion from S(IV) to S(VI) is describedby the conversion efficiency calculated as ε =S(VI)STwhere ST is the total fuel sulphur In the combustor ex-haust the S(VI) increases with increasing FSC In thekinetic model study a value for the efficiency was indi-rectly deduced from comparisons between model resultsand measurements In the PartEmis cases ε ranges be-tween roughly 25 and 6 depending on the com-bustor settings and on the microphysical approach usedFor these values which are close to results from di-rect measurements of gaseous sulphuric acid the kineticmodel reproduces the observation of volatile particles inthe size range 4ndash7 nm for the high FSC case while novolatile particles are found by the model for the low andmedium FSC cases

Particles of size D ge 30 nm are almost entirely in-ternally mixed Smaller particles (D lt 20 nm) exhibitvolume fractions of 10ndash15 for volatile and 4ndash10 for semi-volatile material while the volatile fraction de-creases to lt 5 for D ge 50 nm particles (PETZOLD

et al 2003a NYEKI et al 2003) The decrease of thevolatile particle fraction with increasing particle size isalso reproduced by the kinetic model for the H2O ndashH2SO4 binary mixture (VANCASSEL et al 2004) Thesevalues are valid for the conditions met during PartEmisIf the dilution air is more humid and a higher dilutionforces stronger formation of condensation particles thecondensation particles may grow to larger sizes How-ever as was observed in a young aircraft plume at cruiseconditions condensation particles were also observedmainly in the size range Dlt 10 nm (SCHROumlDER et al2000) similar to the PartEmis results

32 Chemical properties

Primary particles emitted from the combustor arepredominantly composed of carbonaceous material(gt95 ) and of water-soluble inorganic componentsThe carbonaceous material is composed of organic com-pounds and of elemental or black carbon The organiccompounds evaporate and oxidise at different tempera-tures and become classified according to this criterionCompounds which evaporate at temperatures as low as400 K are classified as semi-volatie OC compoundsevaporating compounds evaporating up to 800 K aretermed nonvolatile OC and the remainder is elementalcarbon Fig 5 shows the thermal stability of an exhaustaerosol sampled during operation with low FSC fuel atmodern conditions In the applied version of evolved gasanalysis (SCHMID et al 2001) the blackness of the filtersample is recorded parallel to the heating of the sampleAs is demonstrated in Fig 5 the filter starts to becomeldquowhiterdquo again as soon as the elemental or black carbonstarts to oxidise The evaporation and subsequent oxi-dation of OC compounds leaves the filter black Thusthe onset of increasing filter transmittance is an addi-tional measure for separating organic and elemental car-bon The application of this methods yields an averagecontribution of black carbon to total carbon of 40ndash80 depending on the combustion conditions (PUXBAUM etal 2003) From an aerosol loaded glass wool samplemore than 180 different NMHCs were desorbed underan Helium atmosphere Only 04 wt of the NMHCswere desorbed at T le 400 K whereas 996 wt weredesorbed at T = 400ndash620 K This is in good agreementwith the result of the thermogram in Fig 5 Elemental(EC) or black carbon (BC) respectively is chemicallyalmost inert at low and moderate temperatures and ther-mally stable up to temperatures of about 700 K

As determined from direct measurements of gaseousH2SO4 the efficiency of conversion in the combus-tor exhaust ranges between 09 plusmn 05 and 13 plusmn07 (KATRAGKOU et al 2004) In the HES ex-haust S(VI) increases with increasing FSC fuel flowand from the high-pressure to the low-pressure HESstage For FSC=1270 g kgminus1 the conversion efficiency

472 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

is 23plusmn 12 at the low pressure stage for the mod-ern cruise condition and 14plusmn 07 for the old cruiseThe higher ε observed for the modern cruise conditionssuggests that modern engines have a larger ε comparedto old engines Conversion efficiencies determined fromin-flight data at cruise altitude gave values 04ndash45 (SCHUMANN et al 2002) which is in good agreementwith the PartEmis ground test values

Chemical composition information was obtainedfrom Berner low pressure impactor samples which weretaken at a sample line temperature of 150C (PUXBAUMet al 2003) At these conditions most of the sulphuricacid remains in the gas phase while only a small fractionis more strongly bound to or chemi-sorbed at the par-ticle surface and will be observed in particulate mattersamples The chemical analysis results indicate that themass distribution of the carbonaceous matter matchesthe volume distribution of the aerosol The mass dis-tribution of chemisorbed sulphate however follows thesurface distribution of the aerosol which was calculatedfrom aerosol size spectra assuming a spherical particleshape (PUXBAUM et al 2003) This observation indi-cates that the sulphate is chemisorbed at the particle sur-face The chemisorbed fraction of sulphuric acid corre-sponds to a coverage of max 01 monolayer in the highFSC case This fraction of converted S(VI) is less than01 of total S(IV) present in the fuel The overall par-titioning of sulphur-containing species at 150C is 97 SO2le 27 gaseous H2SO4lt 03 chemisorbedAfter cooling to environmental temperatures the gasphase sulphuric acid can condense and cover the com-bustion particles with about one monolayer of sulphuricacid on average (HITZENBRGER et al 2003a)

33 Particle hygroscopic properties

Particle hygroscopic properties were measured at ambi-ent conditions using a Hygroscopicity Tandem Differ-ential Mobility Analyser (H-TDMA) system (GYSEL etal 2003a 2003b) The instrument measures the hygro-scopic growth factor (g) of particles with initial dry sizesDo = 30 50 and 100 nm over the range RHsim 70minus95 where g(RH) = D(RH)Do Growth factors were found toincrease with increasing FSC at fixed particle size (egDo = 50 nm g (95 ) = 101 and 116 for FSC = lowand high respectively) and to decrease with increasingparticle size at fixed FSC (eg mid FSC g (95 ) =114 110 and 103 for Do = 30 50 and 100 nm respec-tively) The results suggest an increasing amount of sul-phuric acid adsorbed on combustion particles from thegas phase with increasing FSC

Cloud condensation nuclei (CCN) were measured inthe diluted line using a static thermal diffusion typeCCN counter (GIEBL et al 2002) Particles are exposedto controlled supersaturations of water vapour whichenables some of the particles to become activated and

Figure 5 Evolving CO2 as a function of the sample heating temper-ature for a Combustor aerosol sample semi-volatile OC oxidises at

450 K non-volatile OC at 600 K and elemental carbon above 700 K

grow to large (gt 10 microm) droplets During PartEmis thesupersaturation was set tosim 07 The fraction of CCNin the aerosol (relative to the concentration of particlesof Dgt 10 nm) increases with increasing FSC but evenfor low FSC the fraction of CCN (at 0755 supersat-uration) is higher than the fraction of particles with di-ameters larger than 280 nm ie the size at which wet-table insoluble particles are activated according to theKelvin effect At high FSC the fraction of CCN in-creases by about a factor 55 for old and 12 for mod-ern conditions However pure sulphuric acid particlesof the same size which become activated according toKOumlHLER theory would give an activation ratio which isstill larger by a factorsim 15 compared to the observationseven at high FSC Thus the KOumlHLER model cannot re-produce measured activation properties This is in agree-ment with the fact that large combustion particles are notat all completely water soluble The Kelvin model pro-vides good predictions for the low FSC level but acti-vation diameters (ratios) are increasingly overestimated(underestimated) at medium and high FSC This indi-cates that the soluble fraction of the combustion parti-cles increases from low values at low FSC level to rel-atively large values at high FSC An empirical coatedsphere model which equates the measured volatile andsemivolatile fraction with soluble sulphuric acid under-estimates the activation diameter by 45 to 55 underall conditions A small amount of semivolatile organ-ics with low or absent water-solubility adsorbed on thecombustion particles might be one reason for this dis-crepancy Further details are given by HITZENBERGER

et al (2003a 2003b)

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 473

34 Gaseous hydrocarbon emission

More than 100 different non-methane volatile organiccompounds (NMVOCs) such as non-methane hydrocar-bons (NMHCs) and selected partially oxidised hydro-carbons (OHCs) emitted from the aircraft engine com-bustor and the hot end simulator (HES) were identifiedand quantified These species accounted for up to 91wt of the total hydrocarbons (THCs) The NMHCemission indices of 5ndash15 microg kgminus1 in total are compa-rable with the AEROTRACE (Measurement of TraceSpecies in the Exhaust from Aero Engines EC projectAERA-CT94-0003 1994ndash1997 Co-ordinated by Qine-tiQ) study Both studies show up to 1000 times loweremission indices in comparison with emissions from ve-hicular traffic The influence of the operation conditionson NMHC emissions of the combustor and HES was in-vestigated for the identified NMHCs From the six con-ditions which were investigated (2 operation conditions3 HES pressure stages) most compounds show a de-crease of NMHC emissions with increasing combustiontemperature at the same pressure stage This is in agree-ment with results from the combustor exit measurementsand with the results of the AEROTRACE study Emis-sion indices at HES pressure stages high (HP)- inter-mediate (IP)- and low pressure (LP) for fixed FSC andcombustor operation conditions were also investigatedMost compounds show a decrease of NMHC emissionswith decreasing pressure in the HES at similar combus-tor operation conditions and fixed FSC The decrease ofthe pressure in the HES corresponds to an increase ofthe residence time in a turbine and indicates a decreaseof NMHC emissions along the different turbine sections

35 Emission of ions and charged sootparticles

The ions measured during the HES campaign(HAVERKAMP et al 2004 WILHELM et al 2004) areproduct ions formed from primary ions via ion moleculereactions involving either few reaction-steps withmassive neutral molecules or numerous reaction-stepswith relatively small neutral molecules Their lifetimeswith respect to ion-ion recombination are around 9 msafter the high pressure turbine section The ion emissionindex Ei is up to 45 times 1016 ions per kg fuel burntAs is discussed by HAVERKAMP et al (2004) neitherion-ion recombination nor ion removal by attachmentto combustion particles give sufficient explanations forthe observed ion emission index Therefore detailedmodel calculations of the ion chemistry taking place inthe combustor with particular emphasis on the sulphurspecies effect is required (HAVERKAMP et al 2004)It seems that ion-induced nucleation can explain atleast partly the observed volatile aerosols Electricallycharged small positive and negative soot particles

(CSP) have also been detected during both PartEmiscampaigns Charged soot particles are found to havediameters around 6 nm and a total concentration of upto 45times107 cmminus3 (positive and negative) correspondingto a total emission index of ECSP = 23times 1015 CSP perkg fuel burnt (WILHELM et al 2004)

A model which considers the formation and evolu-tion of combustion ions in a combustor of an aircraftengine in dependence on the electron detachment ef-ficiency from negative ions was used to consider theeffect of the transformation of primary negative ionsto more stable secondary negative ions (SOROKIN etal 2003) The formed stable negative ions most prob-ably are sulphur-bearing ions This effect slows downthe charged particle neutralisation rate leading to an in-crease of the concentration of positive and negative ionsat the combustor exit The results of the simulation andtheir comparison with the ground-based experimentaldata obtained within PartEmis support the hypothesisthat the increase of the fuel sulphur content leads to anincrease of the ion concentration at the combustor exit

36 OH measurements

The concentration of OH which is the dominant oxidantin the exhaust gas was detected by laser induced fluo-rescence (LIF) to obtain an absolute OH concentrationmeasurement in order to aid in the modelling of SO3production from the combustor to the engine exit Whenrunning the laser at low laser energies OH in the exhaustduct was successfully detected Calibration experimentsprovide a method to assign the observed OH LIF sig-nals an absolute concentration and from an estimate ormeasurement of the errors for each parameter an overallestimate of the error in the assigned OH concentrationwas made From these calibration experiments very lowvalues of less than 1 ppbv were calculated for the OHconcentration in the exhaust gas after exiting the HESPrevious experiments using a combined Raman scatter-ingLIF technique to look in the exhaust gases 50 cmbehind an aircraft engine exit failed to detect any OHLIF signal and therefore could only infer an upper limitfor OH of 80 ppbv based on the measured detection limitof their apparatus

4 Summary and conclusions

The progress made during PartEmis is manifold Themain conclusions are therefore grouped into differentsubject areas and presented as brief statements For eachsubject area key references are identified if available

41 Combustion aerosol

Refer to WILSON et al (2004) and PETZOLD et al(2003a)

474 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

(i) The combustor behaves like a typical aircraft en-gine combustor with respect to thermodynamic dataand main emissions It may thus be expected that thePartEmis database can be applied as a first order approx-imation to contemporary aircraft engines

(ii) The BC concentration is nearly independent ofFSC level

(iii) The number size distribution of primary particlesin the diameter range 10 lt Dlt 550 nm is almost unin-fluenced by the FSC level or combustor conditions Athigh FSC levels the formation of volatile condensationparticles with diameters Dlt 10 nm occurs

(iv) The volatile volume fraction of particles gt 50 nmincreases from 2 for low FSC to approx 5 for highFSC The volatile volume fraction of particles lt 30 nmis le 15

(v) Volatility measurements indicate an internallymixed aerosol for particles with D gt 15 nm conden-sation particles are below 15 nm in diameter

42 Ultrafine particles

Refer to PETZOLD et al (2003a) and VANCASSEL et al(2004)

(vi) Formation of condensation particles of size D=4minus 7 nm is observed in the Combustor tests for highFSC at old and modern conditions formation occurs ifthe surface area of the combustion aerosol falls below acertain threshold limit the observed number of formedparticles is confirmed by kinetic model studies

(vii) Formation of condensation particles is observedin the HES tests for high FSCmodern conditionslowpressure section only the ldquoformation strengthrdquo in termsof number of condensation particles per number of com-bustion particle is twice as high as for the Combustorcase

(viii) The model studies indicate that condensationparticles are not emitted from the combustor but form inthe sampling line at the sample dilution point

(ix) Ultrafine carbonaceous particles of size 7 nmlt D lt 20 nm are observed independently of the fuelsulphur content a volatility analysis indicates that theseparticles are nonvolatile and may be composed of car-bonaceous compounds

43 Sulphate-containing species chemi-ionsand OH

Refer to KATRAGKOU et al (2004) and HAVERKAMP etal (2004)

(x) S(VI) was measured in the form of sulphuric acidIncreased FSC results in higher S(VI) concentrationsThe measured conversion efficiency for fuel sulphur toS(IV) at the combustor exit varies from 035ndash14 andis independent of the investigated combustor operationconditions and FSC levels

(xi) 10ndash30 of sulphate is chemisorbed on primaryparticles the coverage of primary particles is below 01monolayers of sulphuric acid at 150C and approxi-mately one monolayer at 25C

(xii) Sulphur partitioning at 150C 97 SO2le 27 gaseous H2SO4 lt 03 chemisorbed

(xiii) Positive and negative gaseous ions were de-tected in the exhaust of a combustor with mass numbersup to 1500 The total ion emission index at the HES hpstage is as high as 54times 1017 positive and negative ionsper kg fuel burnt Large ions with mass numbers exceed-ing 4500 contribute about 10 to the total ion numberconcentration

(xiv) OH concentration in the exhaust gas after exit-ing the HES is lt 1 ppbv

44 Hygroscopic growth and CCN activation

Refer to GYSEL et al (2003a) and HITZENBERGER etal (2003a)

(xv) Small differences in particle hygroscopicity areobserved between old and modern cruise conditions

(xvi) Strong effect of the fuel sulphur content is ob-served particle hygroscopicity increases distinctly withincreasing FSC

(xvii) No significant effect of the HES at low andmid FSC level is observed compared to the combustorexit an increased particle hygroscopicity downstreamthe HES compared to the combustor is observed at highFSC indicating continuing transformation of S(IV) toS(VI) during the transfer through the HES

(xviii) CCN activation of combustion particles in-creases with increasing FSC as a result of larger par-ticle hygroscopicity additional effects of adsorbed or-ganic matter cannot be excluded

45 Gaseous organic fraction

(xix) More than 100 different NMVOCs (aliphatic andaromatic hydrocarbons carbonyls and acids) were iden-tified and quantified they account for 91 wt of thetotal detected compounds

(xx) A decrease in NMHC emissions from old tomodern conditions at fixed HES pressure stage and fuelsulphur content (FSC) is found

(xxi) A decrease of NMHC emissions with decreas-ing pressure in the HES at fixed combustor operationconditions and FSC is observed

(xxii) No clear influence on the NMHC emissionswith changing FSC at the same combustor operationcondition is observed

(xxiii) High EIrsquos for organic acids are found Pos-sible explanations are that organic compounds fixed atthe sampling line walls are oxidised by the hot exhaustgases or that the acids are fixed at the surface of thesoot particles which are also trapped in the samplingcartridges

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 475

46 Emission properties

Refer to WILSON et al (2004) and PETZOLD et al(2003a)

(xxiv) Emission properties regarding combustionaerosol mass number and size are determined by thecombustor the influence of the Hot End Simulator isweak

(xxv) The processing of exhaust gas and parti-cles during the transfer through the Hot End Simula-tor increases the particle hygroscopicity and formationstrength of condensation particles both effects are re-lated to gaseous sulphuric acid

The experimental overview of the EU PartEmisproject presents the first comprehensive aerosol and gasmeasurements to be conducted on an aircraft enginecombustor test-rig The database will be used for fur-ther analyses as well as for plume modelling studies ofthe effect of aviation-related particle emissions on theglobal atmosphere and climate

Acknowledgements

The PartEmis project was funded by the European Com-mission under contract no G4RD-CT-2000-00207 Theauthors are very grateful to the test rig operation crew atQinetiQ for their very strong support during the experi-ments

References

ANDERSON BE WR COFER DR BAGWELL JWBARRICK CH HUDGINS KE BRUNKE 1998 Air-borne observations of aircraft aerosol emissions I Totalnonvolatile particle emission indices ndash Geophys Res Lett25 1689ndash1692

BOUCHER O 1999 Influence of air traffic on cirrus occur-rence Nature ndash 397 30ndash31

BROCK CA F SCHROumlDER B KAumlRCHER A PETZOLDR BUSEN M FIEBIG 2000 Ultrafine particle size distri-butions measured in aircraft exhaust plumes ndash J GeophysRes 105 26555ndash26567

CACHIER H MP BREMOND P BIAT-MEacuteNARD 1989Determination of atmospheric soot carbon with a simplethermal method ndash Tellus 41B 379ndash390

DEMOTT PJ Y CHEN SM KREIDENWEIS DCROGERS DE SHERMAN 1999 Ice formation by blackcarbon particles ndash Geophys Res Lett 26 2429ndash2432

GIEBL H A BERNER G REISCHL H PUXBAUM AKASPER-GIEBL R HITZENBERGER 2002 CCN activa-tion of oxalic and malonic acid test aerosols with the Uni-versity of Vienna cloud condensation nuclei counter ndash JAerosol Sci 33 1623ndash1634

GYSEL M E WEINGARTNER U BALTENSPERGER2002 Hygroscopicity of aerosol particles at low tempera-tures 2 Theoretical and experimental hygroscopic proper-ties of laboratory generated aerosols ndash Environ Sci Tech-nol 36 63ndash68

GYSEL M S NYEKI E WEINGARTNER U BAL-TENSPERGER H GIEBL R HITZENBERGER A PET-ZOLD CW WILSON 2003a Properties of jet engine com-bustor particles during the PartEmis experiment Hygro-scopicity at subsaturated conditions ndash Geophys Res Lett30 1566 DOI1010292003GL016896

GYSEL M S NYEKI E WEINGARTNER U BAL-TENSPERGER H GIEBL R HITZENBERGER A PET-ZOLD CW WILSON 2003b Jet engine combustion par-ticle hygroscopicity under subsaturated conditions duringPartEmis ndash J Aerosol Sci 34 (Suppl 2) Abstracts of theEuropean Aerosol Conference S1415ndashS1416

HAGEN DE PD WHITEFIELD H SCHLAGER 1996Particle emissions in the exhaust plume from commercialjet aircraft under cruise conditions ndash J Geophys Res 10119551ndash19557

HAGEN DE PD WHITEFIELD J PALADINO M TRUE-BLOOD H LILENFELD 1998 Particle sizing and emissionindices for a jet engine exhaust sampled at cruise ndash Geo-phys Res Lett 25 1681ndash1684

HAGEN DE PD WHITEFIELD JD PALADINO OSCHMID 2001 Comparative study of jet engine com-bustion emissions and engine operating conditions ndashIn SCHUMANN U G AMANATIDIS (Eds) Aviationaerosols contrails and cirrus clouds Air pollution researchreport 74 106ndash110 European Commission Brussels

HAVERKAMP H S WILHELM A SOROKIN F ARNOLD2004 Positive and negative ion measurements in jet aircraftengine exhaust concentrations sizes and implications foraerosol formation ndash Atmos Environ 38 2879ndash2884

HITZENBERGER R H GIEBL A PETZOLD M GYSELS NYEKI E WEINGARTNER U BALTENSPERGERCW WILSON 2003a Properties of jet engine combus-tor particles during the PartEmis experiment Hygroscopicproperties at supersaturated conditions ndash Geophys ResLett 30 1779 DOI1010292003GL017294

HITZENBERGER R H GIEBL A PETZOILD M GYSELS NYEKI E WEINGARTNER U BALTENSPERGERCW WILSON 2003b CCN activation of jet engine com-bustor particles during PartEmis ndash J Aerosol Sci 34(Suppl 2) Abstracts of the European Aerosol ConferenceS1409ndashS1410

HURLEY CD 1996 AEROTRACE Measurements of Par-ticulates from an Engine Combustor ndash Proceedings Impactof Aircraft Emissions upon the Atmosphere Vol 1 113 In-ternational Colloquium Paris October 1996

KATRAGKOU E S WILHELM F ARNOLD CW WIL-SON 2004 First gaseous Sulfur (VI) measurements in thesimulated internal flow of an aircraft gas turbine engineduring project PartEmis ndash Geophys Res Lett 31 2117DOI1010292003GL018231

NYEKI S M GYSEL E WEINGARTNER U BAL-TENSPERGER R HITZENBERGER A PETZOLD CWWILSON 2003 Volatile properties of jet engine combus-tor particles during PartEmis ndash J Aerosol Sci 34 (Suppl2) Abstracts of the European Aerosol Conference S1411ndashS1412

PETZOLD A FP SCHROumlDER 1998 Jet engine exhaustaerosol characterization ndash Aerosol Sci Technol 28 62ndash76

PETZOLD A A DOumlPELHEUER CA BROCK FPSCHROumlDER 1999 In situ observations and model calcula-

476 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

tions of black carbon emission by aircraft at cruise altitudendash J Geophys Res 104 22171ndash22181

PETZOLD A C STEIN S NYEKI M GYSEL EWEINGARTNER U BALTENSPERGER H GIEBL RHITZENBERGER A DOumlPELHEUER S VRCHOTICKY HPUXBAUM M JOHNSON CD HURLEY R MARSHCW WILSON 2003a Properties of jet engine combus-tor particles during the PartEmis experiment Microphysi-cal and chemical properties ndash Geophys Res Lett 30 1719DOI1010292003GL017283

PETZOLD A CW WILSON U BALTENSPERGERM FIEBIG L FRITZSCHE H GIEBL M GYSELR HITZENBERGER CD HURLEY S NYEKI HPUXBAUM U SCHUMANN C STEIN S VROCHTICKY2003b Particle emissions from aircraft engines ndash anoverview of the European project PartEmis ndash J AerosolSci 34 (Suppl 2) Abstracts of the European Aerosol Con-ference S1405ndashS1406

PETZOLD A AE VRTALA M FIEBIG F FRITZSCHEPH MIRABEL S NYEKI X VANCASSEL CW WIL-SON 2003c Emission of volatile and non-volatile ultra-fine particles from a combustion source during PartEmisndash J Aerosol Sci 34 (Suppl 2) Abstracts of the EuropeanAerosol Conference S1407ndashS1408

PUXBAUM H S VRCHOTICKY A PETZOLD R HITZEN-BERGER S NYEKI CW WILSON 2003 Chemical com-position of jet engine combustor particles during PartEmisndash J Aerosol Sci 34 (Suppl 2) Abstracts of the EuropeanAerosol Conference S1413ndashS1414

SCHMID H L LASKUS H-J ABRAHAM U BAL-TENSPERGER V LAVANCHY M BIZJAK P BURBA HCACHIER D CROW J CHOW T GNAUK A EVENHM TEN BRINK K-P GIESEN R HITZENBERGER CHUEGLIN W MAENHAUT C PIOA CARVALHO J-PPUTAUD D TOOM-SAUNTRY H PUXBAUM 2001 Re-sults of the ldquocarbon conferencerdquo international aerosol car-bon round robin test stage I ndash Atmos Environ 35 2111ndash2121

SCHROumlDER F C BROCK R BAUMANN A PETZOLD RBUSEN P SCHULTE M FIEBIG 2000 In-situ studies onvolatile jet exhaust particle emissions Impact of fuel sul-fur content and thermodynamic conditions on nuclei-modeaerosols ndash J Geophys Res 105 19941ndash19954

SCHUMANN U 1996 Contrail formation from aircraft ex-hausts ndash Meteorol Z NF 5 4ndash23

SCHUMANN U J STROumlM 2001 Aviation impact on atmo-spheric composition and climate ndash In European research inthe stratosphere 1996ndash2000 European Commission EUR19867 Brussels

SCHUMANN U F ARNOLD R BUSEN J CURTIUS BKAumlRCHER A KIENDLER A PETZOLD H SCHLAGERF SCHROumlDER K-H WOHLFROM 2002 Influence of fu-els sulphur on the composition of aircraft exhaust plumesThe experiments SULFUR 1-7 ndash J Geophys Res 107(1010292001JD000813) AAC 2-1 ndash AAC 2-27

SOROKIN A P MIRABEL F ARNOLD 2003 On the In-fluence of fuel fulfur induced stable negative ion formationon the total concentration of ions emitted by an aircraft gasturbine engine Comparison of model and experiment ndash At-mos Chem Phys 3 6001ndash6018

VANCASSEL X A SOROKIN P MIRABEL A PETZOLDCW WILSON 2004 Volatile particles formation duringPartEmis a modelling study ndash Atmos Chem Phys 4 439ndash447

WILHELM S H HAVERKAMP A SOROKIN F ARNOLD2004 Detection of very large ions in aircraft gas turbineengine combustor exhaustcharged small soot particles ndashAtmos Environ 38 4561ndash4569

WILSON CW A PETZOLD S NYEKI U SCHUMANNR ZELLNER 2004 Measurement and Prediction of Emis-sions of Aerosols and Gaseous Precursors from Gas Tur-bine Engines (PartEmis) An Overview ndash Aerosp SciTechnol 8 131ndash143

Page 3: Particle emissions from aircraft engines a survey of the ... · 466 A. Petzold et al.: Particle emissions from aircraft engines Meteorol. Z., 14 , 2005 1 Introduction The role of

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 467

Figure 1 Schematic of the PartEmis combustor and Hot End Simulator HES the HES is composed of three separate heat exchanger

stages referred to as high (hp) intermediate (ip) and low pressure (lp) stages

Table 2 Available equipment and parameters

Parameter TechniqueVolatile condensation particles Condensation Particle Size Analyser (CPSA)(number size) Differential Mobility Analyser (DMA) connected to a ThermodenuderNon-volatile carbonaceous particles Condensation Particle Counter (CPC)(number size) Scanning Mobility Particle Sizer (SMPS)

DMA connected to a ThermodenuderVolatility Tandem-DMA

Surface density Diffusion ChargerAerosol absorption coefficient babs Aerosol Absorption PhotometerSmoke Number SN Smoke number methodWater uptake at RH le 100 Hygroscopicity Tandem-DMACloud condensation nuclei activation Thermal gradient CCN counterChemical composition Berner low pressure impactor (BLPI) samples

filter stack samplesanalysis viagravimetry (total mass)multi-step combustion method (organicelementalcarbon)evolved gas analysis (volatility of carbonaceouscompounds)gas chromatography (aliphatic constituents)ion chromatography (major ions)

CO CO2 HC NOx O2 H2 Standard test rig instrumentationHONO HNO3 Chemical ionisation mass spectrometrySO2 SO3 H2SO4 Paul Ion Trap Mass Spectrometry (PITMAS)Chemi-ionsSO2Gaseous hydrocarbon species Quasi-on-line gas chromatography (GC)

Sampling and off-line analysisPartially oxidised hydrocarbons DNPH methodOH Laser-induced fluorescence (LIF)

level Fuel sulphur content (FSC) levels covered therange from low sulphur fuel (005 g kgminus1) to the con-temporary average (041 g kgminus1) and to a maximum ofabout three times this average (1270 g kgminus1) definedhere as Low Mid and High (L M H) FSC respectivelyThe sulphur content was varied by adding known quanti-ties of benzenethiol (C6H5SH) to the low FSC fuel Theoperation conditions are summarised in Table 1

The employed aerosol measurement methods con-sisted of various size-selective methods aerosol ab-sorption photometers thermodenuder methods a CloudCondensation Nuclei (CCN) Counter (HITZENBERGER

et al 2003a) an extensive set of chemical analyti-cal methods (PETZOLD et al 2003a PUXBAUM et al2003) mass spectrometry methods for the detection ofsulphur-containing species and laser-induced fluores-

cence for OH measurements A summary of deployedmethods is given in Table 2

The black carbon (BC) mass concentration and theaerosol absorption coefficient in the diluted sample gaswere measured using a Particle Soot Absorption Pho-tometer (PSAP Radiance Research Inc Seattle USA)The specific absorption cross section at λ = 055micromσabs = 67 m2 gminus1 was used for conversion of lightabsorption to BC mass The particle number concen-trations were measured in the size intervals D = 4minus7nm 7ndash9 nm 9ndash20 nm and 20ndash250 nm with parallel-operated condensation particle counters (CondensationParticle Size Analyser CPSA) which were set to differ-ent lower instrument cut-off diameters Particles of sizeD gt 9 nm which are detected in CPSA channels 3 and4 are in the following referred to as N10 Particle size

468 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

distributions were measured with a commercial Scan-ning Mobility Particle Sizer (SMPS TSI Model 3071)and a Volatility-Tandem Differential Mobility Analyser(V-TDMA) in the size range Dgt 13 nm

The aerosol mixing state and the volatile fractionof the combustion aerosol were determined using ther-mal denuder methods which differentiate particles bynumber according to their thermal stability As wasinvestigated in previous calibration experiments ther-mal denuder temperatures of 390 K and 573ndash625 Krespectively distinguish volatile (vaporises at 390 K)sulphuric-acid like material from semivolatile (vaporisesbetween 390 and 573 K) and refractory black-carbonlike material The thermal treatment of almost monodis-perse particle size fractions which were pre-selected bya DMA provides information whether the aerosol isinternally or externally mixed The ratio of diametersafter and before thermal treatment yields the particleshrinkage factor D32D

30 This factor describes the par-

ticle volume reduction by the thermal treatment ie thevolatile volume fraction of internally mixed particles is1minusD3D3

0The hygroscopic particle growth factor under sub-

saturated conditions was measured with a Hygroscop-icity Tandem DMA (GYSEL et al 2002 2003a) Dur-ing the measurements the instrument was kept at a con-stant temperature of T sim= 298 K The obtained hygro-scopic growth factor describes the particle growth bywater uptake from the gas phase at a selected relativehumidity Measurements were performed in the rela-tive humidity range from 70 to approximately 95 As was demonstrated by GYSEL at al (2002 2003a)from KOEHLER theory the hygroscopic growth factorof a particle is a function of dry particle size relativehumidity and particle coating with a water-soluble sub-stance For the PartEmis experimental data the coatingsubstance was sulphuric acid The magnitude of the par-ticle growth at a given relative humidity and for a givensize of the dry particle can be directly linked to volumefraction of water-soluble matter contained in the inves-tigated particles During the PartEmis experiments thehygroscopic growth factors measured for fixed dry par-ticle sizes but variable fuel sulphur contents were usedfor the determination of the coating of the combustionaerosol by sulphuric acid The number of particles whichbecome finally activated as cloud condensation nuclei inthe case of water-supersaturation were measured with astatic thermal diffusion chamber CCN counter (GIEBL

et al 2002) The instrument was operated at a super-saturation ratio of 06ndash07 with respect to liquid waterThe activation ratio or fraction of CCN activated par-ticles respectively was determined as the ratio of thenumber of activated CCN to the number of total particles(HITZENBERGER et al 2003a) The ability of combus-tion particles for taking up water from the gas phase is a

key property in the formation processes for liquid waterclouds However it is also of relevance for jet conden-sation trail formation (SCHUMANN 1996) and it mayas well play a role in the formation of ice nuclei fromcombustion particles (DEMOTT et al 1999)

Aerosol samples were collected with filter stack sam-plers and low pressure Berner impactors for total aerosoland size-resolved chemical mass analysis (PUXBAUM etal 2003) The chemical composition of the carbona-ceous fraction of the exhaust particles was determinedusing multi-step combustion methods Total carbon (TC)was determined by a combustion technique involvingcombustion in a vertical furnace and detection of CO2with a commercial NDIR (non dispersive Infrared spec-trometer) unit Aliquots of a collected filter sample wereburnt in the ceramic tube of a combustion furnace withan excess of oxygen at a temperature of 1000C Ele-mental carbon (EC) was determined by the ldquoCachierrdquomethod (CACHIER et al 1989) ie another aliquot ofthe filter sample was treated for 2 h at 320C in an oxy-gen atmosphere to remove organic carbon compoundsand then analysed with the same TC analysis methodThe remaining carbon was attributed to elemental car-bon (EC) Thermograms of the collected carbonaceousmaterial were measured with a thermo-optical evolvedgas analysis method (SCHMID et al 2001) To deter-mine anions eg SO2minus

4 as well as cations two Dionexisocratic systems with electrochemical suppression wereused

In addition the aerosol was sampled on quartz tubescontaining glass wool The aerosol loaded quartz tubeswere analysed by 5 min thermodesorption in a He-lium atmosphere at 130C and 350C respectively Thedesorbed C2ndashC12 non-methane hydrocarbons (NMHC)were analysed with a gas chromatograph (GC HewlettPackard) NMHCs were monitored using a compact GCinstrument (Airmovoc 2010) with a cryotrap as an en-richment system and a flame ionisation detector C2ndashC10 aliphatic and aromatic NMHCswere measured witha time resolution of 10 min and with detection limits inthe pptv range Selected partially oxidised hydrocarbons(OHC) such as carbonyl compounds and organic acidswere collected by using commercial sampling cartridgescoated with 24-Dinitrophenylhydrazine This chemicalcompound forms with the carbonyl molecules a stablecompound called hydrazone This hydrazone is stablefor weeks and can be analysed later in the LaboratoryTo analyse the carbonyl compounds the hydrazones arerinsed out of the cartridge and separated in a High Per-formance Liquid Chromatograph HPLC The OHCweremeasured with a time resolution of 4 min and with de-tection limits in the ppbv range All hydrocarbon com-pound measurements were corrected for line losses andbackground values

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 469

Figure 2 Left size distribution of combustion aerosol with (diamonds) and without condensation particle formation (triangles) RightExternal aerosol mixture composed of volatile condensation particles (grey) and coated nonvolatile combustion particles (black)

Gaseous sulphuric acid was measured using a CIMS(Chemical Ionisation Mass Spectrometer) apparatusequipped with an Ion Trap Mass Spectrometer (ITMS)The CIMS apparatus comprises a stainless steel flow re-actor (inner diameter 4 cm total length 164 cm) heatedup to 80C a high frequency glow discharge ion sourcefor reagent ion production and a Quadrupole Ion TrapMass Spectrometer for ion detection The ITMS allowsCID (Collision-Induced-Dissociation) investigations ofmass selected ions to be performed which in turn allowsmuch better species identification than can be offeredby a conventional mass spectrometer Reactant ions ofthe type NO3 (HNO3)n (n= 012) are produced in aseparate chamber where a high frequency gas dischargeburns in a buffer gas of O2 and NO2 N2 and injectedinto the flow reactor upstream of the ion sampling orificeof the Mass Spectrometer After injection ions are car-ried by the gas stream to the mass spectrometer Whengaseous sulphuric acid is present in the flow reactorsome fraction of the reagent ions reacts via a gas phaseion molecule reaction with sulphuric acid By measur-ing the abundance ratio of product and reagent ions thegaseous sulphuric acid concentration in the flow reactoris determined For more details see KATRAGKOU et al(2004) and references given there

OH laser induced fluorescence (LIF) measurementswere performed in the exhaust duct of the HES The ex-periment comprised a NdYAG pumped dye laser set upto generate light at λ = 281913 nm in order to excite theQ1(1) transition of OH A2Σ(ν prime = 1)larr X2Π(ν primeprime = 0)chosen as it is one of the strongest transitions at the tem-perature expected at the observation point Fluorescencewas detected perpendicular to the path of the laser fromthe ν prime = 0rarr vprimeprime = 0 transition around 309 nm via a lensand a customised interference filter In order to max-imise the signal to noise ratio by discriminating againstRaman scatter expected from N2 and H2O an interfer-ence filter was centred at 30875 nm with a band passof 496 nm A gated CCD camera cooled to ndash30C was

used to collect the signal allowing signal accumulationover long times in order to enhance sensitivity

Since the full set of applied instruments was con-nected to separate sampling lines according to specificinstrument requirements the precise determination ofthe dilution factors was a crucial task After exit fromthe traversing probe the sample was cooled to sim 150Cusing a water-jacket and delivered to the undiluted anddiluted sample lines The undiluted line was insulatedalong its entire length in order to avoid cooling andtherefore wall losses of less volatile combustion prod-ucts A detailed description of the extracting probe andthe sample transportation system is given in WILSON

et al (2004) Dilution with filtered air was necessaryfor the majority of instruments to reduce concentrationsdown to a measurable level The sample dilution factorwas determined from CO2 mixing ratios measured in thesupplied dilution air diluted sample and undiluted sam-ple as well as from temperature and pressure sensorsthroughout the sample delivery system The average di-lution factor was 65 (PETZOLD et al 2003a)

Although stable gaseous species such as CO2 andwater vapour are unaffected by intrusive samplingaerosols may be lost due to diffusion and sedimenta-tion to the surfaces of the sampling system Estimatesof expected line losses were obtained from experiments(HURLEY 1996) performed with aerosol generated bya spark discharge generator (PALAS GfG 1000) Theaerosol was fed into sample lines of different cross sec-tion and length and the sample flow through the lineswas varied From these studies expected line losses arelt 20 by number of particles for the conditions dur-ing the PartEmis experiments Since line losses affectpredominantly small particles lt 20 nm in diameter theaerosol mass is assumed to be almost unaffected by linelosses However the values given for emission indiceshave to be taken as upper values Size distributions re-ported for combustion particles cover a size range of 10ndash250 nm The size spectra used for the investigation of the

470 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

impact of operation conditions and fuels sulphur con-tent were not corrected for line losses because correc-tions are similar for all test points Thus relative trendsconcerning the influence of fuel sulphur and operationconditions on the microphysical properties of the emit-ted particles are not affected by line losses Becauseof the observed size range of exhaust particles countmedian diameters and geometric standard deviations ofmeasured size distributions are also expected to be onlymarginally influenced by line losses

3 Results

31 Microphysical properties

The aerosol emitted from the Combustor is composedof primary particles forming inside the combustor andof volatile condensation particles nucleating in the cool-ing exhaust gas from gaseous precursors The aerosolsize distribution in the diluted exhaust gas of the com-bustor is shown in Fig 2 for conditions with and with-out condensation particle formation While the mode ofcombustion aerosol particles dominates the size rangeD gt 15 nm with a maximum at 40 nm another modeof nanoparticles is present in all cases in the size rangeD= 7minus20 nm Condensation particle formation is oc-curring only at high FSC conditions and contributes tothe size range D= 4minus7 nm

The emission properties with respect to the primaryaerosol are shown in terms of BC mass (EIBC ) TCmass (EITC) and number (EIN10) emissions per kg ofconsumed fuel in Fig 3 The number emission indexEIN10 refers to particles with diameters D ge 10 nm InFig 3 also the count median diameter (CMD) of theexhaust aerosol size distribution is shown Plotted arecorresponding values for the combustor emissions andemissions from the HES low pressure stage In the firstplace it has to be mentioned that all the results are con-sistent At old cruise conditions an increasing numberemission at constant black carbon mass emission is bal-anced by a decreasing particle size with respect to theparticle electrical mobility diameter It should be notedthat these results originate from different measurementtechniques

When operated at old engine conditions with lowcombustor inlet temperature TC EIBC varies from 0051to 0081 g kgminus1 Operated at modern engine condi-tions with the higher inlet temperature EIBC ranges from0028 to 0044 g kgminus1 which is a factor of two lowerSimilar relations hold for the other properties How-ever the amount of emitted particulate matter is withinthe range permitted by the ICAO rules in all cases Allparticle emission indices referring to number and massare in close agreement with values obtained from in-flight measurements which span a range from 05ndash075

Figure 3 Emission indices in terms of number (N10 ) total car-

bon mass (TC) black carbon mass (BC) and count median diameter

(CMD) of the combustion aerosol for the Combustor exit (triangles)

and the Hot End Simulator low pressure stage (diamonds) values are

given for old and modern operation conditions as a function of the

fuel sulphur content error bars indicate signal variability in terms of

plusmn1 stdev

g kgminus1 for a Boeing B707 to 001 g kgminus1 for AirbusA 340 and Boeing B737 equipped with CFM 56 se-ries engines (SCHUMANN et al 2002 PETZOLD et al2003a 2003b) The fleet-averaged mass emission indexof 0038 g kgminus1 (PETZOLD et al 1999) lies within therange of values determined for modern operation condi-tions

Mean CMD values are 37ndash41 nm for old engine and40ndash44 nm for modern engine conditions The geomet-ric standard deviation of the size distribution is 167(16ndash18) for both operation conditions (PETZOLD etal 2003a 2003b) The combustor operating conditionshave only a weak influence on the size number and massof emitted particles Size distribution peaking at similardiameters were observed during several in-flight stud-ies HAGEN et al (1996 1998) reported peak diame-ters of 30ndash60 nm and size distribution widths of approx16 During the SULFUR studies (PETZOLD et al 1999SCHUMANN et al 2002) peak diameters of 35 nm wereobserved The PartEmis size distribution parameters arein close agreement with the reported in-flight observa-tions

The influence of the different HES pressure stages isdemonstrated in Fig 3 and with more detail in Fig 4As is shown in Fig 3 the trends observed in particleemission properties in the combustor experiments withvarying fuel sulphur content and operation conditionsagain observed during the HES experiments Except fora step-like increase from the combustor experiments tothe HES experiments which is most likely due to slightlydifferent operating conditions the HES has no influenceon the emission properties of the Combustor-HES com-bination with respect to combustion aerosol particles

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 471

0

5

10

15

20

25

30

000

005

010

015

020

high intermed low pressure

Combustor Hot End Simulator section

EIBC

EIN

EI

BC

g

kg

-1

EI

N

10

14kg

-1

Figure 4 Effect of the Hot End Simulator on the emission propertiesof the Combustor-HES combination operated at modern conditions

error bars indicate signal variability in terms of plusmn1 stdev

see Fig 4 Increasing the FSC slightly increases the to-tal number of emitted particles

Particles smaller than 10 nm contribute less than10 to the total exhaust aerosol in the low and mediumFSC cases At high FSC levels the formation of volatilecondensation particles with diameters Dlt 10 nm is ob-served (PETZOLD et al 2003c) A kinetic model wasused to investigate the nucleation efficiency of the H2O- H2SO4 binary mixture in the exhaust gas and the par-ticle coagulation processes during the transport throughthe sampling system to the measurement instrumenta-tion (VANCASSEL et al 2004) The model confirmedthat the condensation particles form at the dilutor of thesampling system During combustion the sulphur con-tained in the fuel is oxidised to SO2 which is then par-tially converted to S(VI) in the hot exhaust gas The ef-ficiency of conversion from S(IV) to S(VI) is describedby the conversion efficiency calculated as ε =S(VI)STwhere ST is the total fuel sulphur In the combustor ex-haust the S(VI) increases with increasing FSC In thekinetic model study a value for the efficiency was indi-rectly deduced from comparisons between model resultsand measurements In the PartEmis cases ε ranges be-tween roughly 25 and 6 depending on the com-bustor settings and on the microphysical approach usedFor these values which are close to results from di-rect measurements of gaseous sulphuric acid the kineticmodel reproduces the observation of volatile particles inthe size range 4ndash7 nm for the high FSC case while novolatile particles are found by the model for the low andmedium FSC cases

Particles of size D ge 30 nm are almost entirely in-ternally mixed Smaller particles (D lt 20 nm) exhibitvolume fractions of 10ndash15 for volatile and 4ndash10 for semi-volatile material while the volatile fraction de-creases to lt 5 for D ge 50 nm particles (PETZOLD

et al 2003a NYEKI et al 2003) The decrease of thevolatile particle fraction with increasing particle size isalso reproduced by the kinetic model for the H2O ndashH2SO4 binary mixture (VANCASSEL et al 2004) Thesevalues are valid for the conditions met during PartEmisIf the dilution air is more humid and a higher dilutionforces stronger formation of condensation particles thecondensation particles may grow to larger sizes How-ever as was observed in a young aircraft plume at cruiseconditions condensation particles were also observedmainly in the size range Dlt 10 nm (SCHROumlDER et al2000) similar to the PartEmis results

32 Chemical properties

Primary particles emitted from the combustor arepredominantly composed of carbonaceous material(gt95 ) and of water-soluble inorganic componentsThe carbonaceous material is composed of organic com-pounds and of elemental or black carbon The organiccompounds evaporate and oxidise at different tempera-tures and become classified according to this criterionCompounds which evaporate at temperatures as low as400 K are classified as semi-volatie OC compoundsevaporating compounds evaporating up to 800 K aretermed nonvolatile OC and the remainder is elementalcarbon Fig 5 shows the thermal stability of an exhaustaerosol sampled during operation with low FSC fuel atmodern conditions In the applied version of evolved gasanalysis (SCHMID et al 2001) the blackness of the filtersample is recorded parallel to the heating of the sampleAs is demonstrated in Fig 5 the filter starts to becomeldquowhiterdquo again as soon as the elemental or black carbonstarts to oxidise The evaporation and subsequent oxi-dation of OC compounds leaves the filter black Thusthe onset of increasing filter transmittance is an addi-tional measure for separating organic and elemental car-bon The application of this methods yields an averagecontribution of black carbon to total carbon of 40ndash80 depending on the combustion conditions (PUXBAUM etal 2003) From an aerosol loaded glass wool samplemore than 180 different NMHCs were desorbed underan Helium atmosphere Only 04 wt of the NMHCswere desorbed at T le 400 K whereas 996 wt weredesorbed at T = 400ndash620 K This is in good agreementwith the result of the thermogram in Fig 5 Elemental(EC) or black carbon (BC) respectively is chemicallyalmost inert at low and moderate temperatures and ther-mally stable up to temperatures of about 700 K

As determined from direct measurements of gaseousH2SO4 the efficiency of conversion in the combus-tor exhaust ranges between 09 plusmn 05 and 13 plusmn07 (KATRAGKOU et al 2004) In the HES ex-haust S(VI) increases with increasing FSC fuel flowand from the high-pressure to the low-pressure HESstage For FSC=1270 g kgminus1 the conversion efficiency

472 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

is 23plusmn 12 at the low pressure stage for the mod-ern cruise condition and 14plusmn 07 for the old cruiseThe higher ε observed for the modern cruise conditionssuggests that modern engines have a larger ε comparedto old engines Conversion efficiencies determined fromin-flight data at cruise altitude gave values 04ndash45 (SCHUMANN et al 2002) which is in good agreementwith the PartEmis ground test values

Chemical composition information was obtainedfrom Berner low pressure impactor samples which weretaken at a sample line temperature of 150C (PUXBAUMet al 2003) At these conditions most of the sulphuricacid remains in the gas phase while only a small fractionis more strongly bound to or chemi-sorbed at the par-ticle surface and will be observed in particulate mattersamples The chemical analysis results indicate that themass distribution of the carbonaceous matter matchesthe volume distribution of the aerosol The mass dis-tribution of chemisorbed sulphate however follows thesurface distribution of the aerosol which was calculatedfrom aerosol size spectra assuming a spherical particleshape (PUXBAUM et al 2003) This observation indi-cates that the sulphate is chemisorbed at the particle sur-face The chemisorbed fraction of sulphuric acid corre-sponds to a coverage of max 01 monolayer in the highFSC case This fraction of converted S(VI) is less than01 of total S(IV) present in the fuel The overall par-titioning of sulphur-containing species at 150C is 97 SO2le 27 gaseous H2SO4lt 03 chemisorbedAfter cooling to environmental temperatures the gasphase sulphuric acid can condense and cover the com-bustion particles with about one monolayer of sulphuricacid on average (HITZENBRGER et al 2003a)

33 Particle hygroscopic properties

Particle hygroscopic properties were measured at ambi-ent conditions using a Hygroscopicity Tandem Differ-ential Mobility Analyser (H-TDMA) system (GYSEL etal 2003a 2003b) The instrument measures the hygro-scopic growth factor (g) of particles with initial dry sizesDo = 30 50 and 100 nm over the range RHsim 70minus95 where g(RH) = D(RH)Do Growth factors were found toincrease with increasing FSC at fixed particle size (egDo = 50 nm g (95 ) = 101 and 116 for FSC = lowand high respectively) and to decrease with increasingparticle size at fixed FSC (eg mid FSC g (95 ) =114 110 and 103 for Do = 30 50 and 100 nm respec-tively) The results suggest an increasing amount of sul-phuric acid adsorbed on combustion particles from thegas phase with increasing FSC

Cloud condensation nuclei (CCN) were measured inthe diluted line using a static thermal diffusion typeCCN counter (GIEBL et al 2002) Particles are exposedto controlled supersaturations of water vapour whichenables some of the particles to become activated and

Figure 5 Evolving CO2 as a function of the sample heating temper-ature for a Combustor aerosol sample semi-volatile OC oxidises at

450 K non-volatile OC at 600 K and elemental carbon above 700 K

grow to large (gt 10 microm) droplets During PartEmis thesupersaturation was set tosim 07 The fraction of CCNin the aerosol (relative to the concentration of particlesof Dgt 10 nm) increases with increasing FSC but evenfor low FSC the fraction of CCN (at 0755 supersat-uration) is higher than the fraction of particles with di-ameters larger than 280 nm ie the size at which wet-table insoluble particles are activated according to theKelvin effect At high FSC the fraction of CCN in-creases by about a factor 55 for old and 12 for mod-ern conditions However pure sulphuric acid particlesof the same size which become activated according toKOumlHLER theory would give an activation ratio which isstill larger by a factorsim 15 compared to the observationseven at high FSC Thus the KOumlHLER model cannot re-produce measured activation properties This is in agree-ment with the fact that large combustion particles are notat all completely water soluble The Kelvin model pro-vides good predictions for the low FSC level but acti-vation diameters (ratios) are increasingly overestimated(underestimated) at medium and high FSC This indi-cates that the soluble fraction of the combustion parti-cles increases from low values at low FSC level to rel-atively large values at high FSC An empirical coatedsphere model which equates the measured volatile andsemivolatile fraction with soluble sulphuric acid under-estimates the activation diameter by 45 to 55 underall conditions A small amount of semivolatile organ-ics with low or absent water-solubility adsorbed on thecombustion particles might be one reason for this dis-crepancy Further details are given by HITZENBERGER

et al (2003a 2003b)

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 473

34 Gaseous hydrocarbon emission

More than 100 different non-methane volatile organiccompounds (NMVOCs) such as non-methane hydrocar-bons (NMHCs) and selected partially oxidised hydro-carbons (OHCs) emitted from the aircraft engine com-bustor and the hot end simulator (HES) were identifiedand quantified These species accounted for up to 91wt of the total hydrocarbons (THCs) The NMHCemission indices of 5ndash15 microg kgminus1 in total are compa-rable with the AEROTRACE (Measurement of TraceSpecies in the Exhaust from Aero Engines EC projectAERA-CT94-0003 1994ndash1997 Co-ordinated by Qine-tiQ) study Both studies show up to 1000 times loweremission indices in comparison with emissions from ve-hicular traffic The influence of the operation conditionson NMHC emissions of the combustor and HES was in-vestigated for the identified NMHCs From the six con-ditions which were investigated (2 operation conditions3 HES pressure stages) most compounds show a de-crease of NMHC emissions with increasing combustiontemperature at the same pressure stage This is in agree-ment with results from the combustor exit measurementsand with the results of the AEROTRACE study Emis-sion indices at HES pressure stages high (HP)- inter-mediate (IP)- and low pressure (LP) for fixed FSC andcombustor operation conditions were also investigatedMost compounds show a decrease of NMHC emissionswith decreasing pressure in the HES at similar combus-tor operation conditions and fixed FSC The decrease ofthe pressure in the HES corresponds to an increase ofthe residence time in a turbine and indicates a decreaseof NMHC emissions along the different turbine sections

35 Emission of ions and charged sootparticles

The ions measured during the HES campaign(HAVERKAMP et al 2004 WILHELM et al 2004) areproduct ions formed from primary ions via ion moleculereactions involving either few reaction-steps withmassive neutral molecules or numerous reaction-stepswith relatively small neutral molecules Their lifetimeswith respect to ion-ion recombination are around 9 msafter the high pressure turbine section The ion emissionindex Ei is up to 45 times 1016 ions per kg fuel burntAs is discussed by HAVERKAMP et al (2004) neitherion-ion recombination nor ion removal by attachmentto combustion particles give sufficient explanations forthe observed ion emission index Therefore detailedmodel calculations of the ion chemistry taking place inthe combustor with particular emphasis on the sulphurspecies effect is required (HAVERKAMP et al 2004)It seems that ion-induced nucleation can explain atleast partly the observed volatile aerosols Electricallycharged small positive and negative soot particles

(CSP) have also been detected during both PartEmiscampaigns Charged soot particles are found to havediameters around 6 nm and a total concentration of upto 45times107 cmminus3 (positive and negative) correspondingto a total emission index of ECSP = 23times 1015 CSP perkg fuel burnt (WILHELM et al 2004)

A model which considers the formation and evolu-tion of combustion ions in a combustor of an aircraftengine in dependence on the electron detachment ef-ficiency from negative ions was used to consider theeffect of the transformation of primary negative ionsto more stable secondary negative ions (SOROKIN etal 2003) The formed stable negative ions most prob-ably are sulphur-bearing ions This effect slows downthe charged particle neutralisation rate leading to an in-crease of the concentration of positive and negative ionsat the combustor exit The results of the simulation andtheir comparison with the ground-based experimentaldata obtained within PartEmis support the hypothesisthat the increase of the fuel sulphur content leads to anincrease of the ion concentration at the combustor exit

36 OH measurements

The concentration of OH which is the dominant oxidantin the exhaust gas was detected by laser induced fluo-rescence (LIF) to obtain an absolute OH concentrationmeasurement in order to aid in the modelling of SO3production from the combustor to the engine exit Whenrunning the laser at low laser energies OH in the exhaustduct was successfully detected Calibration experimentsprovide a method to assign the observed OH LIF sig-nals an absolute concentration and from an estimate ormeasurement of the errors for each parameter an overallestimate of the error in the assigned OH concentrationwas made From these calibration experiments very lowvalues of less than 1 ppbv were calculated for the OHconcentration in the exhaust gas after exiting the HESPrevious experiments using a combined Raman scatter-ingLIF technique to look in the exhaust gases 50 cmbehind an aircraft engine exit failed to detect any OHLIF signal and therefore could only infer an upper limitfor OH of 80 ppbv based on the measured detection limitof their apparatus

4 Summary and conclusions

The progress made during PartEmis is manifold Themain conclusions are therefore grouped into differentsubject areas and presented as brief statements For eachsubject area key references are identified if available

41 Combustion aerosol

Refer to WILSON et al (2004) and PETZOLD et al(2003a)

474 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

(i) The combustor behaves like a typical aircraft en-gine combustor with respect to thermodynamic dataand main emissions It may thus be expected that thePartEmis database can be applied as a first order approx-imation to contemporary aircraft engines

(ii) The BC concentration is nearly independent ofFSC level

(iii) The number size distribution of primary particlesin the diameter range 10 lt Dlt 550 nm is almost unin-fluenced by the FSC level or combustor conditions Athigh FSC levels the formation of volatile condensationparticles with diameters Dlt 10 nm occurs

(iv) The volatile volume fraction of particles gt 50 nmincreases from 2 for low FSC to approx 5 for highFSC The volatile volume fraction of particles lt 30 nmis le 15

(v) Volatility measurements indicate an internallymixed aerosol for particles with D gt 15 nm conden-sation particles are below 15 nm in diameter

42 Ultrafine particles

Refer to PETZOLD et al (2003a) and VANCASSEL et al(2004)

(vi) Formation of condensation particles of size D=4minus 7 nm is observed in the Combustor tests for highFSC at old and modern conditions formation occurs ifthe surface area of the combustion aerosol falls below acertain threshold limit the observed number of formedparticles is confirmed by kinetic model studies

(vii) Formation of condensation particles is observedin the HES tests for high FSCmodern conditionslowpressure section only the ldquoformation strengthrdquo in termsof number of condensation particles per number of com-bustion particle is twice as high as for the Combustorcase

(viii) The model studies indicate that condensationparticles are not emitted from the combustor but form inthe sampling line at the sample dilution point

(ix) Ultrafine carbonaceous particles of size 7 nmlt D lt 20 nm are observed independently of the fuelsulphur content a volatility analysis indicates that theseparticles are nonvolatile and may be composed of car-bonaceous compounds

43 Sulphate-containing species chemi-ionsand OH

Refer to KATRAGKOU et al (2004) and HAVERKAMP etal (2004)

(x) S(VI) was measured in the form of sulphuric acidIncreased FSC results in higher S(VI) concentrationsThe measured conversion efficiency for fuel sulphur toS(IV) at the combustor exit varies from 035ndash14 andis independent of the investigated combustor operationconditions and FSC levels

(xi) 10ndash30 of sulphate is chemisorbed on primaryparticles the coverage of primary particles is below 01monolayers of sulphuric acid at 150C and approxi-mately one monolayer at 25C

(xii) Sulphur partitioning at 150C 97 SO2le 27 gaseous H2SO4 lt 03 chemisorbed

(xiii) Positive and negative gaseous ions were de-tected in the exhaust of a combustor with mass numbersup to 1500 The total ion emission index at the HES hpstage is as high as 54times 1017 positive and negative ionsper kg fuel burnt Large ions with mass numbers exceed-ing 4500 contribute about 10 to the total ion numberconcentration

(xiv) OH concentration in the exhaust gas after exit-ing the HES is lt 1 ppbv

44 Hygroscopic growth and CCN activation

Refer to GYSEL et al (2003a) and HITZENBERGER etal (2003a)

(xv) Small differences in particle hygroscopicity areobserved between old and modern cruise conditions

(xvi) Strong effect of the fuel sulphur content is ob-served particle hygroscopicity increases distinctly withincreasing FSC

(xvii) No significant effect of the HES at low andmid FSC level is observed compared to the combustorexit an increased particle hygroscopicity downstreamthe HES compared to the combustor is observed at highFSC indicating continuing transformation of S(IV) toS(VI) during the transfer through the HES

(xviii) CCN activation of combustion particles in-creases with increasing FSC as a result of larger par-ticle hygroscopicity additional effects of adsorbed or-ganic matter cannot be excluded

45 Gaseous organic fraction

(xix) More than 100 different NMVOCs (aliphatic andaromatic hydrocarbons carbonyls and acids) were iden-tified and quantified they account for 91 wt of thetotal detected compounds

(xx) A decrease in NMHC emissions from old tomodern conditions at fixed HES pressure stage and fuelsulphur content (FSC) is found

(xxi) A decrease of NMHC emissions with decreas-ing pressure in the HES at fixed combustor operationconditions and FSC is observed

(xxii) No clear influence on the NMHC emissionswith changing FSC at the same combustor operationcondition is observed

(xxiii) High EIrsquos for organic acids are found Pos-sible explanations are that organic compounds fixed atthe sampling line walls are oxidised by the hot exhaustgases or that the acids are fixed at the surface of thesoot particles which are also trapped in the samplingcartridges

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 475

46 Emission properties

Refer to WILSON et al (2004) and PETZOLD et al(2003a)

(xxiv) Emission properties regarding combustionaerosol mass number and size are determined by thecombustor the influence of the Hot End Simulator isweak

(xxv) The processing of exhaust gas and parti-cles during the transfer through the Hot End Simula-tor increases the particle hygroscopicity and formationstrength of condensation particles both effects are re-lated to gaseous sulphuric acid

The experimental overview of the EU PartEmisproject presents the first comprehensive aerosol and gasmeasurements to be conducted on an aircraft enginecombustor test-rig The database will be used for fur-ther analyses as well as for plume modelling studies ofthe effect of aviation-related particle emissions on theglobal atmosphere and climate

Acknowledgements

The PartEmis project was funded by the European Com-mission under contract no G4RD-CT-2000-00207 Theauthors are very grateful to the test rig operation crew atQinetiQ for their very strong support during the experi-ments

References

ANDERSON BE WR COFER DR BAGWELL JWBARRICK CH HUDGINS KE BRUNKE 1998 Air-borne observations of aircraft aerosol emissions I Totalnonvolatile particle emission indices ndash Geophys Res Lett25 1689ndash1692

BOUCHER O 1999 Influence of air traffic on cirrus occur-rence Nature ndash 397 30ndash31

BROCK CA F SCHROumlDER B KAumlRCHER A PETZOLDR BUSEN M FIEBIG 2000 Ultrafine particle size distri-butions measured in aircraft exhaust plumes ndash J GeophysRes 105 26555ndash26567

CACHIER H MP BREMOND P BIAT-MEacuteNARD 1989Determination of atmospheric soot carbon with a simplethermal method ndash Tellus 41B 379ndash390

DEMOTT PJ Y CHEN SM KREIDENWEIS DCROGERS DE SHERMAN 1999 Ice formation by blackcarbon particles ndash Geophys Res Lett 26 2429ndash2432

GIEBL H A BERNER G REISCHL H PUXBAUM AKASPER-GIEBL R HITZENBERGER 2002 CCN activa-tion of oxalic and malonic acid test aerosols with the Uni-versity of Vienna cloud condensation nuclei counter ndash JAerosol Sci 33 1623ndash1634

GYSEL M E WEINGARTNER U BALTENSPERGER2002 Hygroscopicity of aerosol particles at low tempera-tures 2 Theoretical and experimental hygroscopic proper-ties of laboratory generated aerosols ndash Environ Sci Tech-nol 36 63ndash68

GYSEL M S NYEKI E WEINGARTNER U BAL-TENSPERGER H GIEBL R HITZENBERGER A PET-ZOLD CW WILSON 2003a Properties of jet engine com-bustor particles during the PartEmis experiment Hygro-scopicity at subsaturated conditions ndash Geophys Res Lett30 1566 DOI1010292003GL016896

GYSEL M S NYEKI E WEINGARTNER U BAL-TENSPERGER H GIEBL R HITZENBERGER A PET-ZOLD CW WILSON 2003b Jet engine combustion par-ticle hygroscopicity under subsaturated conditions duringPartEmis ndash J Aerosol Sci 34 (Suppl 2) Abstracts of theEuropean Aerosol Conference S1415ndashS1416

HAGEN DE PD WHITEFIELD H SCHLAGER 1996Particle emissions in the exhaust plume from commercialjet aircraft under cruise conditions ndash J Geophys Res 10119551ndash19557

HAGEN DE PD WHITEFIELD J PALADINO M TRUE-BLOOD H LILENFELD 1998 Particle sizing and emissionindices for a jet engine exhaust sampled at cruise ndash Geo-phys Res Lett 25 1681ndash1684

HAGEN DE PD WHITEFIELD JD PALADINO OSCHMID 2001 Comparative study of jet engine com-bustion emissions and engine operating conditions ndashIn SCHUMANN U G AMANATIDIS (Eds) Aviationaerosols contrails and cirrus clouds Air pollution researchreport 74 106ndash110 European Commission Brussels

HAVERKAMP H S WILHELM A SOROKIN F ARNOLD2004 Positive and negative ion measurements in jet aircraftengine exhaust concentrations sizes and implications foraerosol formation ndash Atmos Environ 38 2879ndash2884

HITZENBERGER R H GIEBL A PETZOLD M GYSELS NYEKI E WEINGARTNER U BALTENSPERGERCW WILSON 2003a Properties of jet engine combus-tor particles during the PartEmis experiment Hygroscopicproperties at supersaturated conditions ndash Geophys ResLett 30 1779 DOI1010292003GL017294

HITZENBERGER R H GIEBL A PETZOILD M GYSELS NYEKI E WEINGARTNER U BALTENSPERGERCW WILSON 2003b CCN activation of jet engine com-bustor particles during PartEmis ndash J Aerosol Sci 34(Suppl 2) Abstracts of the European Aerosol ConferenceS1409ndashS1410

HURLEY CD 1996 AEROTRACE Measurements of Par-ticulates from an Engine Combustor ndash Proceedings Impactof Aircraft Emissions upon the Atmosphere Vol 1 113 In-ternational Colloquium Paris October 1996

KATRAGKOU E S WILHELM F ARNOLD CW WIL-SON 2004 First gaseous Sulfur (VI) measurements in thesimulated internal flow of an aircraft gas turbine engineduring project PartEmis ndash Geophys Res Lett 31 2117DOI1010292003GL018231

NYEKI S M GYSEL E WEINGARTNER U BAL-TENSPERGER R HITZENBERGER A PETZOLD CWWILSON 2003 Volatile properties of jet engine combus-tor particles during PartEmis ndash J Aerosol Sci 34 (Suppl2) Abstracts of the European Aerosol Conference S1411ndashS1412

PETZOLD A FP SCHROumlDER 1998 Jet engine exhaustaerosol characterization ndash Aerosol Sci Technol 28 62ndash76

PETZOLD A A DOumlPELHEUER CA BROCK FPSCHROumlDER 1999 In situ observations and model calcula-

476 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

tions of black carbon emission by aircraft at cruise altitudendash J Geophys Res 104 22171ndash22181

PETZOLD A C STEIN S NYEKI M GYSEL EWEINGARTNER U BALTENSPERGER H GIEBL RHITZENBERGER A DOumlPELHEUER S VRCHOTICKY HPUXBAUM M JOHNSON CD HURLEY R MARSHCW WILSON 2003a Properties of jet engine combus-tor particles during the PartEmis experiment Microphysi-cal and chemical properties ndash Geophys Res Lett 30 1719DOI1010292003GL017283

PETZOLD A CW WILSON U BALTENSPERGERM FIEBIG L FRITZSCHE H GIEBL M GYSELR HITZENBERGER CD HURLEY S NYEKI HPUXBAUM U SCHUMANN C STEIN S VROCHTICKY2003b Particle emissions from aircraft engines ndash anoverview of the European project PartEmis ndash J AerosolSci 34 (Suppl 2) Abstracts of the European Aerosol Con-ference S1405ndashS1406

PETZOLD A AE VRTALA M FIEBIG F FRITZSCHEPH MIRABEL S NYEKI X VANCASSEL CW WIL-SON 2003c Emission of volatile and non-volatile ultra-fine particles from a combustion source during PartEmisndash J Aerosol Sci 34 (Suppl 2) Abstracts of the EuropeanAerosol Conference S1407ndashS1408

PUXBAUM H S VRCHOTICKY A PETZOLD R HITZEN-BERGER S NYEKI CW WILSON 2003 Chemical com-position of jet engine combustor particles during PartEmisndash J Aerosol Sci 34 (Suppl 2) Abstracts of the EuropeanAerosol Conference S1413ndashS1414

SCHMID H L LASKUS H-J ABRAHAM U BAL-TENSPERGER V LAVANCHY M BIZJAK P BURBA HCACHIER D CROW J CHOW T GNAUK A EVENHM TEN BRINK K-P GIESEN R HITZENBERGER CHUEGLIN W MAENHAUT C PIOA CARVALHO J-PPUTAUD D TOOM-SAUNTRY H PUXBAUM 2001 Re-sults of the ldquocarbon conferencerdquo international aerosol car-bon round robin test stage I ndash Atmos Environ 35 2111ndash2121

SCHROumlDER F C BROCK R BAUMANN A PETZOLD RBUSEN P SCHULTE M FIEBIG 2000 In-situ studies onvolatile jet exhaust particle emissions Impact of fuel sul-fur content and thermodynamic conditions on nuclei-modeaerosols ndash J Geophys Res 105 19941ndash19954

SCHUMANN U 1996 Contrail formation from aircraft ex-hausts ndash Meteorol Z NF 5 4ndash23

SCHUMANN U J STROumlM 2001 Aviation impact on atmo-spheric composition and climate ndash In European research inthe stratosphere 1996ndash2000 European Commission EUR19867 Brussels

SCHUMANN U F ARNOLD R BUSEN J CURTIUS BKAumlRCHER A KIENDLER A PETZOLD H SCHLAGERF SCHROumlDER K-H WOHLFROM 2002 Influence of fu-els sulphur on the composition of aircraft exhaust plumesThe experiments SULFUR 1-7 ndash J Geophys Res 107(1010292001JD000813) AAC 2-1 ndash AAC 2-27

SOROKIN A P MIRABEL F ARNOLD 2003 On the In-fluence of fuel fulfur induced stable negative ion formationon the total concentration of ions emitted by an aircraft gasturbine engine Comparison of model and experiment ndash At-mos Chem Phys 3 6001ndash6018

VANCASSEL X A SOROKIN P MIRABEL A PETZOLDCW WILSON 2004 Volatile particles formation duringPartEmis a modelling study ndash Atmos Chem Phys 4 439ndash447

WILHELM S H HAVERKAMP A SOROKIN F ARNOLD2004 Detection of very large ions in aircraft gas turbineengine combustor exhaustcharged small soot particles ndashAtmos Environ 38 4561ndash4569

WILSON CW A PETZOLD S NYEKI U SCHUMANNR ZELLNER 2004 Measurement and Prediction of Emis-sions of Aerosols and Gaseous Precursors from Gas Tur-bine Engines (PartEmis) An Overview ndash Aerosp SciTechnol 8 131ndash143

Page 4: Particle emissions from aircraft engines a survey of the ... · 466 A. Petzold et al.: Particle emissions from aircraft engines Meteorol. Z., 14 , 2005 1 Introduction The role of

468 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

distributions were measured with a commercial Scan-ning Mobility Particle Sizer (SMPS TSI Model 3071)and a Volatility-Tandem Differential Mobility Analyser(V-TDMA) in the size range Dgt 13 nm

The aerosol mixing state and the volatile fractionof the combustion aerosol were determined using ther-mal denuder methods which differentiate particles bynumber according to their thermal stability As wasinvestigated in previous calibration experiments ther-mal denuder temperatures of 390 K and 573ndash625 Krespectively distinguish volatile (vaporises at 390 K)sulphuric-acid like material from semivolatile (vaporisesbetween 390 and 573 K) and refractory black-carbonlike material The thermal treatment of almost monodis-perse particle size fractions which were pre-selected bya DMA provides information whether the aerosol isinternally or externally mixed The ratio of diametersafter and before thermal treatment yields the particleshrinkage factor D32D

30 This factor describes the par-

ticle volume reduction by the thermal treatment ie thevolatile volume fraction of internally mixed particles is1minusD3D3

0The hygroscopic particle growth factor under sub-

saturated conditions was measured with a Hygroscop-icity Tandem DMA (GYSEL et al 2002 2003a) Dur-ing the measurements the instrument was kept at a con-stant temperature of T sim= 298 K The obtained hygro-scopic growth factor describes the particle growth bywater uptake from the gas phase at a selected relativehumidity Measurements were performed in the rela-tive humidity range from 70 to approximately 95 As was demonstrated by GYSEL at al (2002 2003a)from KOEHLER theory the hygroscopic growth factorof a particle is a function of dry particle size relativehumidity and particle coating with a water-soluble sub-stance For the PartEmis experimental data the coatingsubstance was sulphuric acid The magnitude of the par-ticle growth at a given relative humidity and for a givensize of the dry particle can be directly linked to volumefraction of water-soluble matter contained in the inves-tigated particles During the PartEmis experiments thehygroscopic growth factors measured for fixed dry par-ticle sizes but variable fuel sulphur contents were usedfor the determination of the coating of the combustionaerosol by sulphuric acid The number of particles whichbecome finally activated as cloud condensation nuclei inthe case of water-supersaturation were measured with astatic thermal diffusion chamber CCN counter (GIEBL

et al 2002) The instrument was operated at a super-saturation ratio of 06ndash07 with respect to liquid waterThe activation ratio or fraction of CCN activated par-ticles respectively was determined as the ratio of thenumber of activated CCN to the number of total particles(HITZENBERGER et al 2003a) The ability of combus-tion particles for taking up water from the gas phase is a

key property in the formation processes for liquid waterclouds However it is also of relevance for jet conden-sation trail formation (SCHUMANN 1996) and it mayas well play a role in the formation of ice nuclei fromcombustion particles (DEMOTT et al 1999)

Aerosol samples were collected with filter stack sam-plers and low pressure Berner impactors for total aerosoland size-resolved chemical mass analysis (PUXBAUM etal 2003) The chemical composition of the carbona-ceous fraction of the exhaust particles was determinedusing multi-step combustion methods Total carbon (TC)was determined by a combustion technique involvingcombustion in a vertical furnace and detection of CO2with a commercial NDIR (non dispersive Infrared spec-trometer) unit Aliquots of a collected filter sample wereburnt in the ceramic tube of a combustion furnace withan excess of oxygen at a temperature of 1000C Ele-mental carbon (EC) was determined by the ldquoCachierrdquomethod (CACHIER et al 1989) ie another aliquot ofthe filter sample was treated for 2 h at 320C in an oxy-gen atmosphere to remove organic carbon compoundsand then analysed with the same TC analysis methodThe remaining carbon was attributed to elemental car-bon (EC) Thermograms of the collected carbonaceousmaterial were measured with a thermo-optical evolvedgas analysis method (SCHMID et al 2001) To deter-mine anions eg SO2minus

4 as well as cations two Dionexisocratic systems with electrochemical suppression wereused

In addition the aerosol was sampled on quartz tubescontaining glass wool The aerosol loaded quartz tubeswere analysed by 5 min thermodesorption in a He-lium atmosphere at 130C and 350C respectively Thedesorbed C2ndashC12 non-methane hydrocarbons (NMHC)were analysed with a gas chromatograph (GC HewlettPackard) NMHCs were monitored using a compact GCinstrument (Airmovoc 2010) with a cryotrap as an en-richment system and a flame ionisation detector C2ndashC10 aliphatic and aromatic NMHCswere measured witha time resolution of 10 min and with detection limits inthe pptv range Selected partially oxidised hydrocarbons(OHC) such as carbonyl compounds and organic acidswere collected by using commercial sampling cartridgescoated with 24-Dinitrophenylhydrazine This chemicalcompound forms with the carbonyl molecules a stablecompound called hydrazone This hydrazone is stablefor weeks and can be analysed later in the LaboratoryTo analyse the carbonyl compounds the hydrazones arerinsed out of the cartridge and separated in a High Per-formance Liquid Chromatograph HPLC The OHCweremeasured with a time resolution of 4 min and with de-tection limits in the ppbv range All hydrocarbon com-pound measurements were corrected for line losses andbackground values

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 469

Figure 2 Left size distribution of combustion aerosol with (diamonds) and without condensation particle formation (triangles) RightExternal aerosol mixture composed of volatile condensation particles (grey) and coated nonvolatile combustion particles (black)

Gaseous sulphuric acid was measured using a CIMS(Chemical Ionisation Mass Spectrometer) apparatusequipped with an Ion Trap Mass Spectrometer (ITMS)The CIMS apparatus comprises a stainless steel flow re-actor (inner diameter 4 cm total length 164 cm) heatedup to 80C a high frequency glow discharge ion sourcefor reagent ion production and a Quadrupole Ion TrapMass Spectrometer for ion detection The ITMS allowsCID (Collision-Induced-Dissociation) investigations ofmass selected ions to be performed which in turn allowsmuch better species identification than can be offeredby a conventional mass spectrometer Reactant ions ofthe type NO3 (HNO3)n (n= 012) are produced in aseparate chamber where a high frequency gas dischargeburns in a buffer gas of O2 and NO2 N2 and injectedinto the flow reactor upstream of the ion sampling orificeof the Mass Spectrometer After injection ions are car-ried by the gas stream to the mass spectrometer Whengaseous sulphuric acid is present in the flow reactorsome fraction of the reagent ions reacts via a gas phaseion molecule reaction with sulphuric acid By measur-ing the abundance ratio of product and reagent ions thegaseous sulphuric acid concentration in the flow reactoris determined For more details see KATRAGKOU et al(2004) and references given there

OH laser induced fluorescence (LIF) measurementswere performed in the exhaust duct of the HES The ex-periment comprised a NdYAG pumped dye laser set upto generate light at λ = 281913 nm in order to excite theQ1(1) transition of OH A2Σ(ν prime = 1)larr X2Π(ν primeprime = 0)chosen as it is one of the strongest transitions at the tem-perature expected at the observation point Fluorescencewas detected perpendicular to the path of the laser fromthe ν prime = 0rarr vprimeprime = 0 transition around 309 nm via a lensand a customised interference filter In order to max-imise the signal to noise ratio by discriminating againstRaman scatter expected from N2 and H2O an interfer-ence filter was centred at 30875 nm with a band passof 496 nm A gated CCD camera cooled to ndash30C was

used to collect the signal allowing signal accumulationover long times in order to enhance sensitivity

Since the full set of applied instruments was con-nected to separate sampling lines according to specificinstrument requirements the precise determination ofthe dilution factors was a crucial task After exit fromthe traversing probe the sample was cooled to sim 150Cusing a water-jacket and delivered to the undiluted anddiluted sample lines The undiluted line was insulatedalong its entire length in order to avoid cooling andtherefore wall losses of less volatile combustion prod-ucts A detailed description of the extracting probe andthe sample transportation system is given in WILSON

et al (2004) Dilution with filtered air was necessaryfor the majority of instruments to reduce concentrationsdown to a measurable level The sample dilution factorwas determined from CO2 mixing ratios measured in thesupplied dilution air diluted sample and undiluted sam-ple as well as from temperature and pressure sensorsthroughout the sample delivery system The average di-lution factor was 65 (PETZOLD et al 2003a)

Although stable gaseous species such as CO2 andwater vapour are unaffected by intrusive samplingaerosols may be lost due to diffusion and sedimenta-tion to the surfaces of the sampling system Estimatesof expected line losses were obtained from experiments(HURLEY 1996) performed with aerosol generated bya spark discharge generator (PALAS GfG 1000) Theaerosol was fed into sample lines of different cross sec-tion and length and the sample flow through the lineswas varied From these studies expected line losses arelt 20 by number of particles for the conditions dur-ing the PartEmis experiments Since line losses affectpredominantly small particles lt 20 nm in diameter theaerosol mass is assumed to be almost unaffected by linelosses However the values given for emission indiceshave to be taken as upper values Size distributions re-ported for combustion particles cover a size range of 10ndash250 nm The size spectra used for the investigation of the

470 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

impact of operation conditions and fuels sulphur con-tent were not corrected for line losses because correc-tions are similar for all test points Thus relative trendsconcerning the influence of fuel sulphur and operationconditions on the microphysical properties of the emit-ted particles are not affected by line losses Becauseof the observed size range of exhaust particles countmedian diameters and geometric standard deviations ofmeasured size distributions are also expected to be onlymarginally influenced by line losses

3 Results

31 Microphysical properties

The aerosol emitted from the Combustor is composedof primary particles forming inside the combustor andof volatile condensation particles nucleating in the cool-ing exhaust gas from gaseous precursors The aerosolsize distribution in the diluted exhaust gas of the com-bustor is shown in Fig 2 for conditions with and with-out condensation particle formation While the mode ofcombustion aerosol particles dominates the size rangeD gt 15 nm with a maximum at 40 nm another modeof nanoparticles is present in all cases in the size rangeD= 7minus20 nm Condensation particle formation is oc-curring only at high FSC conditions and contributes tothe size range D= 4minus7 nm

The emission properties with respect to the primaryaerosol are shown in terms of BC mass (EIBC ) TCmass (EITC) and number (EIN10) emissions per kg ofconsumed fuel in Fig 3 The number emission indexEIN10 refers to particles with diameters D ge 10 nm InFig 3 also the count median diameter (CMD) of theexhaust aerosol size distribution is shown Plotted arecorresponding values for the combustor emissions andemissions from the HES low pressure stage In the firstplace it has to be mentioned that all the results are con-sistent At old cruise conditions an increasing numberemission at constant black carbon mass emission is bal-anced by a decreasing particle size with respect to theparticle electrical mobility diameter It should be notedthat these results originate from different measurementtechniques

When operated at old engine conditions with lowcombustor inlet temperature TC EIBC varies from 0051to 0081 g kgminus1 Operated at modern engine condi-tions with the higher inlet temperature EIBC ranges from0028 to 0044 g kgminus1 which is a factor of two lowerSimilar relations hold for the other properties How-ever the amount of emitted particulate matter is withinthe range permitted by the ICAO rules in all cases Allparticle emission indices referring to number and massare in close agreement with values obtained from in-flight measurements which span a range from 05ndash075

Figure 3 Emission indices in terms of number (N10 ) total car-

bon mass (TC) black carbon mass (BC) and count median diameter

(CMD) of the combustion aerosol for the Combustor exit (triangles)

and the Hot End Simulator low pressure stage (diamonds) values are

given for old and modern operation conditions as a function of the

fuel sulphur content error bars indicate signal variability in terms of

plusmn1 stdev

g kgminus1 for a Boeing B707 to 001 g kgminus1 for AirbusA 340 and Boeing B737 equipped with CFM 56 se-ries engines (SCHUMANN et al 2002 PETZOLD et al2003a 2003b) The fleet-averaged mass emission indexof 0038 g kgminus1 (PETZOLD et al 1999) lies within therange of values determined for modern operation condi-tions

Mean CMD values are 37ndash41 nm for old engine and40ndash44 nm for modern engine conditions The geomet-ric standard deviation of the size distribution is 167(16ndash18) for both operation conditions (PETZOLD etal 2003a 2003b) The combustor operating conditionshave only a weak influence on the size number and massof emitted particles Size distribution peaking at similardiameters were observed during several in-flight stud-ies HAGEN et al (1996 1998) reported peak diame-ters of 30ndash60 nm and size distribution widths of approx16 During the SULFUR studies (PETZOLD et al 1999SCHUMANN et al 2002) peak diameters of 35 nm wereobserved The PartEmis size distribution parameters arein close agreement with the reported in-flight observa-tions

The influence of the different HES pressure stages isdemonstrated in Fig 3 and with more detail in Fig 4As is shown in Fig 3 the trends observed in particleemission properties in the combustor experiments withvarying fuel sulphur content and operation conditionsagain observed during the HES experiments Except fora step-like increase from the combustor experiments tothe HES experiments which is most likely due to slightlydifferent operating conditions the HES has no influenceon the emission properties of the Combustor-HES com-bination with respect to combustion aerosol particles

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 471

0

5

10

15

20

25

30

000

005

010

015

020

high intermed low pressure

Combustor Hot End Simulator section

EIBC

EIN

EI

BC

g

kg

-1

EI

N

10

14kg

-1

Figure 4 Effect of the Hot End Simulator on the emission propertiesof the Combustor-HES combination operated at modern conditions

error bars indicate signal variability in terms of plusmn1 stdev

see Fig 4 Increasing the FSC slightly increases the to-tal number of emitted particles

Particles smaller than 10 nm contribute less than10 to the total exhaust aerosol in the low and mediumFSC cases At high FSC levels the formation of volatilecondensation particles with diameters Dlt 10 nm is ob-served (PETZOLD et al 2003c) A kinetic model wasused to investigate the nucleation efficiency of the H2O- H2SO4 binary mixture in the exhaust gas and the par-ticle coagulation processes during the transport throughthe sampling system to the measurement instrumenta-tion (VANCASSEL et al 2004) The model confirmedthat the condensation particles form at the dilutor of thesampling system During combustion the sulphur con-tained in the fuel is oxidised to SO2 which is then par-tially converted to S(VI) in the hot exhaust gas The ef-ficiency of conversion from S(IV) to S(VI) is describedby the conversion efficiency calculated as ε =S(VI)STwhere ST is the total fuel sulphur In the combustor ex-haust the S(VI) increases with increasing FSC In thekinetic model study a value for the efficiency was indi-rectly deduced from comparisons between model resultsand measurements In the PartEmis cases ε ranges be-tween roughly 25 and 6 depending on the com-bustor settings and on the microphysical approach usedFor these values which are close to results from di-rect measurements of gaseous sulphuric acid the kineticmodel reproduces the observation of volatile particles inthe size range 4ndash7 nm for the high FSC case while novolatile particles are found by the model for the low andmedium FSC cases

Particles of size D ge 30 nm are almost entirely in-ternally mixed Smaller particles (D lt 20 nm) exhibitvolume fractions of 10ndash15 for volatile and 4ndash10 for semi-volatile material while the volatile fraction de-creases to lt 5 for D ge 50 nm particles (PETZOLD

et al 2003a NYEKI et al 2003) The decrease of thevolatile particle fraction with increasing particle size isalso reproduced by the kinetic model for the H2O ndashH2SO4 binary mixture (VANCASSEL et al 2004) Thesevalues are valid for the conditions met during PartEmisIf the dilution air is more humid and a higher dilutionforces stronger formation of condensation particles thecondensation particles may grow to larger sizes How-ever as was observed in a young aircraft plume at cruiseconditions condensation particles were also observedmainly in the size range Dlt 10 nm (SCHROumlDER et al2000) similar to the PartEmis results

32 Chemical properties

Primary particles emitted from the combustor arepredominantly composed of carbonaceous material(gt95 ) and of water-soluble inorganic componentsThe carbonaceous material is composed of organic com-pounds and of elemental or black carbon The organiccompounds evaporate and oxidise at different tempera-tures and become classified according to this criterionCompounds which evaporate at temperatures as low as400 K are classified as semi-volatie OC compoundsevaporating compounds evaporating up to 800 K aretermed nonvolatile OC and the remainder is elementalcarbon Fig 5 shows the thermal stability of an exhaustaerosol sampled during operation with low FSC fuel atmodern conditions In the applied version of evolved gasanalysis (SCHMID et al 2001) the blackness of the filtersample is recorded parallel to the heating of the sampleAs is demonstrated in Fig 5 the filter starts to becomeldquowhiterdquo again as soon as the elemental or black carbonstarts to oxidise The evaporation and subsequent oxi-dation of OC compounds leaves the filter black Thusthe onset of increasing filter transmittance is an addi-tional measure for separating organic and elemental car-bon The application of this methods yields an averagecontribution of black carbon to total carbon of 40ndash80 depending on the combustion conditions (PUXBAUM etal 2003) From an aerosol loaded glass wool samplemore than 180 different NMHCs were desorbed underan Helium atmosphere Only 04 wt of the NMHCswere desorbed at T le 400 K whereas 996 wt weredesorbed at T = 400ndash620 K This is in good agreementwith the result of the thermogram in Fig 5 Elemental(EC) or black carbon (BC) respectively is chemicallyalmost inert at low and moderate temperatures and ther-mally stable up to temperatures of about 700 K

As determined from direct measurements of gaseousH2SO4 the efficiency of conversion in the combus-tor exhaust ranges between 09 plusmn 05 and 13 plusmn07 (KATRAGKOU et al 2004) In the HES ex-haust S(VI) increases with increasing FSC fuel flowand from the high-pressure to the low-pressure HESstage For FSC=1270 g kgminus1 the conversion efficiency

472 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

is 23plusmn 12 at the low pressure stage for the mod-ern cruise condition and 14plusmn 07 for the old cruiseThe higher ε observed for the modern cruise conditionssuggests that modern engines have a larger ε comparedto old engines Conversion efficiencies determined fromin-flight data at cruise altitude gave values 04ndash45 (SCHUMANN et al 2002) which is in good agreementwith the PartEmis ground test values

Chemical composition information was obtainedfrom Berner low pressure impactor samples which weretaken at a sample line temperature of 150C (PUXBAUMet al 2003) At these conditions most of the sulphuricacid remains in the gas phase while only a small fractionis more strongly bound to or chemi-sorbed at the par-ticle surface and will be observed in particulate mattersamples The chemical analysis results indicate that themass distribution of the carbonaceous matter matchesthe volume distribution of the aerosol The mass dis-tribution of chemisorbed sulphate however follows thesurface distribution of the aerosol which was calculatedfrom aerosol size spectra assuming a spherical particleshape (PUXBAUM et al 2003) This observation indi-cates that the sulphate is chemisorbed at the particle sur-face The chemisorbed fraction of sulphuric acid corre-sponds to a coverage of max 01 monolayer in the highFSC case This fraction of converted S(VI) is less than01 of total S(IV) present in the fuel The overall par-titioning of sulphur-containing species at 150C is 97 SO2le 27 gaseous H2SO4lt 03 chemisorbedAfter cooling to environmental temperatures the gasphase sulphuric acid can condense and cover the com-bustion particles with about one monolayer of sulphuricacid on average (HITZENBRGER et al 2003a)

33 Particle hygroscopic properties

Particle hygroscopic properties were measured at ambi-ent conditions using a Hygroscopicity Tandem Differ-ential Mobility Analyser (H-TDMA) system (GYSEL etal 2003a 2003b) The instrument measures the hygro-scopic growth factor (g) of particles with initial dry sizesDo = 30 50 and 100 nm over the range RHsim 70minus95 where g(RH) = D(RH)Do Growth factors were found toincrease with increasing FSC at fixed particle size (egDo = 50 nm g (95 ) = 101 and 116 for FSC = lowand high respectively) and to decrease with increasingparticle size at fixed FSC (eg mid FSC g (95 ) =114 110 and 103 for Do = 30 50 and 100 nm respec-tively) The results suggest an increasing amount of sul-phuric acid adsorbed on combustion particles from thegas phase with increasing FSC

Cloud condensation nuclei (CCN) were measured inthe diluted line using a static thermal diffusion typeCCN counter (GIEBL et al 2002) Particles are exposedto controlled supersaturations of water vapour whichenables some of the particles to become activated and

Figure 5 Evolving CO2 as a function of the sample heating temper-ature for a Combustor aerosol sample semi-volatile OC oxidises at

450 K non-volatile OC at 600 K and elemental carbon above 700 K

grow to large (gt 10 microm) droplets During PartEmis thesupersaturation was set tosim 07 The fraction of CCNin the aerosol (relative to the concentration of particlesof Dgt 10 nm) increases with increasing FSC but evenfor low FSC the fraction of CCN (at 0755 supersat-uration) is higher than the fraction of particles with di-ameters larger than 280 nm ie the size at which wet-table insoluble particles are activated according to theKelvin effect At high FSC the fraction of CCN in-creases by about a factor 55 for old and 12 for mod-ern conditions However pure sulphuric acid particlesof the same size which become activated according toKOumlHLER theory would give an activation ratio which isstill larger by a factorsim 15 compared to the observationseven at high FSC Thus the KOumlHLER model cannot re-produce measured activation properties This is in agree-ment with the fact that large combustion particles are notat all completely water soluble The Kelvin model pro-vides good predictions for the low FSC level but acti-vation diameters (ratios) are increasingly overestimated(underestimated) at medium and high FSC This indi-cates that the soluble fraction of the combustion parti-cles increases from low values at low FSC level to rel-atively large values at high FSC An empirical coatedsphere model which equates the measured volatile andsemivolatile fraction with soluble sulphuric acid under-estimates the activation diameter by 45 to 55 underall conditions A small amount of semivolatile organ-ics with low or absent water-solubility adsorbed on thecombustion particles might be one reason for this dis-crepancy Further details are given by HITZENBERGER

et al (2003a 2003b)

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 473

34 Gaseous hydrocarbon emission

More than 100 different non-methane volatile organiccompounds (NMVOCs) such as non-methane hydrocar-bons (NMHCs) and selected partially oxidised hydro-carbons (OHCs) emitted from the aircraft engine com-bustor and the hot end simulator (HES) were identifiedand quantified These species accounted for up to 91wt of the total hydrocarbons (THCs) The NMHCemission indices of 5ndash15 microg kgminus1 in total are compa-rable with the AEROTRACE (Measurement of TraceSpecies in the Exhaust from Aero Engines EC projectAERA-CT94-0003 1994ndash1997 Co-ordinated by Qine-tiQ) study Both studies show up to 1000 times loweremission indices in comparison with emissions from ve-hicular traffic The influence of the operation conditionson NMHC emissions of the combustor and HES was in-vestigated for the identified NMHCs From the six con-ditions which were investigated (2 operation conditions3 HES pressure stages) most compounds show a de-crease of NMHC emissions with increasing combustiontemperature at the same pressure stage This is in agree-ment with results from the combustor exit measurementsand with the results of the AEROTRACE study Emis-sion indices at HES pressure stages high (HP)- inter-mediate (IP)- and low pressure (LP) for fixed FSC andcombustor operation conditions were also investigatedMost compounds show a decrease of NMHC emissionswith decreasing pressure in the HES at similar combus-tor operation conditions and fixed FSC The decrease ofthe pressure in the HES corresponds to an increase ofthe residence time in a turbine and indicates a decreaseof NMHC emissions along the different turbine sections

35 Emission of ions and charged sootparticles

The ions measured during the HES campaign(HAVERKAMP et al 2004 WILHELM et al 2004) areproduct ions formed from primary ions via ion moleculereactions involving either few reaction-steps withmassive neutral molecules or numerous reaction-stepswith relatively small neutral molecules Their lifetimeswith respect to ion-ion recombination are around 9 msafter the high pressure turbine section The ion emissionindex Ei is up to 45 times 1016 ions per kg fuel burntAs is discussed by HAVERKAMP et al (2004) neitherion-ion recombination nor ion removal by attachmentto combustion particles give sufficient explanations forthe observed ion emission index Therefore detailedmodel calculations of the ion chemistry taking place inthe combustor with particular emphasis on the sulphurspecies effect is required (HAVERKAMP et al 2004)It seems that ion-induced nucleation can explain atleast partly the observed volatile aerosols Electricallycharged small positive and negative soot particles

(CSP) have also been detected during both PartEmiscampaigns Charged soot particles are found to havediameters around 6 nm and a total concentration of upto 45times107 cmminus3 (positive and negative) correspondingto a total emission index of ECSP = 23times 1015 CSP perkg fuel burnt (WILHELM et al 2004)

A model which considers the formation and evolu-tion of combustion ions in a combustor of an aircraftengine in dependence on the electron detachment ef-ficiency from negative ions was used to consider theeffect of the transformation of primary negative ionsto more stable secondary negative ions (SOROKIN etal 2003) The formed stable negative ions most prob-ably are sulphur-bearing ions This effect slows downthe charged particle neutralisation rate leading to an in-crease of the concentration of positive and negative ionsat the combustor exit The results of the simulation andtheir comparison with the ground-based experimentaldata obtained within PartEmis support the hypothesisthat the increase of the fuel sulphur content leads to anincrease of the ion concentration at the combustor exit

36 OH measurements

The concentration of OH which is the dominant oxidantin the exhaust gas was detected by laser induced fluo-rescence (LIF) to obtain an absolute OH concentrationmeasurement in order to aid in the modelling of SO3production from the combustor to the engine exit Whenrunning the laser at low laser energies OH in the exhaustduct was successfully detected Calibration experimentsprovide a method to assign the observed OH LIF sig-nals an absolute concentration and from an estimate ormeasurement of the errors for each parameter an overallestimate of the error in the assigned OH concentrationwas made From these calibration experiments very lowvalues of less than 1 ppbv were calculated for the OHconcentration in the exhaust gas after exiting the HESPrevious experiments using a combined Raman scatter-ingLIF technique to look in the exhaust gases 50 cmbehind an aircraft engine exit failed to detect any OHLIF signal and therefore could only infer an upper limitfor OH of 80 ppbv based on the measured detection limitof their apparatus

4 Summary and conclusions

The progress made during PartEmis is manifold Themain conclusions are therefore grouped into differentsubject areas and presented as brief statements For eachsubject area key references are identified if available

41 Combustion aerosol

Refer to WILSON et al (2004) and PETZOLD et al(2003a)

474 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

(i) The combustor behaves like a typical aircraft en-gine combustor with respect to thermodynamic dataand main emissions It may thus be expected that thePartEmis database can be applied as a first order approx-imation to contemporary aircraft engines

(ii) The BC concentration is nearly independent ofFSC level

(iii) The number size distribution of primary particlesin the diameter range 10 lt Dlt 550 nm is almost unin-fluenced by the FSC level or combustor conditions Athigh FSC levels the formation of volatile condensationparticles with diameters Dlt 10 nm occurs

(iv) The volatile volume fraction of particles gt 50 nmincreases from 2 for low FSC to approx 5 for highFSC The volatile volume fraction of particles lt 30 nmis le 15

(v) Volatility measurements indicate an internallymixed aerosol for particles with D gt 15 nm conden-sation particles are below 15 nm in diameter

42 Ultrafine particles

Refer to PETZOLD et al (2003a) and VANCASSEL et al(2004)

(vi) Formation of condensation particles of size D=4minus 7 nm is observed in the Combustor tests for highFSC at old and modern conditions formation occurs ifthe surface area of the combustion aerosol falls below acertain threshold limit the observed number of formedparticles is confirmed by kinetic model studies

(vii) Formation of condensation particles is observedin the HES tests for high FSCmodern conditionslowpressure section only the ldquoformation strengthrdquo in termsof number of condensation particles per number of com-bustion particle is twice as high as for the Combustorcase

(viii) The model studies indicate that condensationparticles are not emitted from the combustor but form inthe sampling line at the sample dilution point

(ix) Ultrafine carbonaceous particles of size 7 nmlt D lt 20 nm are observed independently of the fuelsulphur content a volatility analysis indicates that theseparticles are nonvolatile and may be composed of car-bonaceous compounds

43 Sulphate-containing species chemi-ionsand OH

Refer to KATRAGKOU et al (2004) and HAVERKAMP etal (2004)

(x) S(VI) was measured in the form of sulphuric acidIncreased FSC results in higher S(VI) concentrationsThe measured conversion efficiency for fuel sulphur toS(IV) at the combustor exit varies from 035ndash14 andis independent of the investigated combustor operationconditions and FSC levels

(xi) 10ndash30 of sulphate is chemisorbed on primaryparticles the coverage of primary particles is below 01monolayers of sulphuric acid at 150C and approxi-mately one monolayer at 25C

(xii) Sulphur partitioning at 150C 97 SO2le 27 gaseous H2SO4 lt 03 chemisorbed

(xiii) Positive and negative gaseous ions were de-tected in the exhaust of a combustor with mass numbersup to 1500 The total ion emission index at the HES hpstage is as high as 54times 1017 positive and negative ionsper kg fuel burnt Large ions with mass numbers exceed-ing 4500 contribute about 10 to the total ion numberconcentration

(xiv) OH concentration in the exhaust gas after exit-ing the HES is lt 1 ppbv

44 Hygroscopic growth and CCN activation

Refer to GYSEL et al (2003a) and HITZENBERGER etal (2003a)

(xv) Small differences in particle hygroscopicity areobserved between old and modern cruise conditions

(xvi) Strong effect of the fuel sulphur content is ob-served particle hygroscopicity increases distinctly withincreasing FSC

(xvii) No significant effect of the HES at low andmid FSC level is observed compared to the combustorexit an increased particle hygroscopicity downstreamthe HES compared to the combustor is observed at highFSC indicating continuing transformation of S(IV) toS(VI) during the transfer through the HES

(xviii) CCN activation of combustion particles in-creases with increasing FSC as a result of larger par-ticle hygroscopicity additional effects of adsorbed or-ganic matter cannot be excluded

45 Gaseous organic fraction

(xix) More than 100 different NMVOCs (aliphatic andaromatic hydrocarbons carbonyls and acids) were iden-tified and quantified they account for 91 wt of thetotal detected compounds

(xx) A decrease in NMHC emissions from old tomodern conditions at fixed HES pressure stage and fuelsulphur content (FSC) is found

(xxi) A decrease of NMHC emissions with decreas-ing pressure in the HES at fixed combustor operationconditions and FSC is observed

(xxii) No clear influence on the NMHC emissionswith changing FSC at the same combustor operationcondition is observed

(xxiii) High EIrsquos for organic acids are found Pos-sible explanations are that organic compounds fixed atthe sampling line walls are oxidised by the hot exhaustgases or that the acids are fixed at the surface of thesoot particles which are also trapped in the samplingcartridges

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 475

46 Emission properties

Refer to WILSON et al (2004) and PETZOLD et al(2003a)

(xxiv) Emission properties regarding combustionaerosol mass number and size are determined by thecombustor the influence of the Hot End Simulator isweak

(xxv) The processing of exhaust gas and parti-cles during the transfer through the Hot End Simula-tor increases the particle hygroscopicity and formationstrength of condensation particles both effects are re-lated to gaseous sulphuric acid

The experimental overview of the EU PartEmisproject presents the first comprehensive aerosol and gasmeasurements to be conducted on an aircraft enginecombustor test-rig The database will be used for fur-ther analyses as well as for plume modelling studies ofthe effect of aviation-related particle emissions on theglobal atmosphere and climate

Acknowledgements

The PartEmis project was funded by the European Com-mission under contract no G4RD-CT-2000-00207 Theauthors are very grateful to the test rig operation crew atQinetiQ for their very strong support during the experi-ments

References

ANDERSON BE WR COFER DR BAGWELL JWBARRICK CH HUDGINS KE BRUNKE 1998 Air-borne observations of aircraft aerosol emissions I Totalnonvolatile particle emission indices ndash Geophys Res Lett25 1689ndash1692

BOUCHER O 1999 Influence of air traffic on cirrus occur-rence Nature ndash 397 30ndash31

BROCK CA F SCHROumlDER B KAumlRCHER A PETZOLDR BUSEN M FIEBIG 2000 Ultrafine particle size distri-butions measured in aircraft exhaust plumes ndash J GeophysRes 105 26555ndash26567

CACHIER H MP BREMOND P BIAT-MEacuteNARD 1989Determination of atmospheric soot carbon with a simplethermal method ndash Tellus 41B 379ndash390

DEMOTT PJ Y CHEN SM KREIDENWEIS DCROGERS DE SHERMAN 1999 Ice formation by blackcarbon particles ndash Geophys Res Lett 26 2429ndash2432

GIEBL H A BERNER G REISCHL H PUXBAUM AKASPER-GIEBL R HITZENBERGER 2002 CCN activa-tion of oxalic and malonic acid test aerosols with the Uni-versity of Vienna cloud condensation nuclei counter ndash JAerosol Sci 33 1623ndash1634

GYSEL M E WEINGARTNER U BALTENSPERGER2002 Hygroscopicity of aerosol particles at low tempera-tures 2 Theoretical and experimental hygroscopic proper-ties of laboratory generated aerosols ndash Environ Sci Tech-nol 36 63ndash68

GYSEL M S NYEKI E WEINGARTNER U BAL-TENSPERGER H GIEBL R HITZENBERGER A PET-ZOLD CW WILSON 2003a Properties of jet engine com-bustor particles during the PartEmis experiment Hygro-scopicity at subsaturated conditions ndash Geophys Res Lett30 1566 DOI1010292003GL016896

GYSEL M S NYEKI E WEINGARTNER U BAL-TENSPERGER H GIEBL R HITZENBERGER A PET-ZOLD CW WILSON 2003b Jet engine combustion par-ticle hygroscopicity under subsaturated conditions duringPartEmis ndash J Aerosol Sci 34 (Suppl 2) Abstracts of theEuropean Aerosol Conference S1415ndashS1416

HAGEN DE PD WHITEFIELD H SCHLAGER 1996Particle emissions in the exhaust plume from commercialjet aircraft under cruise conditions ndash J Geophys Res 10119551ndash19557

HAGEN DE PD WHITEFIELD J PALADINO M TRUE-BLOOD H LILENFELD 1998 Particle sizing and emissionindices for a jet engine exhaust sampled at cruise ndash Geo-phys Res Lett 25 1681ndash1684

HAGEN DE PD WHITEFIELD JD PALADINO OSCHMID 2001 Comparative study of jet engine com-bustion emissions and engine operating conditions ndashIn SCHUMANN U G AMANATIDIS (Eds) Aviationaerosols contrails and cirrus clouds Air pollution researchreport 74 106ndash110 European Commission Brussels

HAVERKAMP H S WILHELM A SOROKIN F ARNOLD2004 Positive and negative ion measurements in jet aircraftengine exhaust concentrations sizes and implications foraerosol formation ndash Atmos Environ 38 2879ndash2884

HITZENBERGER R H GIEBL A PETZOLD M GYSELS NYEKI E WEINGARTNER U BALTENSPERGERCW WILSON 2003a Properties of jet engine combus-tor particles during the PartEmis experiment Hygroscopicproperties at supersaturated conditions ndash Geophys ResLett 30 1779 DOI1010292003GL017294

HITZENBERGER R H GIEBL A PETZOILD M GYSELS NYEKI E WEINGARTNER U BALTENSPERGERCW WILSON 2003b CCN activation of jet engine com-bustor particles during PartEmis ndash J Aerosol Sci 34(Suppl 2) Abstracts of the European Aerosol ConferenceS1409ndashS1410

HURLEY CD 1996 AEROTRACE Measurements of Par-ticulates from an Engine Combustor ndash Proceedings Impactof Aircraft Emissions upon the Atmosphere Vol 1 113 In-ternational Colloquium Paris October 1996

KATRAGKOU E S WILHELM F ARNOLD CW WIL-SON 2004 First gaseous Sulfur (VI) measurements in thesimulated internal flow of an aircraft gas turbine engineduring project PartEmis ndash Geophys Res Lett 31 2117DOI1010292003GL018231

NYEKI S M GYSEL E WEINGARTNER U BAL-TENSPERGER R HITZENBERGER A PETZOLD CWWILSON 2003 Volatile properties of jet engine combus-tor particles during PartEmis ndash J Aerosol Sci 34 (Suppl2) Abstracts of the European Aerosol Conference S1411ndashS1412

PETZOLD A FP SCHROumlDER 1998 Jet engine exhaustaerosol characterization ndash Aerosol Sci Technol 28 62ndash76

PETZOLD A A DOumlPELHEUER CA BROCK FPSCHROumlDER 1999 In situ observations and model calcula-

476 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

tions of black carbon emission by aircraft at cruise altitudendash J Geophys Res 104 22171ndash22181

PETZOLD A C STEIN S NYEKI M GYSEL EWEINGARTNER U BALTENSPERGER H GIEBL RHITZENBERGER A DOumlPELHEUER S VRCHOTICKY HPUXBAUM M JOHNSON CD HURLEY R MARSHCW WILSON 2003a Properties of jet engine combus-tor particles during the PartEmis experiment Microphysi-cal and chemical properties ndash Geophys Res Lett 30 1719DOI1010292003GL017283

PETZOLD A CW WILSON U BALTENSPERGERM FIEBIG L FRITZSCHE H GIEBL M GYSELR HITZENBERGER CD HURLEY S NYEKI HPUXBAUM U SCHUMANN C STEIN S VROCHTICKY2003b Particle emissions from aircraft engines ndash anoverview of the European project PartEmis ndash J AerosolSci 34 (Suppl 2) Abstracts of the European Aerosol Con-ference S1405ndashS1406

PETZOLD A AE VRTALA M FIEBIG F FRITZSCHEPH MIRABEL S NYEKI X VANCASSEL CW WIL-SON 2003c Emission of volatile and non-volatile ultra-fine particles from a combustion source during PartEmisndash J Aerosol Sci 34 (Suppl 2) Abstracts of the EuropeanAerosol Conference S1407ndashS1408

PUXBAUM H S VRCHOTICKY A PETZOLD R HITZEN-BERGER S NYEKI CW WILSON 2003 Chemical com-position of jet engine combustor particles during PartEmisndash J Aerosol Sci 34 (Suppl 2) Abstracts of the EuropeanAerosol Conference S1413ndashS1414

SCHMID H L LASKUS H-J ABRAHAM U BAL-TENSPERGER V LAVANCHY M BIZJAK P BURBA HCACHIER D CROW J CHOW T GNAUK A EVENHM TEN BRINK K-P GIESEN R HITZENBERGER CHUEGLIN W MAENHAUT C PIOA CARVALHO J-PPUTAUD D TOOM-SAUNTRY H PUXBAUM 2001 Re-sults of the ldquocarbon conferencerdquo international aerosol car-bon round robin test stage I ndash Atmos Environ 35 2111ndash2121

SCHROumlDER F C BROCK R BAUMANN A PETZOLD RBUSEN P SCHULTE M FIEBIG 2000 In-situ studies onvolatile jet exhaust particle emissions Impact of fuel sul-fur content and thermodynamic conditions on nuclei-modeaerosols ndash J Geophys Res 105 19941ndash19954

SCHUMANN U 1996 Contrail formation from aircraft ex-hausts ndash Meteorol Z NF 5 4ndash23

SCHUMANN U J STROumlM 2001 Aviation impact on atmo-spheric composition and climate ndash In European research inthe stratosphere 1996ndash2000 European Commission EUR19867 Brussels

SCHUMANN U F ARNOLD R BUSEN J CURTIUS BKAumlRCHER A KIENDLER A PETZOLD H SCHLAGERF SCHROumlDER K-H WOHLFROM 2002 Influence of fu-els sulphur on the composition of aircraft exhaust plumesThe experiments SULFUR 1-7 ndash J Geophys Res 107(1010292001JD000813) AAC 2-1 ndash AAC 2-27

SOROKIN A P MIRABEL F ARNOLD 2003 On the In-fluence of fuel fulfur induced stable negative ion formationon the total concentration of ions emitted by an aircraft gasturbine engine Comparison of model and experiment ndash At-mos Chem Phys 3 6001ndash6018

VANCASSEL X A SOROKIN P MIRABEL A PETZOLDCW WILSON 2004 Volatile particles formation duringPartEmis a modelling study ndash Atmos Chem Phys 4 439ndash447

WILHELM S H HAVERKAMP A SOROKIN F ARNOLD2004 Detection of very large ions in aircraft gas turbineengine combustor exhaustcharged small soot particles ndashAtmos Environ 38 4561ndash4569

WILSON CW A PETZOLD S NYEKI U SCHUMANNR ZELLNER 2004 Measurement and Prediction of Emis-sions of Aerosols and Gaseous Precursors from Gas Tur-bine Engines (PartEmis) An Overview ndash Aerosp SciTechnol 8 131ndash143

Page 5: Particle emissions from aircraft engines a survey of the ... · 466 A. Petzold et al.: Particle emissions from aircraft engines Meteorol. Z., 14 , 2005 1 Introduction The role of

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 469

Figure 2 Left size distribution of combustion aerosol with (diamonds) and without condensation particle formation (triangles) RightExternal aerosol mixture composed of volatile condensation particles (grey) and coated nonvolatile combustion particles (black)

Gaseous sulphuric acid was measured using a CIMS(Chemical Ionisation Mass Spectrometer) apparatusequipped with an Ion Trap Mass Spectrometer (ITMS)The CIMS apparatus comprises a stainless steel flow re-actor (inner diameter 4 cm total length 164 cm) heatedup to 80C a high frequency glow discharge ion sourcefor reagent ion production and a Quadrupole Ion TrapMass Spectrometer for ion detection The ITMS allowsCID (Collision-Induced-Dissociation) investigations ofmass selected ions to be performed which in turn allowsmuch better species identification than can be offeredby a conventional mass spectrometer Reactant ions ofthe type NO3 (HNO3)n (n= 012) are produced in aseparate chamber where a high frequency gas dischargeburns in a buffer gas of O2 and NO2 N2 and injectedinto the flow reactor upstream of the ion sampling orificeof the Mass Spectrometer After injection ions are car-ried by the gas stream to the mass spectrometer Whengaseous sulphuric acid is present in the flow reactorsome fraction of the reagent ions reacts via a gas phaseion molecule reaction with sulphuric acid By measur-ing the abundance ratio of product and reagent ions thegaseous sulphuric acid concentration in the flow reactoris determined For more details see KATRAGKOU et al(2004) and references given there

OH laser induced fluorescence (LIF) measurementswere performed in the exhaust duct of the HES The ex-periment comprised a NdYAG pumped dye laser set upto generate light at λ = 281913 nm in order to excite theQ1(1) transition of OH A2Σ(ν prime = 1)larr X2Π(ν primeprime = 0)chosen as it is one of the strongest transitions at the tem-perature expected at the observation point Fluorescencewas detected perpendicular to the path of the laser fromthe ν prime = 0rarr vprimeprime = 0 transition around 309 nm via a lensand a customised interference filter In order to max-imise the signal to noise ratio by discriminating againstRaman scatter expected from N2 and H2O an interfer-ence filter was centred at 30875 nm with a band passof 496 nm A gated CCD camera cooled to ndash30C was

used to collect the signal allowing signal accumulationover long times in order to enhance sensitivity

Since the full set of applied instruments was con-nected to separate sampling lines according to specificinstrument requirements the precise determination ofthe dilution factors was a crucial task After exit fromthe traversing probe the sample was cooled to sim 150Cusing a water-jacket and delivered to the undiluted anddiluted sample lines The undiluted line was insulatedalong its entire length in order to avoid cooling andtherefore wall losses of less volatile combustion prod-ucts A detailed description of the extracting probe andthe sample transportation system is given in WILSON

et al (2004) Dilution with filtered air was necessaryfor the majority of instruments to reduce concentrationsdown to a measurable level The sample dilution factorwas determined from CO2 mixing ratios measured in thesupplied dilution air diluted sample and undiluted sam-ple as well as from temperature and pressure sensorsthroughout the sample delivery system The average di-lution factor was 65 (PETZOLD et al 2003a)

Although stable gaseous species such as CO2 andwater vapour are unaffected by intrusive samplingaerosols may be lost due to diffusion and sedimenta-tion to the surfaces of the sampling system Estimatesof expected line losses were obtained from experiments(HURLEY 1996) performed with aerosol generated bya spark discharge generator (PALAS GfG 1000) Theaerosol was fed into sample lines of different cross sec-tion and length and the sample flow through the lineswas varied From these studies expected line losses arelt 20 by number of particles for the conditions dur-ing the PartEmis experiments Since line losses affectpredominantly small particles lt 20 nm in diameter theaerosol mass is assumed to be almost unaffected by linelosses However the values given for emission indiceshave to be taken as upper values Size distributions re-ported for combustion particles cover a size range of 10ndash250 nm The size spectra used for the investigation of the

470 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

impact of operation conditions and fuels sulphur con-tent were not corrected for line losses because correc-tions are similar for all test points Thus relative trendsconcerning the influence of fuel sulphur and operationconditions on the microphysical properties of the emit-ted particles are not affected by line losses Becauseof the observed size range of exhaust particles countmedian diameters and geometric standard deviations ofmeasured size distributions are also expected to be onlymarginally influenced by line losses

3 Results

31 Microphysical properties

The aerosol emitted from the Combustor is composedof primary particles forming inside the combustor andof volatile condensation particles nucleating in the cool-ing exhaust gas from gaseous precursors The aerosolsize distribution in the diluted exhaust gas of the com-bustor is shown in Fig 2 for conditions with and with-out condensation particle formation While the mode ofcombustion aerosol particles dominates the size rangeD gt 15 nm with a maximum at 40 nm another modeof nanoparticles is present in all cases in the size rangeD= 7minus20 nm Condensation particle formation is oc-curring only at high FSC conditions and contributes tothe size range D= 4minus7 nm

The emission properties with respect to the primaryaerosol are shown in terms of BC mass (EIBC ) TCmass (EITC) and number (EIN10) emissions per kg ofconsumed fuel in Fig 3 The number emission indexEIN10 refers to particles with diameters D ge 10 nm InFig 3 also the count median diameter (CMD) of theexhaust aerosol size distribution is shown Plotted arecorresponding values for the combustor emissions andemissions from the HES low pressure stage In the firstplace it has to be mentioned that all the results are con-sistent At old cruise conditions an increasing numberemission at constant black carbon mass emission is bal-anced by a decreasing particle size with respect to theparticle electrical mobility diameter It should be notedthat these results originate from different measurementtechniques

When operated at old engine conditions with lowcombustor inlet temperature TC EIBC varies from 0051to 0081 g kgminus1 Operated at modern engine condi-tions with the higher inlet temperature EIBC ranges from0028 to 0044 g kgminus1 which is a factor of two lowerSimilar relations hold for the other properties How-ever the amount of emitted particulate matter is withinthe range permitted by the ICAO rules in all cases Allparticle emission indices referring to number and massare in close agreement with values obtained from in-flight measurements which span a range from 05ndash075

Figure 3 Emission indices in terms of number (N10 ) total car-

bon mass (TC) black carbon mass (BC) and count median diameter

(CMD) of the combustion aerosol for the Combustor exit (triangles)

and the Hot End Simulator low pressure stage (diamonds) values are

given for old and modern operation conditions as a function of the

fuel sulphur content error bars indicate signal variability in terms of

plusmn1 stdev

g kgminus1 for a Boeing B707 to 001 g kgminus1 for AirbusA 340 and Boeing B737 equipped with CFM 56 se-ries engines (SCHUMANN et al 2002 PETZOLD et al2003a 2003b) The fleet-averaged mass emission indexof 0038 g kgminus1 (PETZOLD et al 1999) lies within therange of values determined for modern operation condi-tions

Mean CMD values are 37ndash41 nm for old engine and40ndash44 nm for modern engine conditions The geomet-ric standard deviation of the size distribution is 167(16ndash18) for both operation conditions (PETZOLD etal 2003a 2003b) The combustor operating conditionshave only a weak influence on the size number and massof emitted particles Size distribution peaking at similardiameters were observed during several in-flight stud-ies HAGEN et al (1996 1998) reported peak diame-ters of 30ndash60 nm and size distribution widths of approx16 During the SULFUR studies (PETZOLD et al 1999SCHUMANN et al 2002) peak diameters of 35 nm wereobserved The PartEmis size distribution parameters arein close agreement with the reported in-flight observa-tions

The influence of the different HES pressure stages isdemonstrated in Fig 3 and with more detail in Fig 4As is shown in Fig 3 the trends observed in particleemission properties in the combustor experiments withvarying fuel sulphur content and operation conditionsagain observed during the HES experiments Except fora step-like increase from the combustor experiments tothe HES experiments which is most likely due to slightlydifferent operating conditions the HES has no influenceon the emission properties of the Combustor-HES com-bination with respect to combustion aerosol particles

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 471

0

5

10

15

20

25

30

000

005

010

015

020

high intermed low pressure

Combustor Hot End Simulator section

EIBC

EIN

EI

BC

g

kg

-1

EI

N

10

14kg

-1

Figure 4 Effect of the Hot End Simulator on the emission propertiesof the Combustor-HES combination operated at modern conditions

error bars indicate signal variability in terms of plusmn1 stdev

see Fig 4 Increasing the FSC slightly increases the to-tal number of emitted particles

Particles smaller than 10 nm contribute less than10 to the total exhaust aerosol in the low and mediumFSC cases At high FSC levels the formation of volatilecondensation particles with diameters Dlt 10 nm is ob-served (PETZOLD et al 2003c) A kinetic model wasused to investigate the nucleation efficiency of the H2O- H2SO4 binary mixture in the exhaust gas and the par-ticle coagulation processes during the transport throughthe sampling system to the measurement instrumenta-tion (VANCASSEL et al 2004) The model confirmedthat the condensation particles form at the dilutor of thesampling system During combustion the sulphur con-tained in the fuel is oxidised to SO2 which is then par-tially converted to S(VI) in the hot exhaust gas The ef-ficiency of conversion from S(IV) to S(VI) is describedby the conversion efficiency calculated as ε =S(VI)STwhere ST is the total fuel sulphur In the combustor ex-haust the S(VI) increases with increasing FSC In thekinetic model study a value for the efficiency was indi-rectly deduced from comparisons between model resultsand measurements In the PartEmis cases ε ranges be-tween roughly 25 and 6 depending on the com-bustor settings and on the microphysical approach usedFor these values which are close to results from di-rect measurements of gaseous sulphuric acid the kineticmodel reproduces the observation of volatile particles inthe size range 4ndash7 nm for the high FSC case while novolatile particles are found by the model for the low andmedium FSC cases

Particles of size D ge 30 nm are almost entirely in-ternally mixed Smaller particles (D lt 20 nm) exhibitvolume fractions of 10ndash15 for volatile and 4ndash10 for semi-volatile material while the volatile fraction de-creases to lt 5 for D ge 50 nm particles (PETZOLD

et al 2003a NYEKI et al 2003) The decrease of thevolatile particle fraction with increasing particle size isalso reproduced by the kinetic model for the H2O ndashH2SO4 binary mixture (VANCASSEL et al 2004) Thesevalues are valid for the conditions met during PartEmisIf the dilution air is more humid and a higher dilutionforces stronger formation of condensation particles thecondensation particles may grow to larger sizes How-ever as was observed in a young aircraft plume at cruiseconditions condensation particles were also observedmainly in the size range Dlt 10 nm (SCHROumlDER et al2000) similar to the PartEmis results

32 Chemical properties

Primary particles emitted from the combustor arepredominantly composed of carbonaceous material(gt95 ) and of water-soluble inorganic componentsThe carbonaceous material is composed of organic com-pounds and of elemental or black carbon The organiccompounds evaporate and oxidise at different tempera-tures and become classified according to this criterionCompounds which evaporate at temperatures as low as400 K are classified as semi-volatie OC compoundsevaporating compounds evaporating up to 800 K aretermed nonvolatile OC and the remainder is elementalcarbon Fig 5 shows the thermal stability of an exhaustaerosol sampled during operation with low FSC fuel atmodern conditions In the applied version of evolved gasanalysis (SCHMID et al 2001) the blackness of the filtersample is recorded parallel to the heating of the sampleAs is demonstrated in Fig 5 the filter starts to becomeldquowhiterdquo again as soon as the elemental or black carbonstarts to oxidise The evaporation and subsequent oxi-dation of OC compounds leaves the filter black Thusthe onset of increasing filter transmittance is an addi-tional measure for separating organic and elemental car-bon The application of this methods yields an averagecontribution of black carbon to total carbon of 40ndash80 depending on the combustion conditions (PUXBAUM etal 2003) From an aerosol loaded glass wool samplemore than 180 different NMHCs were desorbed underan Helium atmosphere Only 04 wt of the NMHCswere desorbed at T le 400 K whereas 996 wt weredesorbed at T = 400ndash620 K This is in good agreementwith the result of the thermogram in Fig 5 Elemental(EC) or black carbon (BC) respectively is chemicallyalmost inert at low and moderate temperatures and ther-mally stable up to temperatures of about 700 K

As determined from direct measurements of gaseousH2SO4 the efficiency of conversion in the combus-tor exhaust ranges between 09 plusmn 05 and 13 plusmn07 (KATRAGKOU et al 2004) In the HES ex-haust S(VI) increases with increasing FSC fuel flowand from the high-pressure to the low-pressure HESstage For FSC=1270 g kgminus1 the conversion efficiency

472 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

is 23plusmn 12 at the low pressure stage for the mod-ern cruise condition and 14plusmn 07 for the old cruiseThe higher ε observed for the modern cruise conditionssuggests that modern engines have a larger ε comparedto old engines Conversion efficiencies determined fromin-flight data at cruise altitude gave values 04ndash45 (SCHUMANN et al 2002) which is in good agreementwith the PartEmis ground test values

Chemical composition information was obtainedfrom Berner low pressure impactor samples which weretaken at a sample line temperature of 150C (PUXBAUMet al 2003) At these conditions most of the sulphuricacid remains in the gas phase while only a small fractionis more strongly bound to or chemi-sorbed at the par-ticle surface and will be observed in particulate mattersamples The chemical analysis results indicate that themass distribution of the carbonaceous matter matchesthe volume distribution of the aerosol The mass dis-tribution of chemisorbed sulphate however follows thesurface distribution of the aerosol which was calculatedfrom aerosol size spectra assuming a spherical particleshape (PUXBAUM et al 2003) This observation indi-cates that the sulphate is chemisorbed at the particle sur-face The chemisorbed fraction of sulphuric acid corre-sponds to a coverage of max 01 monolayer in the highFSC case This fraction of converted S(VI) is less than01 of total S(IV) present in the fuel The overall par-titioning of sulphur-containing species at 150C is 97 SO2le 27 gaseous H2SO4lt 03 chemisorbedAfter cooling to environmental temperatures the gasphase sulphuric acid can condense and cover the com-bustion particles with about one monolayer of sulphuricacid on average (HITZENBRGER et al 2003a)

33 Particle hygroscopic properties

Particle hygroscopic properties were measured at ambi-ent conditions using a Hygroscopicity Tandem Differ-ential Mobility Analyser (H-TDMA) system (GYSEL etal 2003a 2003b) The instrument measures the hygro-scopic growth factor (g) of particles with initial dry sizesDo = 30 50 and 100 nm over the range RHsim 70minus95 where g(RH) = D(RH)Do Growth factors were found toincrease with increasing FSC at fixed particle size (egDo = 50 nm g (95 ) = 101 and 116 for FSC = lowand high respectively) and to decrease with increasingparticle size at fixed FSC (eg mid FSC g (95 ) =114 110 and 103 for Do = 30 50 and 100 nm respec-tively) The results suggest an increasing amount of sul-phuric acid adsorbed on combustion particles from thegas phase with increasing FSC

Cloud condensation nuclei (CCN) were measured inthe diluted line using a static thermal diffusion typeCCN counter (GIEBL et al 2002) Particles are exposedto controlled supersaturations of water vapour whichenables some of the particles to become activated and

Figure 5 Evolving CO2 as a function of the sample heating temper-ature for a Combustor aerosol sample semi-volatile OC oxidises at

450 K non-volatile OC at 600 K and elemental carbon above 700 K

grow to large (gt 10 microm) droplets During PartEmis thesupersaturation was set tosim 07 The fraction of CCNin the aerosol (relative to the concentration of particlesof Dgt 10 nm) increases with increasing FSC but evenfor low FSC the fraction of CCN (at 0755 supersat-uration) is higher than the fraction of particles with di-ameters larger than 280 nm ie the size at which wet-table insoluble particles are activated according to theKelvin effect At high FSC the fraction of CCN in-creases by about a factor 55 for old and 12 for mod-ern conditions However pure sulphuric acid particlesof the same size which become activated according toKOumlHLER theory would give an activation ratio which isstill larger by a factorsim 15 compared to the observationseven at high FSC Thus the KOumlHLER model cannot re-produce measured activation properties This is in agree-ment with the fact that large combustion particles are notat all completely water soluble The Kelvin model pro-vides good predictions for the low FSC level but acti-vation diameters (ratios) are increasingly overestimated(underestimated) at medium and high FSC This indi-cates that the soluble fraction of the combustion parti-cles increases from low values at low FSC level to rel-atively large values at high FSC An empirical coatedsphere model which equates the measured volatile andsemivolatile fraction with soluble sulphuric acid under-estimates the activation diameter by 45 to 55 underall conditions A small amount of semivolatile organ-ics with low or absent water-solubility adsorbed on thecombustion particles might be one reason for this dis-crepancy Further details are given by HITZENBERGER

et al (2003a 2003b)

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 473

34 Gaseous hydrocarbon emission

More than 100 different non-methane volatile organiccompounds (NMVOCs) such as non-methane hydrocar-bons (NMHCs) and selected partially oxidised hydro-carbons (OHCs) emitted from the aircraft engine com-bustor and the hot end simulator (HES) were identifiedand quantified These species accounted for up to 91wt of the total hydrocarbons (THCs) The NMHCemission indices of 5ndash15 microg kgminus1 in total are compa-rable with the AEROTRACE (Measurement of TraceSpecies in the Exhaust from Aero Engines EC projectAERA-CT94-0003 1994ndash1997 Co-ordinated by Qine-tiQ) study Both studies show up to 1000 times loweremission indices in comparison with emissions from ve-hicular traffic The influence of the operation conditionson NMHC emissions of the combustor and HES was in-vestigated for the identified NMHCs From the six con-ditions which were investigated (2 operation conditions3 HES pressure stages) most compounds show a de-crease of NMHC emissions with increasing combustiontemperature at the same pressure stage This is in agree-ment with results from the combustor exit measurementsand with the results of the AEROTRACE study Emis-sion indices at HES pressure stages high (HP)- inter-mediate (IP)- and low pressure (LP) for fixed FSC andcombustor operation conditions were also investigatedMost compounds show a decrease of NMHC emissionswith decreasing pressure in the HES at similar combus-tor operation conditions and fixed FSC The decrease ofthe pressure in the HES corresponds to an increase ofthe residence time in a turbine and indicates a decreaseof NMHC emissions along the different turbine sections

35 Emission of ions and charged sootparticles

The ions measured during the HES campaign(HAVERKAMP et al 2004 WILHELM et al 2004) areproduct ions formed from primary ions via ion moleculereactions involving either few reaction-steps withmassive neutral molecules or numerous reaction-stepswith relatively small neutral molecules Their lifetimeswith respect to ion-ion recombination are around 9 msafter the high pressure turbine section The ion emissionindex Ei is up to 45 times 1016 ions per kg fuel burntAs is discussed by HAVERKAMP et al (2004) neitherion-ion recombination nor ion removal by attachmentto combustion particles give sufficient explanations forthe observed ion emission index Therefore detailedmodel calculations of the ion chemistry taking place inthe combustor with particular emphasis on the sulphurspecies effect is required (HAVERKAMP et al 2004)It seems that ion-induced nucleation can explain atleast partly the observed volatile aerosols Electricallycharged small positive and negative soot particles

(CSP) have also been detected during both PartEmiscampaigns Charged soot particles are found to havediameters around 6 nm and a total concentration of upto 45times107 cmminus3 (positive and negative) correspondingto a total emission index of ECSP = 23times 1015 CSP perkg fuel burnt (WILHELM et al 2004)

A model which considers the formation and evolu-tion of combustion ions in a combustor of an aircraftengine in dependence on the electron detachment ef-ficiency from negative ions was used to consider theeffect of the transformation of primary negative ionsto more stable secondary negative ions (SOROKIN etal 2003) The formed stable negative ions most prob-ably are sulphur-bearing ions This effect slows downthe charged particle neutralisation rate leading to an in-crease of the concentration of positive and negative ionsat the combustor exit The results of the simulation andtheir comparison with the ground-based experimentaldata obtained within PartEmis support the hypothesisthat the increase of the fuel sulphur content leads to anincrease of the ion concentration at the combustor exit

36 OH measurements

The concentration of OH which is the dominant oxidantin the exhaust gas was detected by laser induced fluo-rescence (LIF) to obtain an absolute OH concentrationmeasurement in order to aid in the modelling of SO3production from the combustor to the engine exit Whenrunning the laser at low laser energies OH in the exhaustduct was successfully detected Calibration experimentsprovide a method to assign the observed OH LIF sig-nals an absolute concentration and from an estimate ormeasurement of the errors for each parameter an overallestimate of the error in the assigned OH concentrationwas made From these calibration experiments very lowvalues of less than 1 ppbv were calculated for the OHconcentration in the exhaust gas after exiting the HESPrevious experiments using a combined Raman scatter-ingLIF technique to look in the exhaust gases 50 cmbehind an aircraft engine exit failed to detect any OHLIF signal and therefore could only infer an upper limitfor OH of 80 ppbv based on the measured detection limitof their apparatus

4 Summary and conclusions

The progress made during PartEmis is manifold Themain conclusions are therefore grouped into differentsubject areas and presented as brief statements For eachsubject area key references are identified if available

41 Combustion aerosol

Refer to WILSON et al (2004) and PETZOLD et al(2003a)

474 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

(i) The combustor behaves like a typical aircraft en-gine combustor with respect to thermodynamic dataand main emissions It may thus be expected that thePartEmis database can be applied as a first order approx-imation to contemporary aircraft engines

(ii) The BC concentration is nearly independent ofFSC level

(iii) The number size distribution of primary particlesin the diameter range 10 lt Dlt 550 nm is almost unin-fluenced by the FSC level or combustor conditions Athigh FSC levels the formation of volatile condensationparticles with diameters Dlt 10 nm occurs

(iv) The volatile volume fraction of particles gt 50 nmincreases from 2 for low FSC to approx 5 for highFSC The volatile volume fraction of particles lt 30 nmis le 15

(v) Volatility measurements indicate an internallymixed aerosol for particles with D gt 15 nm conden-sation particles are below 15 nm in diameter

42 Ultrafine particles

Refer to PETZOLD et al (2003a) and VANCASSEL et al(2004)

(vi) Formation of condensation particles of size D=4minus 7 nm is observed in the Combustor tests for highFSC at old and modern conditions formation occurs ifthe surface area of the combustion aerosol falls below acertain threshold limit the observed number of formedparticles is confirmed by kinetic model studies

(vii) Formation of condensation particles is observedin the HES tests for high FSCmodern conditionslowpressure section only the ldquoformation strengthrdquo in termsof number of condensation particles per number of com-bustion particle is twice as high as for the Combustorcase

(viii) The model studies indicate that condensationparticles are not emitted from the combustor but form inthe sampling line at the sample dilution point

(ix) Ultrafine carbonaceous particles of size 7 nmlt D lt 20 nm are observed independently of the fuelsulphur content a volatility analysis indicates that theseparticles are nonvolatile and may be composed of car-bonaceous compounds

43 Sulphate-containing species chemi-ionsand OH

Refer to KATRAGKOU et al (2004) and HAVERKAMP etal (2004)

(x) S(VI) was measured in the form of sulphuric acidIncreased FSC results in higher S(VI) concentrationsThe measured conversion efficiency for fuel sulphur toS(IV) at the combustor exit varies from 035ndash14 andis independent of the investigated combustor operationconditions and FSC levels

(xi) 10ndash30 of sulphate is chemisorbed on primaryparticles the coverage of primary particles is below 01monolayers of sulphuric acid at 150C and approxi-mately one monolayer at 25C

(xii) Sulphur partitioning at 150C 97 SO2le 27 gaseous H2SO4 lt 03 chemisorbed

(xiii) Positive and negative gaseous ions were de-tected in the exhaust of a combustor with mass numbersup to 1500 The total ion emission index at the HES hpstage is as high as 54times 1017 positive and negative ionsper kg fuel burnt Large ions with mass numbers exceed-ing 4500 contribute about 10 to the total ion numberconcentration

(xiv) OH concentration in the exhaust gas after exit-ing the HES is lt 1 ppbv

44 Hygroscopic growth and CCN activation

Refer to GYSEL et al (2003a) and HITZENBERGER etal (2003a)

(xv) Small differences in particle hygroscopicity areobserved between old and modern cruise conditions

(xvi) Strong effect of the fuel sulphur content is ob-served particle hygroscopicity increases distinctly withincreasing FSC

(xvii) No significant effect of the HES at low andmid FSC level is observed compared to the combustorexit an increased particle hygroscopicity downstreamthe HES compared to the combustor is observed at highFSC indicating continuing transformation of S(IV) toS(VI) during the transfer through the HES

(xviii) CCN activation of combustion particles in-creases with increasing FSC as a result of larger par-ticle hygroscopicity additional effects of adsorbed or-ganic matter cannot be excluded

45 Gaseous organic fraction

(xix) More than 100 different NMVOCs (aliphatic andaromatic hydrocarbons carbonyls and acids) were iden-tified and quantified they account for 91 wt of thetotal detected compounds

(xx) A decrease in NMHC emissions from old tomodern conditions at fixed HES pressure stage and fuelsulphur content (FSC) is found

(xxi) A decrease of NMHC emissions with decreas-ing pressure in the HES at fixed combustor operationconditions and FSC is observed

(xxii) No clear influence on the NMHC emissionswith changing FSC at the same combustor operationcondition is observed

(xxiii) High EIrsquos for organic acids are found Pos-sible explanations are that organic compounds fixed atthe sampling line walls are oxidised by the hot exhaustgases or that the acids are fixed at the surface of thesoot particles which are also trapped in the samplingcartridges

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 475

46 Emission properties

Refer to WILSON et al (2004) and PETZOLD et al(2003a)

(xxiv) Emission properties regarding combustionaerosol mass number and size are determined by thecombustor the influence of the Hot End Simulator isweak

(xxv) The processing of exhaust gas and parti-cles during the transfer through the Hot End Simula-tor increases the particle hygroscopicity and formationstrength of condensation particles both effects are re-lated to gaseous sulphuric acid

The experimental overview of the EU PartEmisproject presents the first comprehensive aerosol and gasmeasurements to be conducted on an aircraft enginecombustor test-rig The database will be used for fur-ther analyses as well as for plume modelling studies ofthe effect of aviation-related particle emissions on theglobal atmosphere and climate

Acknowledgements

The PartEmis project was funded by the European Com-mission under contract no G4RD-CT-2000-00207 Theauthors are very grateful to the test rig operation crew atQinetiQ for their very strong support during the experi-ments

References

ANDERSON BE WR COFER DR BAGWELL JWBARRICK CH HUDGINS KE BRUNKE 1998 Air-borne observations of aircraft aerosol emissions I Totalnonvolatile particle emission indices ndash Geophys Res Lett25 1689ndash1692

BOUCHER O 1999 Influence of air traffic on cirrus occur-rence Nature ndash 397 30ndash31

BROCK CA F SCHROumlDER B KAumlRCHER A PETZOLDR BUSEN M FIEBIG 2000 Ultrafine particle size distri-butions measured in aircraft exhaust plumes ndash J GeophysRes 105 26555ndash26567

CACHIER H MP BREMOND P BIAT-MEacuteNARD 1989Determination of atmospheric soot carbon with a simplethermal method ndash Tellus 41B 379ndash390

DEMOTT PJ Y CHEN SM KREIDENWEIS DCROGERS DE SHERMAN 1999 Ice formation by blackcarbon particles ndash Geophys Res Lett 26 2429ndash2432

GIEBL H A BERNER G REISCHL H PUXBAUM AKASPER-GIEBL R HITZENBERGER 2002 CCN activa-tion of oxalic and malonic acid test aerosols with the Uni-versity of Vienna cloud condensation nuclei counter ndash JAerosol Sci 33 1623ndash1634

GYSEL M E WEINGARTNER U BALTENSPERGER2002 Hygroscopicity of aerosol particles at low tempera-tures 2 Theoretical and experimental hygroscopic proper-ties of laboratory generated aerosols ndash Environ Sci Tech-nol 36 63ndash68

GYSEL M S NYEKI E WEINGARTNER U BAL-TENSPERGER H GIEBL R HITZENBERGER A PET-ZOLD CW WILSON 2003a Properties of jet engine com-bustor particles during the PartEmis experiment Hygro-scopicity at subsaturated conditions ndash Geophys Res Lett30 1566 DOI1010292003GL016896

GYSEL M S NYEKI E WEINGARTNER U BAL-TENSPERGER H GIEBL R HITZENBERGER A PET-ZOLD CW WILSON 2003b Jet engine combustion par-ticle hygroscopicity under subsaturated conditions duringPartEmis ndash J Aerosol Sci 34 (Suppl 2) Abstracts of theEuropean Aerosol Conference S1415ndashS1416

HAGEN DE PD WHITEFIELD H SCHLAGER 1996Particle emissions in the exhaust plume from commercialjet aircraft under cruise conditions ndash J Geophys Res 10119551ndash19557

HAGEN DE PD WHITEFIELD J PALADINO M TRUE-BLOOD H LILENFELD 1998 Particle sizing and emissionindices for a jet engine exhaust sampled at cruise ndash Geo-phys Res Lett 25 1681ndash1684

HAGEN DE PD WHITEFIELD JD PALADINO OSCHMID 2001 Comparative study of jet engine com-bustion emissions and engine operating conditions ndashIn SCHUMANN U G AMANATIDIS (Eds) Aviationaerosols contrails and cirrus clouds Air pollution researchreport 74 106ndash110 European Commission Brussels

HAVERKAMP H S WILHELM A SOROKIN F ARNOLD2004 Positive and negative ion measurements in jet aircraftengine exhaust concentrations sizes and implications foraerosol formation ndash Atmos Environ 38 2879ndash2884

HITZENBERGER R H GIEBL A PETZOLD M GYSELS NYEKI E WEINGARTNER U BALTENSPERGERCW WILSON 2003a Properties of jet engine combus-tor particles during the PartEmis experiment Hygroscopicproperties at supersaturated conditions ndash Geophys ResLett 30 1779 DOI1010292003GL017294

HITZENBERGER R H GIEBL A PETZOILD M GYSELS NYEKI E WEINGARTNER U BALTENSPERGERCW WILSON 2003b CCN activation of jet engine com-bustor particles during PartEmis ndash J Aerosol Sci 34(Suppl 2) Abstracts of the European Aerosol ConferenceS1409ndashS1410

HURLEY CD 1996 AEROTRACE Measurements of Par-ticulates from an Engine Combustor ndash Proceedings Impactof Aircraft Emissions upon the Atmosphere Vol 1 113 In-ternational Colloquium Paris October 1996

KATRAGKOU E S WILHELM F ARNOLD CW WIL-SON 2004 First gaseous Sulfur (VI) measurements in thesimulated internal flow of an aircraft gas turbine engineduring project PartEmis ndash Geophys Res Lett 31 2117DOI1010292003GL018231

NYEKI S M GYSEL E WEINGARTNER U BAL-TENSPERGER R HITZENBERGER A PETZOLD CWWILSON 2003 Volatile properties of jet engine combus-tor particles during PartEmis ndash J Aerosol Sci 34 (Suppl2) Abstracts of the European Aerosol Conference S1411ndashS1412

PETZOLD A FP SCHROumlDER 1998 Jet engine exhaustaerosol characterization ndash Aerosol Sci Technol 28 62ndash76

PETZOLD A A DOumlPELHEUER CA BROCK FPSCHROumlDER 1999 In situ observations and model calcula-

476 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

tions of black carbon emission by aircraft at cruise altitudendash J Geophys Res 104 22171ndash22181

PETZOLD A C STEIN S NYEKI M GYSEL EWEINGARTNER U BALTENSPERGER H GIEBL RHITZENBERGER A DOumlPELHEUER S VRCHOTICKY HPUXBAUM M JOHNSON CD HURLEY R MARSHCW WILSON 2003a Properties of jet engine combus-tor particles during the PartEmis experiment Microphysi-cal and chemical properties ndash Geophys Res Lett 30 1719DOI1010292003GL017283

PETZOLD A CW WILSON U BALTENSPERGERM FIEBIG L FRITZSCHE H GIEBL M GYSELR HITZENBERGER CD HURLEY S NYEKI HPUXBAUM U SCHUMANN C STEIN S VROCHTICKY2003b Particle emissions from aircraft engines ndash anoverview of the European project PartEmis ndash J AerosolSci 34 (Suppl 2) Abstracts of the European Aerosol Con-ference S1405ndashS1406

PETZOLD A AE VRTALA M FIEBIG F FRITZSCHEPH MIRABEL S NYEKI X VANCASSEL CW WIL-SON 2003c Emission of volatile and non-volatile ultra-fine particles from a combustion source during PartEmisndash J Aerosol Sci 34 (Suppl 2) Abstracts of the EuropeanAerosol Conference S1407ndashS1408

PUXBAUM H S VRCHOTICKY A PETZOLD R HITZEN-BERGER S NYEKI CW WILSON 2003 Chemical com-position of jet engine combustor particles during PartEmisndash J Aerosol Sci 34 (Suppl 2) Abstracts of the EuropeanAerosol Conference S1413ndashS1414

SCHMID H L LASKUS H-J ABRAHAM U BAL-TENSPERGER V LAVANCHY M BIZJAK P BURBA HCACHIER D CROW J CHOW T GNAUK A EVENHM TEN BRINK K-P GIESEN R HITZENBERGER CHUEGLIN W MAENHAUT C PIOA CARVALHO J-PPUTAUD D TOOM-SAUNTRY H PUXBAUM 2001 Re-sults of the ldquocarbon conferencerdquo international aerosol car-bon round robin test stage I ndash Atmos Environ 35 2111ndash2121

SCHROumlDER F C BROCK R BAUMANN A PETZOLD RBUSEN P SCHULTE M FIEBIG 2000 In-situ studies onvolatile jet exhaust particle emissions Impact of fuel sul-fur content and thermodynamic conditions on nuclei-modeaerosols ndash J Geophys Res 105 19941ndash19954

SCHUMANN U 1996 Contrail formation from aircraft ex-hausts ndash Meteorol Z NF 5 4ndash23

SCHUMANN U J STROumlM 2001 Aviation impact on atmo-spheric composition and climate ndash In European research inthe stratosphere 1996ndash2000 European Commission EUR19867 Brussels

SCHUMANN U F ARNOLD R BUSEN J CURTIUS BKAumlRCHER A KIENDLER A PETZOLD H SCHLAGERF SCHROumlDER K-H WOHLFROM 2002 Influence of fu-els sulphur on the composition of aircraft exhaust plumesThe experiments SULFUR 1-7 ndash J Geophys Res 107(1010292001JD000813) AAC 2-1 ndash AAC 2-27

SOROKIN A P MIRABEL F ARNOLD 2003 On the In-fluence of fuel fulfur induced stable negative ion formationon the total concentration of ions emitted by an aircraft gasturbine engine Comparison of model and experiment ndash At-mos Chem Phys 3 6001ndash6018

VANCASSEL X A SOROKIN P MIRABEL A PETZOLDCW WILSON 2004 Volatile particles formation duringPartEmis a modelling study ndash Atmos Chem Phys 4 439ndash447

WILHELM S H HAVERKAMP A SOROKIN F ARNOLD2004 Detection of very large ions in aircraft gas turbineengine combustor exhaustcharged small soot particles ndashAtmos Environ 38 4561ndash4569

WILSON CW A PETZOLD S NYEKI U SCHUMANNR ZELLNER 2004 Measurement and Prediction of Emis-sions of Aerosols and Gaseous Precursors from Gas Tur-bine Engines (PartEmis) An Overview ndash Aerosp SciTechnol 8 131ndash143

Page 6: Particle emissions from aircraft engines a survey of the ... · 466 A. Petzold et al.: Particle emissions from aircraft engines Meteorol. Z., 14 , 2005 1 Introduction The role of

470 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

impact of operation conditions and fuels sulphur con-tent were not corrected for line losses because correc-tions are similar for all test points Thus relative trendsconcerning the influence of fuel sulphur and operationconditions on the microphysical properties of the emit-ted particles are not affected by line losses Becauseof the observed size range of exhaust particles countmedian diameters and geometric standard deviations ofmeasured size distributions are also expected to be onlymarginally influenced by line losses

3 Results

31 Microphysical properties

The aerosol emitted from the Combustor is composedof primary particles forming inside the combustor andof volatile condensation particles nucleating in the cool-ing exhaust gas from gaseous precursors The aerosolsize distribution in the diluted exhaust gas of the com-bustor is shown in Fig 2 for conditions with and with-out condensation particle formation While the mode ofcombustion aerosol particles dominates the size rangeD gt 15 nm with a maximum at 40 nm another modeof nanoparticles is present in all cases in the size rangeD= 7minus20 nm Condensation particle formation is oc-curring only at high FSC conditions and contributes tothe size range D= 4minus7 nm

The emission properties with respect to the primaryaerosol are shown in terms of BC mass (EIBC ) TCmass (EITC) and number (EIN10) emissions per kg ofconsumed fuel in Fig 3 The number emission indexEIN10 refers to particles with diameters D ge 10 nm InFig 3 also the count median diameter (CMD) of theexhaust aerosol size distribution is shown Plotted arecorresponding values for the combustor emissions andemissions from the HES low pressure stage In the firstplace it has to be mentioned that all the results are con-sistent At old cruise conditions an increasing numberemission at constant black carbon mass emission is bal-anced by a decreasing particle size with respect to theparticle electrical mobility diameter It should be notedthat these results originate from different measurementtechniques

When operated at old engine conditions with lowcombustor inlet temperature TC EIBC varies from 0051to 0081 g kgminus1 Operated at modern engine condi-tions with the higher inlet temperature EIBC ranges from0028 to 0044 g kgminus1 which is a factor of two lowerSimilar relations hold for the other properties How-ever the amount of emitted particulate matter is withinthe range permitted by the ICAO rules in all cases Allparticle emission indices referring to number and massare in close agreement with values obtained from in-flight measurements which span a range from 05ndash075

Figure 3 Emission indices in terms of number (N10 ) total car-

bon mass (TC) black carbon mass (BC) and count median diameter

(CMD) of the combustion aerosol for the Combustor exit (triangles)

and the Hot End Simulator low pressure stage (diamonds) values are

given for old and modern operation conditions as a function of the

fuel sulphur content error bars indicate signal variability in terms of

plusmn1 stdev

g kgminus1 for a Boeing B707 to 001 g kgminus1 for AirbusA 340 and Boeing B737 equipped with CFM 56 se-ries engines (SCHUMANN et al 2002 PETZOLD et al2003a 2003b) The fleet-averaged mass emission indexof 0038 g kgminus1 (PETZOLD et al 1999) lies within therange of values determined for modern operation condi-tions

Mean CMD values are 37ndash41 nm for old engine and40ndash44 nm for modern engine conditions The geomet-ric standard deviation of the size distribution is 167(16ndash18) for both operation conditions (PETZOLD etal 2003a 2003b) The combustor operating conditionshave only a weak influence on the size number and massof emitted particles Size distribution peaking at similardiameters were observed during several in-flight stud-ies HAGEN et al (1996 1998) reported peak diame-ters of 30ndash60 nm and size distribution widths of approx16 During the SULFUR studies (PETZOLD et al 1999SCHUMANN et al 2002) peak diameters of 35 nm wereobserved The PartEmis size distribution parameters arein close agreement with the reported in-flight observa-tions

The influence of the different HES pressure stages isdemonstrated in Fig 3 and with more detail in Fig 4As is shown in Fig 3 the trends observed in particleemission properties in the combustor experiments withvarying fuel sulphur content and operation conditionsagain observed during the HES experiments Except fora step-like increase from the combustor experiments tothe HES experiments which is most likely due to slightlydifferent operating conditions the HES has no influenceon the emission properties of the Combustor-HES com-bination with respect to combustion aerosol particles

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 471

0

5

10

15

20

25

30

000

005

010

015

020

high intermed low pressure

Combustor Hot End Simulator section

EIBC

EIN

EI

BC

g

kg

-1

EI

N

10

14kg

-1

Figure 4 Effect of the Hot End Simulator on the emission propertiesof the Combustor-HES combination operated at modern conditions

error bars indicate signal variability in terms of plusmn1 stdev

see Fig 4 Increasing the FSC slightly increases the to-tal number of emitted particles

Particles smaller than 10 nm contribute less than10 to the total exhaust aerosol in the low and mediumFSC cases At high FSC levels the formation of volatilecondensation particles with diameters Dlt 10 nm is ob-served (PETZOLD et al 2003c) A kinetic model wasused to investigate the nucleation efficiency of the H2O- H2SO4 binary mixture in the exhaust gas and the par-ticle coagulation processes during the transport throughthe sampling system to the measurement instrumenta-tion (VANCASSEL et al 2004) The model confirmedthat the condensation particles form at the dilutor of thesampling system During combustion the sulphur con-tained in the fuel is oxidised to SO2 which is then par-tially converted to S(VI) in the hot exhaust gas The ef-ficiency of conversion from S(IV) to S(VI) is describedby the conversion efficiency calculated as ε =S(VI)STwhere ST is the total fuel sulphur In the combustor ex-haust the S(VI) increases with increasing FSC In thekinetic model study a value for the efficiency was indi-rectly deduced from comparisons between model resultsand measurements In the PartEmis cases ε ranges be-tween roughly 25 and 6 depending on the com-bustor settings and on the microphysical approach usedFor these values which are close to results from di-rect measurements of gaseous sulphuric acid the kineticmodel reproduces the observation of volatile particles inthe size range 4ndash7 nm for the high FSC case while novolatile particles are found by the model for the low andmedium FSC cases

Particles of size D ge 30 nm are almost entirely in-ternally mixed Smaller particles (D lt 20 nm) exhibitvolume fractions of 10ndash15 for volatile and 4ndash10 for semi-volatile material while the volatile fraction de-creases to lt 5 for D ge 50 nm particles (PETZOLD

et al 2003a NYEKI et al 2003) The decrease of thevolatile particle fraction with increasing particle size isalso reproduced by the kinetic model for the H2O ndashH2SO4 binary mixture (VANCASSEL et al 2004) Thesevalues are valid for the conditions met during PartEmisIf the dilution air is more humid and a higher dilutionforces stronger formation of condensation particles thecondensation particles may grow to larger sizes How-ever as was observed in a young aircraft plume at cruiseconditions condensation particles were also observedmainly in the size range Dlt 10 nm (SCHROumlDER et al2000) similar to the PartEmis results

32 Chemical properties

Primary particles emitted from the combustor arepredominantly composed of carbonaceous material(gt95 ) and of water-soluble inorganic componentsThe carbonaceous material is composed of organic com-pounds and of elemental or black carbon The organiccompounds evaporate and oxidise at different tempera-tures and become classified according to this criterionCompounds which evaporate at temperatures as low as400 K are classified as semi-volatie OC compoundsevaporating compounds evaporating up to 800 K aretermed nonvolatile OC and the remainder is elementalcarbon Fig 5 shows the thermal stability of an exhaustaerosol sampled during operation with low FSC fuel atmodern conditions In the applied version of evolved gasanalysis (SCHMID et al 2001) the blackness of the filtersample is recorded parallel to the heating of the sampleAs is demonstrated in Fig 5 the filter starts to becomeldquowhiterdquo again as soon as the elemental or black carbonstarts to oxidise The evaporation and subsequent oxi-dation of OC compounds leaves the filter black Thusthe onset of increasing filter transmittance is an addi-tional measure for separating organic and elemental car-bon The application of this methods yields an averagecontribution of black carbon to total carbon of 40ndash80 depending on the combustion conditions (PUXBAUM etal 2003) From an aerosol loaded glass wool samplemore than 180 different NMHCs were desorbed underan Helium atmosphere Only 04 wt of the NMHCswere desorbed at T le 400 K whereas 996 wt weredesorbed at T = 400ndash620 K This is in good agreementwith the result of the thermogram in Fig 5 Elemental(EC) or black carbon (BC) respectively is chemicallyalmost inert at low and moderate temperatures and ther-mally stable up to temperatures of about 700 K

As determined from direct measurements of gaseousH2SO4 the efficiency of conversion in the combus-tor exhaust ranges between 09 plusmn 05 and 13 plusmn07 (KATRAGKOU et al 2004) In the HES ex-haust S(VI) increases with increasing FSC fuel flowand from the high-pressure to the low-pressure HESstage For FSC=1270 g kgminus1 the conversion efficiency

472 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

is 23plusmn 12 at the low pressure stage for the mod-ern cruise condition and 14plusmn 07 for the old cruiseThe higher ε observed for the modern cruise conditionssuggests that modern engines have a larger ε comparedto old engines Conversion efficiencies determined fromin-flight data at cruise altitude gave values 04ndash45 (SCHUMANN et al 2002) which is in good agreementwith the PartEmis ground test values

Chemical composition information was obtainedfrom Berner low pressure impactor samples which weretaken at a sample line temperature of 150C (PUXBAUMet al 2003) At these conditions most of the sulphuricacid remains in the gas phase while only a small fractionis more strongly bound to or chemi-sorbed at the par-ticle surface and will be observed in particulate mattersamples The chemical analysis results indicate that themass distribution of the carbonaceous matter matchesthe volume distribution of the aerosol The mass dis-tribution of chemisorbed sulphate however follows thesurface distribution of the aerosol which was calculatedfrom aerosol size spectra assuming a spherical particleshape (PUXBAUM et al 2003) This observation indi-cates that the sulphate is chemisorbed at the particle sur-face The chemisorbed fraction of sulphuric acid corre-sponds to a coverage of max 01 monolayer in the highFSC case This fraction of converted S(VI) is less than01 of total S(IV) present in the fuel The overall par-titioning of sulphur-containing species at 150C is 97 SO2le 27 gaseous H2SO4lt 03 chemisorbedAfter cooling to environmental temperatures the gasphase sulphuric acid can condense and cover the com-bustion particles with about one monolayer of sulphuricacid on average (HITZENBRGER et al 2003a)

33 Particle hygroscopic properties

Particle hygroscopic properties were measured at ambi-ent conditions using a Hygroscopicity Tandem Differ-ential Mobility Analyser (H-TDMA) system (GYSEL etal 2003a 2003b) The instrument measures the hygro-scopic growth factor (g) of particles with initial dry sizesDo = 30 50 and 100 nm over the range RHsim 70minus95 where g(RH) = D(RH)Do Growth factors were found toincrease with increasing FSC at fixed particle size (egDo = 50 nm g (95 ) = 101 and 116 for FSC = lowand high respectively) and to decrease with increasingparticle size at fixed FSC (eg mid FSC g (95 ) =114 110 and 103 for Do = 30 50 and 100 nm respec-tively) The results suggest an increasing amount of sul-phuric acid adsorbed on combustion particles from thegas phase with increasing FSC

Cloud condensation nuclei (CCN) were measured inthe diluted line using a static thermal diffusion typeCCN counter (GIEBL et al 2002) Particles are exposedto controlled supersaturations of water vapour whichenables some of the particles to become activated and

Figure 5 Evolving CO2 as a function of the sample heating temper-ature for a Combustor aerosol sample semi-volatile OC oxidises at

450 K non-volatile OC at 600 K and elemental carbon above 700 K

grow to large (gt 10 microm) droplets During PartEmis thesupersaturation was set tosim 07 The fraction of CCNin the aerosol (relative to the concentration of particlesof Dgt 10 nm) increases with increasing FSC but evenfor low FSC the fraction of CCN (at 0755 supersat-uration) is higher than the fraction of particles with di-ameters larger than 280 nm ie the size at which wet-table insoluble particles are activated according to theKelvin effect At high FSC the fraction of CCN in-creases by about a factor 55 for old and 12 for mod-ern conditions However pure sulphuric acid particlesof the same size which become activated according toKOumlHLER theory would give an activation ratio which isstill larger by a factorsim 15 compared to the observationseven at high FSC Thus the KOumlHLER model cannot re-produce measured activation properties This is in agree-ment with the fact that large combustion particles are notat all completely water soluble The Kelvin model pro-vides good predictions for the low FSC level but acti-vation diameters (ratios) are increasingly overestimated(underestimated) at medium and high FSC This indi-cates that the soluble fraction of the combustion parti-cles increases from low values at low FSC level to rel-atively large values at high FSC An empirical coatedsphere model which equates the measured volatile andsemivolatile fraction with soluble sulphuric acid under-estimates the activation diameter by 45 to 55 underall conditions A small amount of semivolatile organ-ics with low or absent water-solubility adsorbed on thecombustion particles might be one reason for this dis-crepancy Further details are given by HITZENBERGER

et al (2003a 2003b)

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 473

34 Gaseous hydrocarbon emission

More than 100 different non-methane volatile organiccompounds (NMVOCs) such as non-methane hydrocar-bons (NMHCs) and selected partially oxidised hydro-carbons (OHCs) emitted from the aircraft engine com-bustor and the hot end simulator (HES) were identifiedand quantified These species accounted for up to 91wt of the total hydrocarbons (THCs) The NMHCemission indices of 5ndash15 microg kgminus1 in total are compa-rable with the AEROTRACE (Measurement of TraceSpecies in the Exhaust from Aero Engines EC projectAERA-CT94-0003 1994ndash1997 Co-ordinated by Qine-tiQ) study Both studies show up to 1000 times loweremission indices in comparison with emissions from ve-hicular traffic The influence of the operation conditionson NMHC emissions of the combustor and HES was in-vestigated for the identified NMHCs From the six con-ditions which were investigated (2 operation conditions3 HES pressure stages) most compounds show a de-crease of NMHC emissions with increasing combustiontemperature at the same pressure stage This is in agree-ment with results from the combustor exit measurementsand with the results of the AEROTRACE study Emis-sion indices at HES pressure stages high (HP)- inter-mediate (IP)- and low pressure (LP) for fixed FSC andcombustor operation conditions were also investigatedMost compounds show a decrease of NMHC emissionswith decreasing pressure in the HES at similar combus-tor operation conditions and fixed FSC The decrease ofthe pressure in the HES corresponds to an increase ofthe residence time in a turbine and indicates a decreaseof NMHC emissions along the different turbine sections

35 Emission of ions and charged sootparticles

The ions measured during the HES campaign(HAVERKAMP et al 2004 WILHELM et al 2004) areproduct ions formed from primary ions via ion moleculereactions involving either few reaction-steps withmassive neutral molecules or numerous reaction-stepswith relatively small neutral molecules Their lifetimeswith respect to ion-ion recombination are around 9 msafter the high pressure turbine section The ion emissionindex Ei is up to 45 times 1016 ions per kg fuel burntAs is discussed by HAVERKAMP et al (2004) neitherion-ion recombination nor ion removal by attachmentto combustion particles give sufficient explanations forthe observed ion emission index Therefore detailedmodel calculations of the ion chemistry taking place inthe combustor with particular emphasis on the sulphurspecies effect is required (HAVERKAMP et al 2004)It seems that ion-induced nucleation can explain atleast partly the observed volatile aerosols Electricallycharged small positive and negative soot particles

(CSP) have also been detected during both PartEmiscampaigns Charged soot particles are found to havediameters around 6 nm and a total concentration of upto 45times107 cmminus3 (positive and negative) correspondingto a total emission index of ECSP = 23times 1015 CSP perkg fuel burnt (WILHELM et al 2004)

A model which considers the formation and evolu-tion of combustion ions in a combustor of an aircraftengine in dependence on the electron detachment ef-ficiency from negative ions was used to consider theeffect of the transformation of primary negative ionsto more stable secondary negative ions (SOROKIN etal 2003) The formed stable negative ions most prob-ably are sulphur-bearing ions This effect slows downthe charged particle neutralisation rate leading to an in-crease of the concentration of positive and negative ionsat the combustor exit The results of the simulation andtheir comparison with the ground-based experimentaldata obtained within PartEmis support the hypothesisthat the increase of the fuel sulphur content leads to anincrease of the ion concentration at the combustor exit

36 OH measurements

The concentration of OH which is the dominant oxidantin the exhaust gas was detected by laser induced fluo-rescence (LIF) to obtain an absolute OH concentrationmeasurement in order to aid in the modelling of SO3production from the combustor to the engine exit Whenrunning the laser at low laser energies OH in the exhaustduct was successfully detected Calibration experimentsprovide a method to assign the observed OH LIF sig-nals an absolute concentration and from an estimate ormeasurement of the errors for each parameter an overallestimate of the error in the assigned OH concentrationwas made From these calibration experiments very lowvalues of less than 1 ppbv were calculated for the OHconcentration in the exhaust gas after exiting the HESPrevious experiments using a combined Raman scatter-ingLIF technique to look in the exhaust gases 50 cmbehind an aircraft engine exit failed to detect any OHLIF signal and therefore could only infer an upper limitfor OH of 80 ppbv based on the measured detection limitof their apparatus

4 Summary and conclusions

The progress made during PartEmis is manifold Themain conclusions are therefore grouped into differentsubject areas and presented as brief statements For eachsubject area key references are identified if available

41 Combustion aerosol

Refer to WILSON et al (2004) and PETZOLD et al(2003a)

474 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

(i) The combustor behaves like a typical aircraft en-gine combustor with respect to thermodynamic dataand main emissions It may thus be expected that thePartEmis database can be applied as a first order approx-imation to contemporary aircraft engines

(ii) The BC concentration is nearly independent ofFSC level

(iii) The number size distribution of primary particlesin the diameter range 10 lt Dlt 550 nm is almost unin-fluenced by the FSC level or combustor conditions Athigh FSC levels the formation of volatile condensationparticles with diameters Dlt 10 nm occurs

(iv) The volatile volume fraction of particles gt 50 nmincreases from 2 for low FSC to approx 5 for highFSC The volatile volume fraction of particles lt 30 nmis le 15

(v) Volatility measurements indicate an internallymixed aerosol for particles with D gt 15 nm conden-sation particles are below 15 nm in diameter

42 Ultrafine particles

Refer to PETZOLD et al (2003a) and VANCASSEL et al(2004)

(vi) Formation of condensation particles of size D=4minus 7 nm is observed in the Combustor tests for highFSC at old and modern conditions formation occurs ifthe surface area of the combustion aerosol falls below acertain threshold limit the observed number of formedparticles is confirmed by kinetic model studies

(vii) Formation of condensation particles is observedin the HES tests for high FSCmodern conditionslowpressure section only the ldquoformation strengthrdquo in termsof number of condensation particles per number of com-bustion particle is twice as high as for the Combustorcase

(viii) The model studies indicate that condensationparticles are not emitted from the combustor but form inthe sampling line at the sample dilution point

(ix) Ultrafine carbonaceous particles of size 7 nmlt D lt 20 nm are observed independently of the fuelsulphur content a volatility analysis indicates that theseparticles are nonvolatile and may be composed of car-bonaceous compounds

43 Sulphate-containing species chemi-ionsand OH

Refer to KATRAGKOU et al (2004) and HAVERKAMP etal (2004)

(x) S(VI) was measured in the form of sulphuric acidIncreased FSC results in higher S(VI) concentrationsThe measured conversion efficiency for fuel sulphur toS(IV) at the combustor exit varies from 035ndash14 andis independent of the investigated combustor operationconditions and FSC levels

(xi) 10ndash30 of sulphate is chemisorbed on primaryparticles the coverage of primary particles is below 01monolayers of sulphuric acid at 150C and approxi-mately one monolayer at 25C

(xii) Sulphur partitioning at 150C 97 SO2le 27 gaseous H2SO4 lt 03 chemisorbed

(xiii) Positive and negative gaseous ions were de-tected in the exhaust of a combustor with mass numbersup to 1500 The total ion emission index at the HES hpstage is as high as 54times 1017 positive and negative ionsper kg fuel burnt Large ions with mass numbers exceed-ing 4500 contribute about 10 to the total ion numberconcentration

(xiv) OH concentration in the exhaust gas after exit-ing the HES is lt 1 ppbv

44 Hygroscopic growth and CCN activation

Refer to GYSEL et al (2003a) and HITZENBERGER etal (2003a)

(xv) Small differences in particle hygroscopicity areobserved between old and modern cruise conditions

(xvi) Strong effect of the fuel sulphur content is ob-served particle hygroscopicity increases distinctly withincreasing FSC

(xvii) No significant effect of the HES at low andmid FSC level is observed compared to the combustorexit an increased particle hygroscopicity downstreamthe HES compared to the combustor is observed at highFSC indicating continuing transformation of S(IV) toS(VI) during the transfer through the HES

(xviii) CCN activation of combustion particles in-creases with increasing FSC as a result of larger par-ticle hygroscopicity additional effects of adsorbed or-ganic matter cannot be excluded

45 Gaseous organic fraction

(xix) More than 100 different NMVOCs (aliphatic andaromatic hydrocarbons carbonyls and acids) were iden-tified and quantified they account for 91 wt of thetotal detected compounds

(xx) A decrease in NMHC emissions from old tomodern conditions at fixed HES pressure stage and fuelsulphur content (FSC) is found

(xxi) A decrease of NMHC emissions with decreas-ing pressure in the HES at fixed combustor operationconditions and FSC is observed

(xxii) No clear influence on the NMHC emissionswith changing FSC at the same combustor operationcondition is observed

(xxiii) High EIrsquos for organic acids are found Pos-sible explanations are that organic compounds fixed atthe sampling line walls are oxidised by the hot exhaustgases or that the acids are fixed at the surface of thesoot particles which are also trapped in the samplingcartridges

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 475

46 Emission properties

Refer to WILSON et al (2004) and PETZOLD et al(2003a)

(xxiv) Emission properties regarding combustionaerosol mass number and size are determined by thecombustor the influence of the Hot End Simulator isweak

(xxv) The processing of exhaust gas and parti-cles during the transfer through the Hot End Simula-tor increases the particle hygroscopicity and formationstrength of condensation particles both effects are re-lated to gaseous sulphuric acid

The experimental overview of the EU PartEmisproject presents the first comprehensive aerosol and gasmeasurements to be conducted on an aircraft enginecombustor test-rig The database will be used for fur-ther analyses as well as for plume modelling studies ofthe effect of aviation-related particle emissions on theglobal atmosphere and climate

Acknowledgements

The PartEmis project was funded by the European Com-mission under contract no G4RD-CT-2000-00207 Theauthors are very grateful to the test rig operation crew atQinetiQ for their very strong support during the experi-ments

References

ANDERSON BE WR COFER DR BAGWELL JWBARRICK CH HUDGINS KE BRUNKE 1998 Air-borne observations of aircraft aerosol emissions I Totalnonvolatile particle emission indices ndash Geophys Res Lett25 1689ndash1692

BOUCHER O 1999 Influence of air traffic on cirrus occur-rence Nature ndash 397 30ndash31

BROCK CA F SCHROumlDER B KAumlRCHER A PETZOLDR BUSEN M FIEBIG 2000 Ultrafine particle size distri-butions measured in aircraft exhaust plumes ndash J GeophysRes 105 26555ndash26567

CACHIER H MP BREMOND P BIAT-MEacuteNARD 1989Determination of atmospheric soot carbon with a simplethermal method ndash Tellus 41B 379ndash390

DEMOTT PJ Y CHEN SM KREIDENWEIS DCROGERS DE SHERMAN 1999 Ice formation by blackcarbon particles ndash Geophys Res Lett 26 2429ndash2432

GIEBL H A BERNER G REISCHL H PUXBAUM AKASPER-GIEBL R HITZENBERGER 2002 CCN activa-tion of oxalic and malonic acid test aerosols with the Uni-versity of Vienna cloud condensation nuclei counter ndash JAerosol Sci 33 1623ndash1634

GYSEL M E WEINGARTNER U BALTENSPERGER2002 Hygroscopicity of aerosol particles at low tempera-tures 2 Theoretical and experimental hygroscopic proper-ties of laboratory generated aerosols ndash Environ Sci Tech-nol 36 63ndash68

GYSEL M S NYEKI E WEINGARTNER U BAL-TENSPERGER H GIEBL R HITZENBERGER A PET-ZOLD CW WILSON 2003a Properties of jet engine com-bustor particles during the PartEmis experiment Hygro-scopicity at subsaturated conditions ndash Geophys Res Lett30 1566 DOI1010292003GL016896

GYSEL M S NYEKI E WEINGARTNER U BAL-TENSPERGER H GIEBL R HITZENBERGER A PET-ZOLD CW WILSON 2003b Jet engine combustion par-ticle hygroscopicity under subsaturated conditions duringPartEmis ndash J Aerosol Sci 34 (Suppl 2) Abstracts of theEuropean Aerosol Conference S1415ndashS1416

HAGEN DE PD WHITEFIELD H SCHLAGER 1996Particle emissions in the exhaust plume from commercialjet aircraft under cruise conditions ndash J Geophys Res 10119551ndash19557

HAGEN DE PD WHITEFIELD J PALADINO M TRUE-BLOOD H LILENFELD 1998 Particle sizing and emissionindices for a jet engine exhaust sampled at cruise ndash Geo-phys Res Lett 25 1681ndash1684

HAGEN DE PD WHITEFIELD JD PALADINO OSCHMID 2001 Comparative study of jet engine com-bustion emissions and engine operating conditions ndashIn SCHUMANN U G AMANATIDIS (Eds) Aviationaerosols contrails and cirrus clouds Air pollution researchreport 74 106ndash110 European Commission Brussels

HAVERKAMP H S WILHELM A SOROKIN F ARNOLD2004 Positive and negative ion measurements in jet aircraftengine exhaust concentrations sizes and implications foraerosol formation ndash Atmos Environ 38 2879ndash2884

HITZENBERGER R H GIEBL A PETZOLD M GYSELS NYEKI E WEINGARTNER U BALTENSPERGERCW WILSON 2003a Properties of jet engine combus-tor particles during the PartEmis experiment Hygroscopicproperties at supersaturated conditions ndash Geophys ResLett 30 1779 DOI1010292003GL017294

HITZENBERGER R H GIEBL A PETZOILD M GYSELS NYEKI E WEINGARTNER U BALTENSPERGERCW WILSON 2003b CCN activation of jet engine com-bustor particles during PartEmis ndash J Aerosol Sci 34(Suppl 2) Abstracts of the European Aerosol ConferenceS1409ndashS1410

HURLEY CD 1996 AEROTRACE Measurements of Par-ticulates from an Engine Combustor ndash Proceedings Impactof Aircraft Emissions upon the Atmosphere Vol 1 113 In-ternational Colloquium Paris October 1996

KATRAGKOU E S WILHELM F ARNOLD CW WIL-SON 2004 First gaseous Sulfur (VI) measurements in thesimulated internal flow of an aircraft gas turbine engineduring project PartEmis ndash Geophys Res Lett 31 2117DOI1010292003GL018231

NYEKI S M GYSEL E WEINGARTNER U BAL-TENSPERGER R HITZENBERGER A PETZOLD CWWILSON 2003 Volatile properties of jet engine combus-tor particles during PartEmis ndash J Aerosol Sci 34 (Suppl2) Abstracts of the European Aerosol Conference S1411ndashS1412

PETZOLD A FP SCHROumlDER 1998 Jet engine exhaustaerosol characterization ndash Aerosol Sci Technol 28 62ndash76

PETZOLD A A DOumlPELHEUER CA BROCK FPSCHROumlDER 1999 In situ observations and model calcula-

476 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

tions of black carbon emission by aircraft at cruise altitudendash J Geophys Res 104 22171ndash22181

PETZOLD A C STEIN S NYEKI M GYSEL EWEINGARTNER U BALTENSPERGER H GIEBL RHITZENBERGER A DOumlPELHEUER S VRCHOTICKY HPUXBAUM M JOHNSON CD HURLEY R MARSHCW WILSON 2003a Properties of jet engine combus-tor particles during the PartEmis experiment Microphysi-cal and chemical properties ndash Geophys Res Lett 30 1719DOI1010292003GL017283

PETZOLD A CW WILSON U BALTENSPERGERM FIEBIG L FRITZSCHE H GIEBL M GYSELR HITZENBERGER CD HURLEY S NYEKI HPUXBAUM U SCHUMANN C STEIN S VROCHTICKY2003b Particle emissions from aircraft engines ndash anoverview of the European project PartEmis ndash J AerosolSci 34 (Suppl 2) Abstracts of the European Aerosol Con-ference S1405ndashS1406

PETZOLD A AE VRTALA M FIEBIG F FRITZSCHEPH MIRABEL S NYEKI X VANCASSEL CW WIL-SON 2003c Emission of volatile and non-volatile ultra-fine particles from a combustion source during PartEmisndash J Aerosol Sci 34 (Suppl 2) Abstracts of the EuropeanAerosol Conference S1407ndashS1408

PUXBAUM H S VRCHOTICKY A PETZOLD R HITZEN-BERGER S NYEKI CW WILSON 2003 Chemical com-position of jet engine combustor particles during PartEmisndash J Aerosol Sci 34 (Suppl 2) Abstracts of the EuropeanAerosol Conference S1413ndashS1414

SCHMID H L LASKUS H-J ABRAHAM U BAL-TENSPERGER V LAVANCHY M BIZJAK P BURBA HCACHIER D CROW J CHOW T GNAUK A EVENHM TEN BRINK K-P GIESEN R HITZENBERGER CHUEGLIN W MAENHAUT C PIOA CARVALHO J-PPUTAUD D TOOM-SAUNTRY H PUXBAUM 2001 Re-sults of the ldquocarbon conferencerdquo international aerosol car-bon round robin test stage I ndash Atmos Environ 35 2111ndash2121

SCHROumlDER F C BROCK R BAUMANN A PETZOLD RBUSEN P SCHULTE M FIEBIG 2000 In-situ studies onvolatile jet exhaust particle emissions Impact of fuel sul-fur content and thermodynamic conditions on nuclei-modeaerosols ndash J Geophys Res 105 19941ndash19954

SCHUMANN U 1996 Contrail formation from aircraft ex-hausts ndash Meteorol Z NF 5 4ndash23

SCHUMANN U J STROumlM 2001 Aviation impact on atmo-spheric composition and climate ndash In European research inthe stratosphere 1996ndash2000 European Commission EUR19867 Brussels

SCHUMANN U F ARNOLD R BUSEN J CURTIUS BKAumlRCHER A KIENDLER A PETZOLD H SCHLAGERF SCHROumlDER K-H WOHLFROM 2002 Influence of fu-els sulphur on the composition of aircraft exhaust plumesThe experiments SULFUR 1-7 ndash J Geophys Res 107(1010292001JD000813) AAC 2-1 ndash AAC 2-27

SOROKIN A P MIRABEL F ARNOLD 2003 On the In-fluence of fuel fulfur induced stable negative ion formationon the total concentration of ions emitted by an aircraft gasturbine engine Comparison of model and experiment ndash At-mos Chem Phys 3 6001ndash6018

VANCASSEL X A SOROKIN P MIRABEL A PETZOLDCW WILSON 2004 Volatile particles formation duringPartEmis a modelling study ndash Atmos Chem Phys 4 439ndash447

WILHELM S H HAVERKAMP A SOROKIN F ARNOLD2004 Detection of very large ions in aircraft gas turbineengine combustor exhaustcharged small soot particles ndashAtmos Environ 38 4561ndash4569

WILSON CW A PETZOLD S NYEKI U SCHUMANNR ZELLNER 2004 Measurement and Prediction of Emis-sions of Aerosols and Gaseous Precursors from Gas Tur-bine Engines (PartEmis) An Overview ndash Aerosp SciTechnol 8 131ndash143

Page 7: Particle emissions from aircraft engines a survey of the ... · 466 A. Petzold et al.: Particle emissions from aircraft engines Meteorol. Z., 14 , 2005 1 Introduction The role of

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 471

0

5

10

15

20

25

30

000

005

010

015

020

high intermed low pressure

Combustor Hot End Simulator section

EIBC

EIN

EI

BC

g

kg

-1

EI

N

10

14kg

-1

Figure 4 Effect of the Hot End Simulator on the emission propertiesof the Combustor-HES combination operated at modern conditions

error bars indicate signal variability in terms of plusmn1 stdev

see Fig 4 Increasing the FSC slightly increases the to-tal number of emitted particles

Particles smaller than 10 nm contribute less than10 to the total exhaust aerosol in the low and mediumFSC cases At high FSC levels the formation of volatilecondensation particles with diameters Dlt 10 nm is ob-served (PETZOLD et al 2003c) A kinetic model wasused to investigate the nucleation efficiency of the H2O- H2SO4 binary mixture in the exhaust gas and the par-ticle coagulation processes during the transport throughthe sampling system to the measurement instrumenta-tion (VANCASSEL et al 2004) The model confirmedthat the condensation particles form at the dilutor of thesampling system During combustion the sulphur con-tained in the fuel is oxidised to SO2 which is then par-tially converted to S(VI) in the hot exhaust gas The ef-ficiency of conversion from S(IV) to S(VI) is describedby the conversion efficiency calculated as ε =S(VI)STwhere ST is the total fuel sulphur In the combustor ex-haust the S(VI) increases with increasing FSC In thekinetic model study a value for the efficiency was indi-rectly deduced from comparisons between model resultsand measurements In the PartEmis cases ε ranges be-tween roughly 25 and 6 depending on the com-bustor settings and on the microphysical approach usedFor these values which are close to results from di-rect measurements of gaseous sulphuric acid the kineticmodel reproduces the observation of volatile particles inthe size range 4ndash7 nm for the high FSC case while novolatile particles are found by the model for the low andmedium FSC cases

Particles of size D ge 30 nm are almost entirely in-ternally mixed Smaller particles (D lt 20 nm) exhibitvolume fractions of 10ndash15 for volatile and 4ndash10 for semi-volatile material while the volatile fraction de-creases to lt 5 for D ge 50 nm particles (PETZOLD

et al 2003a NYEKI et al 2003) The decrease of thevolatile particle fraction with increasing particle size isalso reproduced by the kinetic model for the H2O ndashH2SO4 binary mixture (VANCASSEL et al 2004) Thesevalues are valid for the conditions met during PartEmisIf the dilution air is more humid and a higher dilutionforces stronger formation of condensation particles thecondensation particles may grow to larger sizes How-ever as was observed in a young aircraft plume at cruiseconditions condensation particles were also observedmainly in the size range Dlt 10 nm (SCHROumlDER et al2000) similar to the PartEmis results

32 Chemical properties

Primary particles emitted from the combustor arepredominantly composed of carbonaceous material(gt95 ) and of water-soluble inorganic componentsThe carbonaceous material is composed of organic com-pounds and of elemental or black carbon The organiccompounds evaporate and oxidise at different tempera-tures and become classified according to this criterionCompounds which evaporate at temperatures as low as400 K are classified as semi-volatie OC compoundsevaporating compounds evaporating up to 800 K aretermed nonvolatile OC and the remainder is elementalcarbon Fig 5 shows the thermal stability of an exhaustaerosol sampled during operation with low FSC fuel atmodern conditions In the applied version of evolved gasanalysis (SCHMID et al 2001) the blackness of the filtersample is recorded parallel to the heating of the sampleAs is demonstrated in Fig 5 the filter starts to becomeldquowhiterdquo again as soon as the elemental or black carbonstarts to oxidise The evaporation and subsequent oxi-dation of OC compounds leaves the filter black Thusthe onset of increasing filter transmittance is an addi-tional measure for separating organic and elemental car-bon The application of this methods yields an averagecontribution of black carbon to total carbon of 40ndash80 depending on the combustion conditions (PUXBAUM etal 2003) From an aerosol loaded glass wool samplemore than 180 different NMHCs were desorbed underan Helium atmosphere Only 04 wt of the NMHCswere desorbed at T le 400 K whereas 996 wt weredesorbed at T = 400ndash620 K This is in good agreementwith the result of the thermogram in Fig 5 Elemental(EC) or black carbon (BC) respectively is chemicallyalmost inert at low and moderate temperatures and ther-mally stable up to temperatures of about 700 K

As determined from direct measurements of gaseousH2SO4 the efficiency of conversion in the combus-tor exhaust ranges between 09 plusmn 05 and 13 plusmn07 (KATRAGKOU et al 2004) In the HES ex-haust S(VI) increases with increasing FSC fuel flowand from the high-pressure to the low-pressure HESstage For FSC=1270 g kgminus1 the conversion efficiency

472 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

is 23plusmn 12 at the low pressure stage for the mod-ern cruise condition and 14plusmn 07 for the old cruiseThe higher ε observed for the modern cruise conditionssuggests that modern engines have a larger ε comparedto old engines Conversion efficiencies determined fromin-flight data at cruise altitude gave values 04ndash45 (SCHUMANN et al 2002) which is in good agreementwith the PartEmis ground test values

Chemical composition information was obtainedfrom Berner low pressure impactor samples which weretaken at a sample line temperature of 150C (PUXBAUMet al 2003) At these conditions most of the sulphuricacid remains in the gas phase while only a small fractionis more strongly bound to or chemi-sorbed at the par-ticle surface and will be observed in particulate mattersamples The chemical analysis results indicate that themass distribution of the carbonaceous matter matchesthe volume distribution of the aerosol The mass dis-tribution of chemisorbed sulphate however follows thesurface distribution of the aerosol which was calculatedfrom aerosol size spectra assuming a spherical particleshape (PUXBAUM et al 2003) This observation indi-cates that the sulphate is chemisorbed at the particle sur-face The chemisorbed fraction of sulphuric acid corre-sponds to a coverage of max 01 monolayer in the highFSC case This fraction of converted S(VI) is less than01 of total S(IV) present in the fuel The overall par-titioning of sulphur-containing species at 150C is 97 SO2le 27 gaseous H2SO4lt 03 chemisorbedAfter cooling to environmental temperatures the gasphase sulphuric acid can condense and cover the com-bustion particles with about one monolayer of sulphuricacid on average (HITZENBRGER et al 2003a)

33 Particle hygroscopic properties

Particle hygroscopic properties were measured at ambi-ent conditions using a Hygroscopicity Tandem Differ-ential Mobility Analyser (H-TDMA) system (GYSEL etal 2003a 2003b) The instrument measures the hygro-scopic growth factor (g) of particles with initial dry sizesDo = 30 50 and 100 nm over the range RHsim 70minus95 where g(RH) = D(RH)Do Growth factors were found toincrease with increasing FSC at fixed particle size (egDo = 50 nm g (95 ) = 101 and 116 for FSC = lowand high respectively) and to decrease with increasingparticle size at fixed FSC (eg mid FSC g (95 ) =114 110 and 103 for Do = 30 50 and 100 nm respec-tively) The results suggest an increasing amount of sul-phuric acid adsorbed on combustion particles from thegas phase with increasing FSC

Cloud condensation nuclei (CCN) were measured inthe diluted line using a static thermal diffusion typeCCN counter (GIEBL et al 2002) Particles are exposedto controlled supersaturations of water vapour whichenables some of the particles to become activated and

Figure 5 Evolving CO2 as a function of the sample heating temper-ature for a Combustor aerosol sample semi-volatile OC oxidises at

450 K non-volatile OC at 600 K and elemental carbon above 700 K

grow to large (gt 10 microm) droplets During PartEmis thesupersaturation was set tosim 07 The fraction of CCNin the aerosol (relative to the concentration of particlesof Dgt 10 nm) increases with increasing FSC but evenfor low FSC the fraction of CCN (at 0755 supersat-uration) is higher than the fraction of particles with di-ameters larger than 280 nm ie the size at which wet-table insoluble particles are activated according to theKelvin effect At high FSC the fraction of CCN in-creases by about a factor 55 for old and 12 for mod-ern conditions However pure sulphuric acid particlesof the same size which become activated according toKOumlHLER theory would give an activation ratio which isstill larger by a factorsim 15 compared to the observationseven at high FSC Thus the KOumlHLER model cannot re-produce measured activation properties This is in agree-ment with the fact that large combustion particles are notat all completely water soluble The Kelvin model pro-vides good predictions for the low FSC level but acti-vation diameters (ratios) are increasingly overestimated(underestimated) at medium and high FSC This indi-cates that the soluble fraction of the combustion parti-cles increases from low values at low FSC level to rel-atively large values at high FSC An empirical coatedsphere model which equates the measured volatile andsemivolatile fraction with soluble sulphuric acid under-estimates the activation diameter by 45 to 55 underall conditions A small amount of semivolatile organ-ics with low or absent water-solubility adsorbed on thecombustion particles might be one reason for this dis-crepancy Further details are given by HITZENBERGER

et al (2003a 2003b)

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 473

34 Gaseous hydrocarbon emission

More than 100 different non-methane volatile organiccompounds (NMVOCs) such as non-methane hydrocar-bons (NMHCs) and selected partially oxidised hydro-carbons (OHCs) emitted from the aircraft engine com-bustor and the hot end simulator (HES) were identifiedand quantified These species accounted for up to 91wt of the total hydrocarbons (THCs) The NMHCemission indices of 5ndash15 microg kgminus1 in total are compa-rable with the AEROTRACE (Measurement of TraceSpecies in the Exhaust from Aero Engines EC projectAERA-CT94-0003 1994ndash1997 Co-ordinated by Qine-tiQ) study Both studies show up to 1000 times loweremission indices in comparison with emissions from ve-hicular traffic The influence of the operation conditionson NMHC emissions of the combustor and HES was in-vestigated for the identified NMHCs From the six con-ditions which were investigated (2 operation conditions3 HES pressure stages) most compounds show a de-crease of NMHC emissions with increasing combustiontemperature at the same pressure stage This is in agree-ment with results from the combustor exit measurementsand with the results of the AEROTRACE study Emis-sion indices at HES pressure stages high (HP)- inter-mediate (IP)- and low pressure (LP) for fixed FSC andcombustor operation conditions were also investigatedMost compounds show a decrease of NMHC emissionswith decreasing pressure in the HES at similar combus-tor operation conditions and fixed FSC The decrease ofthe pressure in the HES corresponds to an increase ofthe residence time in a turbine and indicates a decreaseof NMHC emissions along the different turbine sections

35 Emission of ions and charged sootparticles

The ions measured during the HES campaign(HAVERKAMP et al 2004 WILHELM et al 2004) areproduct ions formed from primary ions via ion moleculereactions involving either few reaction-steps withmassive neutral molecules or numerous reaction-stepswith relatively small neutral molecules Their lifetimeswith respect to ion-ion recombination are around 9 msafter the high pressure turbine section The ion emissionindex Ei is up to 45 times 1016 ions per kg fuel burntAs is discussed by HAVERKAMP et al (2004) neitherion-ion recombination nor ion removal by attachmentto combustion particles give sufficient explanations forthe observed ion emission index Therefore detailedmodel calculations of the ion chemistry taking place inthe combustor with particular emphasis on the sulphurspecies effect is required (HAVERKAMP et al 2004)It seems that ion-induced nucleation can explain atleast partly the observed volatile aerosols Electricallycharged small positive and negative soot particles

(CSP) have also been detected during both PartEmiscampaigns Charged soot particles are found to havediameters around 6 nm and a total concentration of upto 45times107 cmminus3 (positive and negative) correspondingto a total emission index of ECSP = 23times 1015 CSP perkg fuel burnt (WILHELM et al 2004)

A model which considers the formation and evolu-tion of combustion ions in a combustor of an aircraftengine in dependence on the electron detachment ef-ficiency from negative ions was used to consider theeffect of the transformation of primary negative ionsto more stable secondary negative ions (SOROKIN etal 2003) The formed stable negative ions most prob-ably are sulphur-bearing ions This effect slows downthe charged particle neutralisation rate leading to an in-crease of the concentration of positive and negative ionsat the combustor exit The results of the simulation andtheir comparison with the ground-based experimentaldata obtained within PartEmis support the hypothesisthat the increase of the fuel sulphur content leads to anincrease of the ion concentration at the combustor exit

36 OH measurements

The concentration of OH which is the dominant oxidantin the exhaust gas was detected by laser induced fluo-rescence (LIF) to obtain an absolute OH concentrationmeasurement in order to aid in the modelling of SO3production from the combustor to the engine exit Whenrunning the laser at low laser energies OH in the exhaustduct was successfully detected Calibration experimentsprovide a method to assign the observed OH LIF sig-nals an absolute concentration and from an estimate ormeasurement of the errors for each parameter an overallestimate of the error in the assigned OH concentrationwas made From these calibration experiments very lowvalues of less than 1 ppbv were calculated for the OHconcentration in the exhaust gas after exiting the HESPrevious experiments using a combined Raman scatter-ingLIF technique to look in the exhaust gases 50 cmbehind an aircraft engine exit failed to detect any OHLIF signal and therefore could only infer an upper limitfor OH of 80 ppbv based on the measured detection limitof their apparatus

4 Summary and conclusions

The progress made during PartEmis is manifold Themain conclusions are therefore grouped into differentsubject areas and presented as brief statements For eachsubject area key references are identified if available

41 Combustion aerosol

Refer to WILSON et al (2004) and PETZOLD et al(2003a)

474 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

(i) The combustor behaves like a typical aircraft en-gine combustor with respect to thermodynamic dataand main emissions It may thus be expected that thePartEmis database can be applied as a first order approx-imation to contemporary aircraft engines

(ii) The BC concentration is nearly independent ofFSC level

(iii) The number size distribution of primary particlesin the diameter range 10 lt Dlt 550 nm is almost unin-fluenced by the FSC level or combustor conditions Athigh FSC levels the formation of volatile condensationparticles with diameters Dlt 10 nm occurs

(iv) The volatile volume fraction of particles gt 50 nmincreases from 2 for low FSC to approx 5 for highFSC The volatile volume fraction of particles lt 30 nmis le 15

(v) Volatility measurements indicate an internallymixed aerosol for particles with D gt 15 nm conden-sation particles are below 15 nm in diameter

42 Ultrafine particles

Refer to PETZOLD et al (2003a) and VANCASSEL et al(2004)

(vi) Formation of condensation particles of size D=4minus 7 nm is observed in the Combustor tests for highFSC at old and modern conditions formation occurs ifthe surface area of the combustion aerosol falls below acertain threshold limit the observed number of formedparticles is confirmed by kinetic model studies

(vii) Formation of condensation particles is observedin the HES tests for high FSCmodern conditionslowpressure section only the ldquoformation strengthrdquo in termsof number of condensation particles per number of com-bustion particle is twice as high as for the Combustorcase

(viii) The model studies indicate that condensationparticles are not emitted from the combustor but form inthe sampling line at the sample dilution point

(ix) Ultrafine carbonaceous particles of size 7 nmlt D lt 20 nm are observed independently of the fuelsulphur content a volatility analysis indicates that theseparticles are nonvolatile and may be composed of car-bonaceous compounds

43 Sulphate-containing species chemi-ionsand OH

Refer to KATRAGKOU et al (2004) and HAVERKAMP etal (2004)

(x) S(VI) was measured in the form of sulphuric acidIncreased FSC results in higher S(VI) concentrationsThe measured conversion efficiency for fuel sulphur toS(IV) at the combustor exit varies from 035ndash14 andis independent of the investigated combustor operationconditions and FSC levels

(xi) 10ndash30 of sulphate is chemisorbed on primaryparticles the coverage of primary particles is below 01monolayers of sulphuric acid at 150C and approxi-mately one monolayer at 25C

(xii) Sulphur partitioning at 150C 97 SO2le 27 gaseous H2SO4 lt 03 chemisorbed

(xiii) Positive and negative gaseous ions were de-tected in the exhaust of a combustor with mass numbersup to 1500 The total ion emission index at the HES hpstage is as high as 54times 1017 positive and negative ionsper kg fuel burnt Large ions with mass numbers exceed-ing 4500 contribute about 10 to the total ion numberconcentration

(xiv) OH concentration in the exhaust gas after exit-ing the HES is lt 1 ppbv

44 Hygroscopic growth and CCN activation

Refer to GYSEL et al (2003a) and HITZENBERGER etal (2003a)

(xv) Small differences in particle hygroscopicity areobserved between old and modern cruise conditions

(xvi) Strong effect of the fuel sulphur content is ob-served particle hygroscopicity increases distinctly withincreasing FSC

(xvii) No significant effect of the HES at low andmid FSC level is observed compared to the combustorexit an increased particle hygroscopicity downstreamthe HES compared to the combustor is observed at highFSC indicating continuing transformation of S(IV) toS(VI) during the transfer through the HES

(xviii) CCN activation of combustion particles in-creases with increasing FSC as a result of larger par-ticle hygroscopicity additional effects of adsorbed or-ganic matter cannot be excluded

45 Gaseous organic fraction

(xix) More than 100 different NMVOCs (aliphatic andaromatic hydrocarbons carbonyls and acids) were iden-tified and quantified they account for 91 wt of thetotal detected compounds

(xx) A decrease in NMHC emissions from old tomodern conditions at fixed HES pressure stage and fuelsulphur content (FSC) is found

(xxi) A decrease of NMHC emissions with decreas-ing pressure in the HES at fixed combustor operationconditions and FSC is observed

(xxii) No clear influence on the NMHC emissionswith changing FSC at the same combustor operationcondition is observed

(xxiii) High EIrsquos for organic acids are found Pos-sible explanations are that organic compounds fixed atthe sampling line walls are oxidised by the hot exhaustgases or that the acids are fixed at the surface of thesoot particles which are also trapped in the samplingcartridges

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 475

46 Emission properties

Refer to WILSON et al (2004) and PETZOLD et al(2003a)

(xxiv) Emission properties regarding combustionaerosol mass number and size are determined by thecombustor the influence of the Hot End Simulator isweak

(xxv) The processing of exhaust gas and parti-cles during the transfer through the Hot End Simula-tor increases the particle hygroscopicity and formationstrength of condensation particles both effects are re-lated to gaseous sulphuric acid

The experimental overview of the EU PartEmisproject presents the first comprehensive aerosol and gasmeasurements to be conducted on an aircraft enginecombustor test-rig The database will be used for fur-ther analyses as well as for plume modelling studies ofthe effect of aviation-related particle emissions on theglobal atmosphere and climate

Acknowledgements

The PartEmis project was funded by the European Com-mission under contract no G4RD-CT-2000-00207 Theauthors are very grateful to the test rig operation crew atQinetiQ for their very strong support during the experi-ments

References

ANDERSON BE WR COFER DR BAGWELL JWBARRICK CH HUDGINS KE BRUNKE 1998 Air-borne observations of aircraft aerosol emissions I Totalnonvolatile particle emission indices ndash Geophys Res Lett25 1689ndash1692

BOUCHER O 1999 Influence of air traffic on cirrus occur-rence Nature ndash 397 30ndash31

BROCK CA F SCHROumlDER B KAumlRCHER A PETZOLDR BUSEN M FIEBIG 2000 Ultrafine particle size distri-butions measured in aircraft exhaust plumes ndash J GeophysRes 105 26555ndash26567

CACHIER H MP BREMOND P BIAT-MEacuteNARD 1989Determination of atmospheric soot carbon with a simplethermal method ndash Tellus 41B 379ndash390

DEMOTT PJ Y CHEN SM KREIDENWEIS DCROGERS DE SHERMAN 1999 Ice formation by blackcarbon particles ndash Geophys Res Lett 26 2429ndash2432

GIEBL H A BERNER G REISCHL H PUXBAUM AKASPER-GIEBL R HITZENBERGER 2002 CCN activa-tion of oxalic and malonic acid test aerosols with the Uni-versity of Vienna cloud condensation nuclei counter ndash JAerosol Sci 33 1623ndash1634

GYSEL M E WEINGARTNER U BALTENSPERGER2002 Hygroscopicity of aerosol particles at low tempera-tures 2 Theoretical and experimental hygroscopic proper-ties of laboratory generated aerosols ndash Environ Sci Tech-nol 36 63ndash68

GYSEL M S NYEKI E WEINGARTNER U BAL-TENSPERGER H GIEBL R HITZENBERGER A PET-ZOLD CW WILSON 2003a Properties of jet engine com-bustor particles during the PartEmis experiment Hygro-scopicity at subsaturated conditions ndash Geophys Res Lett30 1566 DOI1010292003GL016896

GYSEL M S NYEKI E WEINGARTNER U BAL-TENSPERGER H GIEBL R HITZENBERGER A PET-ZOLD CW WILSON 2003b Jet engine combustion par-ticle hygroscopicity under subsaturated conditions duringPartEmis ndash J Aerosol Sci 34 (Suppl 2) Abstracts of theEuropean Aerosol Conference S1415ndashS1416

HAGEN DE PD WHITEFIELD H SCHLAGER 1996Particle emissions in the exhaust plume from commercialjet aircraft under cruise conditions ndash J Geophys Res 10119551ndash19557

HAGEN DE PD WHITEFIELD J PALADINO M TRUE-BLOOD H LILENFELD 1998 Particle sizing and emissionindices for a jet engine exhaust sampled at cruise ndash Geo-phys Res Lett 25 1681ndash1684

HAGEN DE PD WHITEFIELD JD PALADINO OSCHMID 2001 Comparative study of jet engine com-bustion emissions and engine operating conditions ndashIn SCHUMANN U G AMANATIDIS (Eds) Aviationaerosols contrails and cirrus clouds Air pollution researchreport 74 106ndash110 European Commission Brussels

HAVERKAMP H S WILHELM A SOROKIN F ARNOLD2004 Positive and negative ion measurements in jet aircraftengine exhaust concentrations sizes and implications foraerosol formation ndash Atmos Environ 38 2879ndash2884

HITZENBERGER R H GIEBL A PETZOLD M GYSELS NYEKI E WEINGARTNER U BALTENSPERGERCW WILSON 2003a Properties of jet engine combus-tor particles during the PartEmis experiment Hygroscopicproperties at supersaturated conditions ndash Geophys ResLett 30 1779 DOI1010292003GL017294

HITZENBERGER R H GIEBL A PETZOILD M GYSELS NYEKI E WEINGARTNER U BALTENSPERGERCW WILSON 2003b CCN activation of jet engine com-bustor particles during PartEmis ndash J Aerosol Sci 34(Suppl 2) Abstracts of the European Aerosol ConferenceS1409ndashS1410

HURLEY CD 1996 AEROTRACE Measurements of Par-ticulates from an Engine Combustor ndash Proceedings Impactof Aircraft Emissions upon the Atmosphere Vol 1 113 In-ternational Colloquium Paris October 1996

KATRAGKOU E S WILHELM F ARNOLD CW WIL-SON 2004 First gaseous Sulfur (VI) measurements in thesimulated internal flow of an aircraft gas turbine engineduring project PartEmis ndash Geophys Res Lett 31 2117DOI1010292003GL018231

NYEKI S M GYSEL E WEINGARTNER U BAL-TENSPERGER R HITZENBERGER A PETZOLD CWWILSON 2003 Volatile properties of jet engine combus-tor particles during PartEmis ndash J Aerosol Sci 34 (Suppl2) Abstracts of the European Aerosol Conference S1411ndashS1412

PETZOLD A FP SCHROumlDER 1998 Jet engine exhaustaerosol characterization ndash Aerosol Sci Technol 28 62ndash76

PETZOLD A A DOumlPELHEUER CA BROCK FPSCHROumlDER 1999 In situ observations and model calcula-

476 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

tions of black carbon emission by aircraft at cruise altitudendash J Geophys Res 104 22171ndash22181

PETZOLD A C STEIN S NYEKI M GYSEL EWEINGARTNER U BALTENSPERGER H GIEBL RHITZENBERGER A DOumlPELHEUER S VRCHOTICKY HPUXBAUM M JOHNSON CD HURLEY R MARSHCW WILSON 2003a Properties of jet engine combus-tor particles during the PartEmis experiment Microphysi-cal and chemical properties ndash Geophys Res Lett 30 1719DOI1010292003GL017283

PETZOLD A CW WILSON U BALTENSPERGERM FIEBIG L FRITZSCHE H GIEBL M GYSELR HITZENBERGER CD HURLEY S NYEKI HPUXBAUM U SCHUMANN C STEIN S VROCHTICKY2003b Particle emissions from aircraft engines ndash anoverview of the European project PartEmis ndash J AerosolSci 34 (Suppl 2) Abstracts of the European Aerosol Con-ference S1405ndashS1406

PETZOLD A AE VRTALA M FIEBIG F FRITZSCHEPH MIRABEL S NYEKI X VANCASSEL CW WIL-SON 2003c Emission of volatile and non-volatile ultra-fine particles from a combustion source during PartEmisndash J Aerosol Sci 34 (Suppl 2) Abstracts of the EuropeanAerosol Conference S1407ndashS1408

PUXBAUM H S VRCHOTICKY A PETZOLD R HITZEN-BERGER S NYEKI CW WILSON 2003 Chemical com-position of jet engine combustor particles during PartEmisndash J Aerosol Sci 34 (Suppl 2) Abstracts of the EuropeanAerosol Conference S1413ndashS1414

SCHMID H L LASKUS H-J ABRAHAM U BAL-TENSPERGER V LAVANCHY M BIZJAK P BURBA HCACHIER D CROW J CHOW T GNAUK A EVENHM TEN BRINK K-P GIESEN R HITZENBERGER CHUEGLIN W MAENHAUT C PIOA CARVALHO J-PPUTAUD D TOOM-SAUNTRY H PUXBAUM 2001 Re-sults of the ldquocarbon conferencerdquo international aerosol car-bon round robin test stage I ndash Atmos Environ 35 2111ndash2121

SCHROumlDER F C BROCK R BAUMANN A PETZOLD RBUSEN P SCHULTE M FIEBIG 2000 In-situ studies onvolatile jet exhaust particle emissions Impact of fuel sul-fur content and thermodynamic conditions on nuclei-modeaerosols ndash J Geophys Res 105 19941ndash19954

SCHUMANN U 1996 Contrail formation from aircraft ex-hausts ndash Meteorol Z NF 5 4ndash23

SCHUMANN U J STROumlM 2001 Aviation impact on atmo-spheric composition and climate ndash In European research inthe stratosphere 1996ndash2000 European Commission EUR19867 Brussels

SCHUMANN U F ARNOLD R BUSEN J CURTIUS BKAumlRCHER A KIENDLER A PETZOLD H SCHLAGERF SCHROumlDER K-H WOHLFROM 2002 Influence of fu-els sulphur on the composition of aircraft exhaust plumesThe experiments SULFUR 1-7 ndash J Geophys Res 107(1010292001JD000813) AAC 2-1 ndash AAC 2-27

SOROKIN A P MIRABEL F ARNOLD 2003 On the In-fluence of fuel fulfur induced stable negative ion formationon the total concentration of ions emitted by an aircraft gasturbine engine Comparison of model and experiment ndash At-mos Chem Phys 3 6001ndash6018

VANCASSEL X A SOROKIN P MIRABEL A PETZOLDCW WILSON 2004 Volatile particles formation duringPartEmis a modelling study ndash Atmos Chem Phys 4 439ndash447

WILHELM S H HAVERKAMP A SOROKIN F ARNOLD2004 Detection of very large ions in aircraft gas turbineengine combustor exhaustcharged small soot particles ndashAtmos Environ 38 4561ndash4569

WILSON CW A PETZOLD S NYEKI U SCHUMANNR ZELLNER 2004 Measurement and Prediction of Emis-sions of Aerosols and Gaseous Precursors from Gas Tur-bine Engines (PartEmis) An Overview ndash Aerosp SciTechnol 8 131ndash143

Page 8: Particle emissions from aircraft engines a survey of the ... · 466 A. Petzold et al.: Particle emissions from aircraft engines Meteorol. Z., 14 , 2005 1 Introduction The role of

472 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

is 23plusmn 12 at the low pressure stage for the mod-ern cruise condition and 14plusmn 07 for the old cruiseThe higher ε observed for the modern cruise conditionssuggests that modern engines have a larger ε comparedto old engines Conversion efficiencies determined fromin-flight data at cruise altitude gave values 04ndash45 (SCHUMANN et al 2002) which is in good agreementwith the PartEmis ground test values

Chemical composition information was obtainedfrom Berner low pressure impactor samples which weretaken at a sample line temperature of 150C (PUXBAUMet al 2003) At these conditions most of the sulphuricacid remains in the gas phase while only a small fractionis more strongly bound to or chemi-sorbed at the par-ticle surface and will be observed in particulate mattersamples The chemical analysis results indicate that themass distribution of the carbonaceous matter matchesthe volume distribution of the aerosol The mass dis-tribution of chemisorbed sulphate however follows thesurface distribution of the aerosol which was calculatedfrom aerosol size spectra assuming a spherical particleshape (PUXBAUM et al 2003) This observation indi-cates that the sulphate is chemisorbed at the particle sur-face The chemisorbed fraction of sulphuric acid corre-sponds to a coverage of max 01 monolayer in the highFSC case This fraction of converted S(VI) is less than01 of total S(IV) present in the fuel The overall par-titioning of sulphur-containing species at 150C is 97 SO2le 27 gaseous H2SO4lt 03 chemisorbedAfter cooling to environmental temperatures the gasphase sulphuric acid can condense and cover the com-bustion particles with about one monolayer of sulphuricacid on average (HITZENBRGER et al 2003a)

33 Particle hygroscopic properties

Particle hygroscopic properties were measured at ambi-ent conditions using a Hygroscopicity Tandem Differ-ential Mobility Analyser (H-TDMA) system (GYSEL etal 2003a 2003b) The instrument measures the hygro-scopic growth factor (g) of particles with initial dry sizesDo = 30 50 and 100 nm over the range RHsim 70minus95 where g(RH) = D(RH)Do Growth factors were found toincrease with increasing FSC at fixed particle size (egDo = 50 nm g (95 ) = 101 and 116 for FSC = lowand high respectively) and to decrease with increasingparticle size at fixed FSC (eg mid FSC g (95 ) =114 110 and 103 for Do = 30 50 and 100 nm respec-tively) The results suggest an increasing amount of sul-phuric acid adsorbed on combustion particles from thegas phase with increasing FSC

Cloud condensation nuclei (CCN) were measured inthe diluted line using a static thermal diffusion typeCCN counter (GIEBL et al 2002) Particles are exposedto controlled supersaturations of water vapour whichenables some of the particles to become activated and

Figure 5 Evolving CO2 as a function of the sample heating temper-ature for a Combustor aerosol sample semi-volatile OC oxidises at

450 K non-volatile OC at 600 K and elemental carbon above 700 K

grow to large (gt 10 microm) droplets During PartEmis thesupersaturation was set tosim 07 The fraction of CCNin the aerosol (relative to the concentration of particlesof Dgt 10 nm) increases with increasing FSC but evenfor low FSC the fraction of CCN (at 0755 supersat-uration) is higher than the fraction of particles with di-ameters larger than 280 nm ie the size at which wet-table insoluble particles are activated according to theKelvin effect At high FSC the fraction of CCN in-creases by about a factor 55 for old and 12 for mod-ern conditions However pure sulphuric acid particlesof the same size which become activated according toKOumlHLER theory would give an activation ratio which isstill larger by a factorsim 15 compared to the observationseven at high FSC Thus the KOumlHLER model cannot re-produce measured activation properties This is in agree-ment with the fact that large combustion particles are notat all completely water soluble The Kelvin model pro-vides good predictions for the low FSC level but acti-vation diameters (ratios) are increasingly overestimated(underestimated) at medium and high FSC This indi-cates that the soluble fraction of the combustion parti-cles increases from low values at low FSC level to rel-atively large values at high FSC An empirical coatedsphere model which equates the measured volatile andsemivolatile fraction with soluble sulphuric acid under-estimates the activation diameter by 45 to 55 underall conditions A small amount of semivolatile organ-ics with low or absent water-solubility adsorbed on thecombustion particles might be one reason for this dis-crepancy Further details are given by HITZENBERGER

et al (2003a 2003b)

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 473

34 Gaseous hydrocarbon emission

More than 100 different non-methane volatile organiccompounds (NMVOCs) such as non-methane hydrocar-bons (NMHCs) and selected partially oxidised hydro-carbons (OHCs) emitted from the aircraft engine com-bustor and the hot end simulator (HES) were identifiedand quantified These species accounted for up to 91wt of the total hydrocarbons (THCs) The NMHCemission indices of 5ndash15 microg kgminus1 in total are compa-rable with the AEROTRACE (Measurement of TraceSpecies in the Exhaust from Aero Engines EC projectAERA-CT94-0003 1994ndash1997 Co-ordinated by Qine-tiQ) study Both studies show up to 1000 times loweremission indices in comparison with emissions from ve-hicular traffic The influence of the operation conditionson NMHC emissions of the combustor and HES was in-vestigated for the identified NMHCs From the six con-ditions which were investigated (2 operation conditions3 HES pressure stages) most compounds show a de-crease of NMHC emissions with increasing combustiontemperature at the same pressure stage This is in agree-ment with results from the combustor exit measurementsand with the results of the AEROTRACE study Emis-sion indices at HES pressure stages high (HP)- inter-mediate (IP)- and low pressure (LP) for fixed FSC andcombustor operation conditions were also investigatedMost compounds show a decrease of NMHC emissionswith decreasing pressure in the HES at similar combus-tor operation conditions and fixed FSC The decrease ofthe pressure in the HES corresponds to an increase ofthe residence time in a turbine and indicates a decreaseof NMHC emissions along the different turbine sections

35 Emission of ions and charged sootparticles

The ions measured during the HES campaign(HAVERKAMP et al 2004 WILHELM et al 2004) areproduct ions formed from primary ions via ion moleculereactions involving either few reaction-steps withmassive neutral molecules or numerous reaction-stepswith relatively small neutral molecules Their lifetimeswith respect to ion-ion recombination are around 9 msafter the high pressure turbine section The ion emissionindex Ei is up to 45 times 1016 ions per kg fuel burntAs is discussed by HAVERKAMP et al (2004) neitherion-ion recombination nor ion removal by attachmentto combustion particles give sufficient explanations forthe observed ion emission index Therefore detailedmodel calculations of the ion chemistry taking place inthe combustor with particular emphasis on the sulphurspecies effect is required (HAVERKAMP et al 2004)It seems that ion-induced nucleation can explain atleast partly the observed volatile aerosols Electricallycharged small positive and negative soot particles

(CSP) have also been detected during both PartEmiscampaigns Charged soot particles are found to havediameters around 6 nm and a total concentration of upto 45times107 cmminus3 (positive and negative) correspondingto a total emission index of ECSP = 23times 1015 CSP perkg fuel burnt (WILHELM et al 2004)

A model which considers the formation and evolu-tion of combustion ions in a combustor of an aircraftengine in dependence on the electron detachment ef-ficiency from negative ions was used to consider theeffect of the transformation of primary negative ionsto more stable secondary negative ions (SOROKIN etal 2003) The formed stable negative ions most prob-ably are sulphur-bearing ions This effect slows downthe charged particle neutralisation rate leading to an in-crease of the concentration of positive and negative ionsat the combustor exit The results of the simulation andtheir comparison with the ground-based experimentaldata obtained within PartEmis support the hypothesisthat the increase of the fuel sulphur content leads to anincrease of the ion concentration at the combustor exit

36 OH measurements

The concentration of OH which is the dominant oxidantin the exhaust gas was detected by laser induced fluo-rescence (LIF) to obtain an absolute OH concentrationmeasurement in order to aid in the modelling of SO3production from the combustor to the engine exit Whenrunning the laser at low laser energies OH in the exhaustduct was successfully detected Calibration experimentsprovide a method to assign the observed OH LIF sig-nals an absolute concentration and from an estimate ormeasurement of the errors for each parameter an overallestimate of the error in the assigned OH concentrationwas made From these calibration experiments very lowvalues of less than 1 ppbv were calculated for the OHconcentration in the exhaust gas after exiting the HESPrevious experiments using a combined Raman scatter-ingLIF technique to look in the exhaust gases 50 cmbehind an aircraft engine exit failed to detect any OHLIF signal and therefore could only infer an upper limitfor OH of 80 ppbv based on the measured detection limitof their apparatus

4 Summary and conclusions

The progress made during PartEmis is manifold Themain conclusions are therefore grouped into differentsubject areas and presented as brief statements For eachsubject area key references are identified if available

41 Combustion aerosol

Refer to WILSON et al (2004) and PETZOLD et al(2003a)

474 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

(i) The combustor behaves like a typical aircraft en-gine combustor with respect to thermodynamic dataand main emissions It may thus be expected that thePartEmis database can be applied as a first order approx-imation to contemporary aircraft engines

(ii) The BC concentration is nearly independent ofFSC level

(iii) The number size distribution of primary particlesin the diameter range 10 lt Dlt 550 nm is almost unin-fluenced by the FSC level or combustor conditions Athigh FSC levels the formation of volatile condensationparticles with diameters Dlt 10 nm occurs

(iv) The volatile volume fraction of particles gt 50 nmincreases from 2 for low FSC to approx 5 for highFSC The volatile volume fraction of particles lt 30 nmis le 15

(v) Volatility measurements indicate an internallymixed aerosol for particles with D gt 15 nm conden-sation particles are below 15 nm in diameter

42 Ultrafine particles

Refer to PETZOLD et al (2003a) and VANCASSEL et al(2004)

(vi) Formation of condensation particles of size D=4minus 7 nm is observed in the Combustor tests for highFSC at old and modern conditions formation occurs ifthe surface area of the combustion aerosol falls below acertain threshold limit the observed number of formedparticles is confirmed by kinetic model studies

(vii) Formation of condensation particles is observedin the HES tests for high FSCmodern conditionslowpressure section only the ldquoformation strengthrdquo in termsof number of condensation particles per number of com-bustion particle is twice as high as for the Combustorcase

(viii) The model studies indicate that condensationparticles are not emitted from the combustor but form inthe sampling line at the sample dilution point

(ix) Ultrafine carbonaceous particles of size 7 nmlt D lt 20 nm are observed independently of the fuelsulphur content a volatility analysis indicates that theseparticles are nonvolatile and may be composed of car-bonaceous compounds

43 Sulphate-containing species chemi-ionsand OH

Refer to KATRAGKOU et al (2004) and HAVERKAMP etal (2004)

(x) S(VI) was measured in the form of sulphuric acidIncreased FSC results in higher S(VI) concentrationsThe measured conversion efficiency for fuel sulphur toS(IV) at the combustor exit varies from 035ndash14 andis independent of the investigated combustor operationconditions and FSC levels

(xi) 10ndash30 of sulphate is chemisorbed on primaryparticles the coverage of primary particles is below 01monolayers of sulphuric acid at 150C and approxi-mately one monolayer at 25C

(xii) Sulphur partitioning at 150C 97 SO2le 27 gaseous H2SO4 lt 03 chemisorbed

(xiii) Positive and negative gaseous ions were de-tected in the exhaust of a combustor with mass numbersup to 1500 The total ion emission index at the HES hpstage is as high as 54times 1017 positive and negative ionsper kg fuel burnt Large ions with mass numbers exceed-ing 4500 contribute about 10 to the total ion numberconcentration

(xiv) OH concentration in the exhaust gas after exit-ing the HES is lt 1 ppbv

44 Hygroscopic growth and CCN activation

Refer to GYSEL et al (2003a) and HITZENBERGER etal (2003a)

(xv) Small differences in particle hygroscopicity areobserved between old and modern cruise conditions

(xvi) Strong effect of the fuel sulphur content is ob-served particle hygroscopicity increases distinctly withincreasing FSC

(xvii) No significant effect of the HES at low andmid FSC level is observed compared to the combustorexit an increased particle hygroscopicity downstreamthe HES compared to the combustor is observed at highFSC indicating continuing transformation of S(IV) toS(VI) during the transfer through the HES

(xviii) CCN activation of combustion particles in-creases with increasing FSC as a result of larger par-ticle hygroscopicity additional effects of adsorbed or-ganic matter cannot be excluded

45 Gaseous organic fraction

(xix) More than 100 different NMVOCs (aliphatic andaromatic hydrocarbons carbonyls and acids) were iden-tified and quantified they account for 91 wt of thetotal detected compounds

(xx) A decrease in NMHC emissions from old tomodern conditions at fixed HES pressure stage and fuelsulphur content (FSC) is found

(xxi) A decrease of NMHC emissions with decreas-ing pressure in the HES at fixed combustor operationconditions and FSC is observed

(xxii) No clear influence on the NMHC emissionswith changing FSC at the same combustor operationcondition is observed

(xxiii) High EIrsquos for organic acids are found Pos-sible explanations are that organic compounds fixed atthe sampling line walls are oxidised by the hot exhaustgases or that the acids are fixed at the surface of thesoot particles which are also trapped in the samplingcartridges

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 475

46 Emission properties

Refer to WILSON et al (2004) and PETZOLD et al(2003a)

(xxiv) Emission properties regarding combustionaerosol mass number and size are determined by thecombustor the influence of the Hot End Simulator isweak

(xxv) The processing of exhaust gas and parti-cles during the transfer through the Hot End Simula-tor increases the particle hygroscopicity and formationstrength of condensation particles both effects are re-lated to gaseous sulphuric acid

The experimental overview of the EU PartEmisproject presents the first comprehensive aerosol and gasmeasurements to be conducted on an aircraft enginecombustor test-rig The database will be used for fur-ther analyses as well as for plume modelling studies ofthe effect of aviation-related particle emissions on theglobal atmosphere and climate

Acknowledgements

The PartEmis project was funded by the European Com-mission under contract no G4RD-CT-2000-00207 Theauthors are very grateful to the test rig operation crew atQinetiQ for their very strong support during the experi-ments

References

ANDERSON BE WR COFER DR BAGWELL JWBARRICK CH HUDGINS KE BRUNKE 1998 Air-borne observations of aircraft aerosol emissions I Totalnonvolatile particle emission indices ndash Geophys Res Lett25 1689ndash1692

BOUCHER O 1999 Influence of air traffic on cirrus occur-rence Nature ndash 397 30ndash31

BROCK CA F SCHROumlDER B KAumlRCHER A PETZOLDR BUSEN M FIEBIG 2000 Ultrafine particle size distri-butions measured in aircraft exhaust plumes ndash J GeophysRes 105 26555ndash26567

CACHIER H MP BREMOND P BIAT-MEacuteNARD 1989Determination of atmospheric soot carbon with a simplethermal method ndash Tellus 41B 379ndash390

DEMOTT PJ Y CHEN SM KREIDENWEIS DCROGERS DE SHERMAN 1999 Ice formation by blackcarbon particles ndash Geophys Res Lett 26 2429ndash2432

GIEBL H A BERNER G REISCHL H PUXBAUM AKASPER-GIEBL R HITZENBERGER 2002 CCN activa-tion of oxalic and malonic acid test aerosols with the Uni-versity of Vienna cloud condensation nuclei counter ndash JAerosol Sci 33 1623ndash1634

GYSEL M E WEINGARTNER U BALTENSPERGER2002 Hygroscopicity of aerosol particles at low tempera-tures 2 Theoretical and experimental hygroscopic proper-ties of laboratory generated aerosols ndash Environ Sci Tech-nol 36 63ndash68

GYSEL M S NYEKI E WEINGARTNER U BAL-TENSPERGER H GIEBL R HITZENBERGER A PET-ZOLD CW WILSON 2003a Properties of jet engine com-bustor particles during the PartEmis experiment Hygro-scopicity at subsaturated conditions ndash Geophys Res Lett30 1566 DOI1010292003GL016896

GYSEL M S NYEKI E WEINGARTNER U BAL-TENSPERGER H GIEBL R HITZENBERGER A PET-ZOLD CW WILSON 2003b Jet engine combustion par-ticle hygroscopicity under subsaturated conditions duringPartEmis ndash J Aerosol Sci 34 (Suppl 2) Abstracts of theEuropean Aerosol Conference S1415ndashS1416

HAGEN DE PD WHITEFIELD H SCHLAGER 1996Particle emissions in the exhaust plume from commercialjet aircraft under cruise conditions ndash J Geophys Res 10119551ndash19557

HAGEN DE PD WHITEFIELD J PALADINO M TRUE-BLOOD H LILENFELD 1998 Particle sizing and emissionindices for a jet engine exhaust sampled at cruise ndash Geo-phys Res Lett 25 1681ndash1684

HAGEN DE PD WHITEFIELD JD PALADINO OSCHMID 2001 Comparative study of jet engine com-bustion emissions and engine operating conditions ndashIn SCHUMANN U G AMANATIDIS (Eds) Aviationaerosols contrails and cirrus clouds Air pollution researchreport 74 106ndash110 European Commission Brussels

HAVERKAMP H S WILHELM A SOROKIN F ARNOLD2004 Positive and negative ion measurements in jet aircraftengine exhaust concentrations sizes and implications foraerosol formation ndash Atmos Environ 38 2879ndash2884

HITZENBERGER R H GIEBL A PETZOLD M GYSELS NYEKI E WEINGARTNER U BALTENSPERGERCW WILSON 2003a Properties of jet engine combus-tor particles during the PartEmis experiment Hygroscopicproperties at supersaturated conditions ndash Geophys ResLett 30 1779 DOI1010292003GL017294

HITZENBERGER R H GIEBL A PETZOILD M GYSELS NYEKI E WEINGARTNER U BALTENSPERGERCW WILSON 2003b CCN activation of jet engine com-bustor particles during PartEmis ndash J Aerosol Sci 34(Suppl 2) Abstracts of the European Aerosol ConferenceS1409ndashS1410

HURLEY CD 1996 AEROTRACE Measurements of Par-ticulates from an Engine Combustor ndash Proceedings Impactof Aircraft Emissions upon the Atmosphere Vol 1 113 In-ternational Colloquium Paris October 1996

KATRAGKOU E S WILHELM F ARNOLD CW WIL-SON 2004 First gaseous Sulfur (VI) measurements in thesimulated internal flow of an aircraft gas turbine engineduring project PartEmis ndash Geophys Res Lett 31 2117DOI1010292003GL018231

NYEKI S M GYSEL E WEINGARTNER U BAL-TENSPERGER R HITZENBERGER A PETZOLD CWWILSON 2003 Volatile properties of jet engine combus-tor particles during PartEmis ndash J Aerosol Sci 34 (Suppl2) Abstracts of the European Aerosol Conference S1411ndashS1412

PETZOLD A FP SCHROumlDER 1998 Jet engine exhaustaerosol characterization ndash Aerosol Sci Technol 28 62ndash76

PETZOLD A A DOumlPELHEUER CA BROCK FPSCHROumlDER 1999 In situ observations and model calcula-

476 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

tions of black carbon emission by aircraft at cruise altitudendash J Geophys Res 104 22171ndash22181

PETZOLD A C STEIN S NYEKI M GYSEL EWEINGARTNER U BALTENSPERGER H GIEBL RHITZENBERGER A DOumlPELHEUER S VRCHOTICKY HPUXBAUM M JOHNSON CD HURLEY R MARSHCW WILSON 2003a Properties of jet engine combus-tor particles during the PartEmis experiment Microphysi-cal and chemical properties ndash Geophys Res Lett 30 1719DOI1010292003GL017283

PETZOLD A CW WILSON U BALTENSPERGERM FIEBIG L FRITZSCHE H GIEBL M GYSELR HITZENBERGER CD HURLEY S NYEKI HPUXBAUM U SCHUMANN C STEIN S VROCHTICKY2003b Particle emissions from aircraft engines ndash anoverview of the European project PartEmis ndash J AerosolSci 34 (Suppl 2) Abstracts of the European Aerosol Con-ference S1405ndashS1406

PETZOLD A AE VRTALA M FIEBIG F FRITZSCHEPH MIRABEL S NYEKI X VANCASSEL CW WIL-SON 2003c Emission of volatile and non-volatile ultra-fine particles from a combustion source during PartEmisndash J Aerosol Sci 34 (Suppl 2) Abstracts of the EuropeanAerosol Conference S1407ndashS1408

PUXBAUM H S VRCHOTICKY A PETZOLD R HITZEN-BERGER S NYEKI CW WILSON 2003 Chemical com-position of jet engine combustor particles during PartEmisndash J Aerosol Sci 34 (Suppl 2) Abstracts of the EuropeanAerosol Conference S1413ndashS1414

SCHMID H L LASKUS H-J ABRAHAM U BAL-TENSPERGER V LAVANCHY M BIZJAK P BURBA HCACHIER D CROW J CHOW T GNAUK A EVENHM TEN BRINK K-P GIESEN R HITZENBERGER CHUEGLIN W MAENHAUT C PIOA CARVALHO J-PPUTAUD D TOOM-SAUNTRY H PUXBAUM 2001 Re-sults of the ldquocarbon conferencerdquo international aerosol car-bon round robin test stage I ndash Atmos Environ 35 2111ndash2121

SCHROumlDER F C BROCK R BAUMANN A PETZOLD RBUSEN P SCHULTE M FIEBIG 2000 In-situ studies onvolatile jet exhaust particle emissions Impact of fuel sul-fur content and thermodynamic conditions on nuclei-modeaerosols ndash J Geophys Res 105 19941ndash19954

SCHUMANN U 1996 Contrail formation from aircraft ex-hausts ndash Meteorol Z NF 5 4ndash23

SCHUMANN U J STROumlM 2001 Aviation impact on atmo-spheric composition and climate ndash In European research inthe stratosphere 1996ndash2000 European Commission EUR19867 Brussels

SCHUMANN U F ARNOLD R BUSEN J CURTIUS BKAumlRCHER A KIENDLER A PETZOLD H SCHLAGERF SCHROumlDER K-H WOHLFROM 2002 Influence of fu-els sulphur on the composition of aircraft exhaust plumesThe experiments SULFUR 1-7 ndash J Geophys Res 107(1010292001JD000813) AAC 2-1 ndash AAC 2-27

SOROKIN A P MIRABEL F ARNOLD 2003 On the In-fluence of fuel fulfur induced stable negative ion formationon the total concentration of ions emitted by an aircraft gasturbine engine Comparison of model and experiment ndash At-mos Chem Phys 3 6001ndash6018

VANCASSEL X A SOROKIN P MIRABEL A PETZOLDCW WILSON 2004 Volatile particles formation duringPartEmis a modelling study ndash Atmos Chem Phys 4 439ndash447

WILHELM S H HAVERKAMP A SOROKIN F ARNOLD2004 Detection of very large ions in aircraft gas turbineengine combustor exhaustcharged small soot particles ndashAtmos Environ 38 4561ndash4569

WILSON CW A PETZOLD S NYEKI U SCHUMANNR ZELLNER 2004 Measurement and Prediction of Emis-sions of Aerosols and Gaseous Precursors from Gas Tur-bine Engines (PartEmis) An Overview ndash Aerosp SciTechnol 8 131ndash143

Page 9: Particle emissions from aircraft engines a survey of the ... · 466 A. Petzold et al.: Particle emissions from aircraft engines Meteorol. Z., 14 , 2005 1 Introduction The role of

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 473

34 Gaseous hydrocarbon emission

More than 100 different non-methane volatile organiccompounds (NMVOCs) such as non-methane hydrocar-bons (NMHCs) and selected partially oxidised hydro-carbons (OHCs) emitted from the aircraft engine com-bustor and the hot end simulator (HES) were identifiedand quantified These species accounted for up to 91wt of the total hydrocarbons (THCs) The NMHCemission indices of 5ndash15 microg kgminus1 in total are compa-rable with the AEROTRACE (Measurement of TraceSpecies in the Exhaust from Aero Engines EC projectAERA-CT94-0003 1994ndash1997 Co-ordinated by Qine-tiQ) study Both studies show up to 1000 times loweremission indices in comparison with emissions from ve-hicular traffic The influence of the operation conditionson NMHC emissions of the combustor and HES was in-vestigated for the identified NMHCs From the six con-ditions which were investigated (2 operation conditions3 HES pressure stages) most compounds show a de-crease of NMHC emissions with increasing combustiontemperature at the same pressure stage This is in agree-ment with results from the combustor exit measurementsand with the results of the AEROTRACE study Emis-sion indices at HES pressure stages high (HP)- inter-mediate (IP)- and low pressure (LP) for fixed FSC andcombustor operation conditions were also investigatedMost compounds show a decrease of NMHC emissionswith decreasing pressure in the HES at similar combus-tor operation conditions and fixed FSC The decrease ofthe pressure in the HES corresponds to an increase ofthe residence time in a turbine and indicates a decreaseof NMHC emissions along the different turbine sections

35 Emission of ions and charged sootparticles

The ions measured during the HES campaign(HAVERKAMP et al 2004 WILHELM et al 2004) areproduct ions formed from primary ions via ion moleculereactions involving either few reaction-steps withmassive neutral molecules or numerous reaction-stepswith relatively small neutral molecules Their lifetimeswith respect to ion-ion recombination are around 9 msafter the high pressure turbine section The ion emissionindex Ei is up to 45 times 1016 ions per kg fuel burntAs is discussed by HAVERKAMP et al (2004) neitherion-ion recombination nor ion removal by attachmentto combustion particles give sufficient explanations forthe observed ion emission index Therefore detailedmodel calculations of the ion chemistry taking place inthe combustor with particular emphasis on the sulphurspecies effect is required (HAVERKAMP et al 2004)It seems that ion-induced nucleation can explain atleast partly the observed volatile aerosols Electricallycharged small positive and negative soot particles

(CSP) have also been detected during both PartEmiscampaigns Charged soot particles are found to havediameters around 6 nm and a total concentration of upto 45times107 cmminus3 (positive and negative) correspondingto a total emission index of ECSP = 23times 1015 CSP perkg fuel burnt (WILHELM et al 2004)

A model which considers the formation and evolu-tion of combustion ions in a combustor of an aircraftengine in dependence on the electron detachment ef-ficiency from negative ions was used to consider theeffect of the transformation of primary negative ionsto more stable secondary negative ions (SOROKIN etal 2003) The formed stable negative ions most prob-ably are sulphur-bearing ions This effect slows downthe charged particle neutralisation rate leading to an in-crease of the concentration of positive and negative ionsat the combustor exit The results of the simulation andtheir comparison with the ground-based experimentaldata obtained within PartEmis support the hypothesisthat the increase of the fuel sulphur content leads to anincrease of the ion concentration at the combustor exit

36 OH measurements

The concentration of OH which is the dominant oxidantin the exhaust gas was detected by laser induced fluo-rescence (LIF) to obtain an absolute OH concentrationmeasurement in order to aid in the modelling of SO3production from the combustor to the engine exit Whenrunning the laser at low laser energies OH in the exhaustduct was successfully detected Calibration experimentsprovide a method to assign the observed OH LIF sig-nals an absolute concentration and from an estimate ormeasurement of the errors for each parameter an overallestimate of the error in the assigned OH concentrationwas made From these calibration experiments very lowvalues of less than 1 ppbv were calculated for the OHconcentration in the exhaust gas after exiting the HESPrevious experiments using a combined Raman scatter-ingLIF technique to look in the exhaust gases 50 cmbehind an aircraft engine exit failed to detect any OHLIF signal and therefore could only infer an upper limitfor OH of 80 ppbv based on the measured detection limitof their apparatus

4 Summary and conclusions

The progress made during PartEmis is manifold Themain conclusions are therefore grouped into differentsubject areas and presented as brief statements For eachsubject area key references are identified if available

41 Combustion aerosol

Refer to WILSON et al (2004) and PETZOLD et al(2003a)

474 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

(i) The combustor behaves like a typical aircraft en-gine combustor with respect to thermodynamic dataand main emissions It may thus be expected that thePartEmis database can be applied as a first order approx-imation to contemporary aircraft engines

(ii) The BC concentration is nearly independent ofFSC level

(iii) The number size distribution of primary particlesin the diameter range 10 lt Dlt 550 nm is almost unin-fluenced by the FSC level or combustor conditions Athigh FSC levels the formation of volatile condensationparticles with diameters Dlt 10 nm occurs

(iv) The volatile volume fraction of particles gt 50 nmincreases from 2 for low FSC to approx 5 for highFSC The volatile volume fraction of particles lt 30 nmis le 15

(v) Volatility measurements indicate an internallymixed aerosol for particles with D gt 15 nm conden-sation particles are below 15 nm in diameter

42 Ultrafine particles

Refer to PETZOLD et al (2003a) and VANCASSEL et al(2004)

(vi) Formation of condensation particles of size D=4minus 7 nm is observed in the Combustor tests for highFSC at old and modern conditions formation occurs ifthe surface area of the combustion aerosol falls below acertain threshold limit the observed number of formedparticles is confirmed by kinetic model studies

(vii) Formation of condensation particles is observedin the HES tests for high FSCmodern conditionslowpressure section only the ldquoformation strengthrdquo in termsof number of condensation particles per number of com-bustion particle is twice as high as for the Combustorcase

(viii) The model studies indicate that condensationparticles are not emitted from the combustor but form inthe sampling line at the sample dilution point

(ix) Ultrafine carbonaceous particles of size 7 nmlt D lt 20 nm are observed independently of the fuelsulphur content a volatility analysis indicates that theseparticles are nonvolatile and may be composed of car-bonaceous compounds

43 Sulphate-containing species chemi-ionsand OH

Refer to KATRAGKOU et al (2004) and HAVERKAMP etal (2004)

(x) S(VI) was measured in the form of sulphuric acidIncreased FSC results in higher S(VI) concentrationsThe measured conversion efficiency for fuel sulphur toS(IV) at the combustor exit varies from 035ndash14 andis independent of the investigated combustor operationconditions and FSC levels

(xi) 10ndash30 of sulphate is chemisorbed on primaryparticles the coverage of primary particles is below 01monolayers of sulphuric acid at 150C and approxi-mately one monolayer at 25C

(xii) Sulphur partitioning at 150C 97 SO2le 27 gaseous H2SO4 lt 03 chemisorbed

(xiii) Positive and negative gaseous ions were de-tected in the exhaust of a combustor with mass numbersup to 1500 The total ion emission index at the HES hpstage is as high as 54times 1017 positive and negative ionsper kg fuel burnt Large ions with mass numbers exceed-ing 4500 contribute about 10 to the total ion numberconcentration

(xiv) OH concentration in the exhaust gas after exit-ing the HES is lt 1 ppbv

44 Hygroscopic growth and CCN activation

Refer to GYSEL et al (2003a) and HITZENBERGER etal (2003a)

(xv) Small differences in particle hygroscopicity areobserved between old and modern cruise conditions

(xvi) Strong effect of the fuel sulphur content is ob-served particle hygroscopicity increases distinctly withincreasing FSC

(xvii) No significant effect of the HES at low andmid FSC level is observed compared to the combustorexit an increased particle hygroscopicity downstreamthe HES compared to the combustor is observed at highFSC indicating continuing transformation of S(IV) toS(VI) during the transfer through the HES

(xviii) CCN activation of combustion particles in-creases with increasing FSC as a result of larger par-ticle hygroscopicity additional effects of adsorbed or-ganic matter cannot be excluded

45 Gaseous organic fraction

(xix) More than 100 different NMVOCs (aliphatic andaromatic hydrocarbons carbonyls and acids) were iden-tified and quantified they account for 91 wt of thetotal detected compounds

(xx) A decrease in NMHC emissions from old tomodern conditions at fixed HES pressure stage and fuelsulphur content (FSC) is found

(xxi) A decrease of NMHC emissions with decreas-ing pressure in the HES at fixed combustor operationconditions and FSC is observed

(xxii) No clear influence on the NMHC emissionswith changing FSC at the same combustor operationcondition is observed

(xxiii) High EIrsquos for organic acids are found Pos-sible explanations are that organic compounds fixed atthe sampling line walls are oxidised by the hot exhaustgases or that the acids are fixed at the surface of thesoot particles which are also trapped in the samplingcartridges

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 475

46 Emission properties

Refer to WILSON et al (2004) and PETZOLD et al(2003a)

(xxiv) Emission properties regarding combustionaerosol mass number and size are determined by thecombustor the influence of the Hot End Simulator isweak

(xxv) The processing of exhaust gas and parti-cles during the transfer through the Hot End Simula-tor increases the particle hygroscopicity and formationstrength of condensation particles both effects are re-lated to gaseous sulphuric acid

The experimental overview of the EU PartEmisproject presents the first comprehensive aerosol and gasmeasurements to be conducted on an aircraft enginecombustor test-rig The database will be used for fur-ther analyses as well as for plume modelling studies ofthe effect of aviation-related particle emissions on theglobal atmosphere and climate

Acknowledgements

The PartEmis project was funded by the European Com-mission under contract no G4RD-CT-2000-00207 Theauthors are very grateful to the test rig operation crew atQinetiQ for their very strong support during the experi-ments

References

ANDERSON BE WR COFER DR BAGWELL JWBARRICK CH HUDGINS KE BRUNKE 1998 Air-borne observations of aircraft aerosol emissions I Totalnonvolatile particle emission indices ndash Geophys Res Lett25 1689ndash1692

BOUCHER O 1999 Influence of air traffic on cirrus occur-rence Nature ndash 397 30ndash31

BROCK CA F SCHROumlDER B KAumlRCHER A PETZOLDR BUSEN M FIEBIG 2000 Ultrafine particle size distri-butions measured in aircraft exhaust plumes ndash J GeophysRes 105 26555ndash26567

CACHIER H MP BREMOND P BIAT-MEacuteNARD 1989Determination of atmospheric soot carbon with a simplethermal method ndash Tellus 41B 379ndash390

DEMOTT PJ Y CHEN SM KREIDENWEIS DCROGERS DE SHERMAN 1999 Ice formation by blackcarbon particles ndash Geophys Res Lett 26 2429ndash2432

GIEBL H A BERNER G REISCHL H PUXBAUM AKASPER-GIEBL R HITZENBERGER 2002 CCN activa-tion of oxalic and malonic acid test aerosols with the Uni-versity of Vienna cloud condensation nuclei counter ndash JAerosol Sci 33 1623ndash1634

GYSEL M E WEINGARTNER U BALTENSPERGER2002 Hygroscopicity of aerosol particles at low tempera-tures 2 Theoretical and experimental hygroscopic proper-ties of laboratory generated aerosols ndash Environ Sci Tech-nol 36 63ndash68

GYSEL M S NYEKI E WEINGARTNER U BAL-TENSPERGER H GIEBL R HITZENBERGER A PET-ZOLD CW WILSON 2003a Properties of jet engine com-bustor particles during the PartEmis experiment Hygro-scopicity at subsaturated conditions ndash Geophys Res Lett30 1566 DOI1010292003GL016896

GYSEL M S NYEKI E WEINGARTNER U BAL-TENSPERGER H GIEBL R HITZENBERGER A PET-ZOLD CW WILSON 2003b Jet engine combustion par-ticle hygroscopicity under subsaturated conditions duringPartEmis ndash J Aerosol Sci 34 (Suppl 2) Abstracts of theEuropean Aerosol Conference S1415ndashS1416

HAGEN DE PD WHITEFIELD H SCHLAGER 1996Particle emissions in the exhaust plume from commercialjet aircraft under cruise conditions ndash J Geophys Res 10119551ndash19557

HAGEN DE PD WHITEFIELD J PALADINO M TRUE-BLOOD H LILENFELD 1998 Particle sizing and emissionindices for a jet engine exhaust sampled at cruise ndash Geo-phys Res Lett 25 1681ndash1684

HAGEN DE PD WHITEFIELD JD PALADINO OSCHMID 2001 Comparative study of jet engine com-bustion emissions and engine operating conditions ndashIn SCHUMANN U G AMANATIDIS (Eds) Aviationaerosols contrails and cirrus clouds Air pollution researchreport 74 106ndash110 European Commission Brussels

HAVERKAMP H S WILHELM A SOROKIN F ARNOLD2004 Positive and negative ion measurements in jet aircraftengine exhaust concentrations sizes and implications foraerosol formation ndash Atmos Environ 38 2879ndash2884

HITZENBERGER R H GIEBL A PETZOLD M GYSELS NYEKI E WEINGARTNER U BALTENSPERGERCW WILSON 2003a Properties of jet engine combus-tor particles during the PartEmis experiment Hygroscopicproperties at supersaturated conditions ndash Geophys ResLett 30 1779 DOI1010292003GL017294

HITZENBERGER R H GIEBL A PETZOILD M GYSELS NYEKI E WEINGARTNER U BALTENSPERGERCW WILSON 2003b CCN activation of jet engine com-bustor particles during PartEmis ndash J Aerosol Sci 34(Suppl 2) Abstracts of the European Aerosol ConferenceS1409ndashS1410

HURLEY CD 1996 AEROTRACE Measurements of Par-ticulates from an Engine Combustor ndash Proceedings Impactof Aircraft Emissions upon the Atmosphere Vol 1 113 In-ternational Colloquium Paris October 1996

KATRAGKOU E S WILHELM F ARNOLD CW WIL-SON 2004 First gaseous Sulfur (VI) measurements in thesimulated internal flow of an aircraft gas turbine engineduring project PartEmis ndash Geophys Res Lett 31 2117DOI1010292003GL018231

NYEKI S M GYSEL E WEINGARTNER U BAL-TENSPERGER R HITZENBERGER A PETZOLD CWWILSON 2003 Volatile properties of jet engine combus-tor particles during PartEmis ndash J Aerosol Sci 34 (Suppl2) Abstracts of the European Aerosol Conference S1411ndashS1412

PETZOLD A FP SCHROumlDER 1998 Jet engine exhaustaerosol characterization ndash Aerosol Sci Technol 28 62ndash76

PETZOLD A A DOumlPELHEUER CA BROCK FPSCHROumlDER 1999 In situ observations and model calcula-

476 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

tions of black carbon emission by aircraft at cruise altitudendash J Geophys Res 104 22171ndash22181

PETZOLD A C STEIN S NYEKI M GYSEL EWEINGARTNER U BALTENSPERGER H GIEBL RHITZENBERGER A DOumlPELHEUER S VRCHOTICKY HPUXBAUM M JOHNSON CD HURLEY R MARSHCW WILSON 2003a Properties of jet engine combus-tor particles during the PartEmis experiment Microphysi-cal and chemical properties ndash Geophys Res Lett 30 1719DOI1010292003GL017283

PETZOLD A CW WILSON U BALTENSPERGERM FIEBIG L FRITZSCHE H GIEBL M GYSELR HITZENBERGER CD HURLEY S NYEKI HPUXBAUM U SCHUMANN C STEIN S VROCHTICKY2003b Particle emissions from aircraft engines ndash anoverview of the European project PartEmis ndash J AerosolSci 34 (Suppl 2) Abstracts of the European Aerosol Con-ference S1405ndashS1406

PETZOLD A AE VRTALA M FIEBIG F FRITZSCHEPH MIRABEL S NYEKI X VANCASSEL CW WIL-SON 2003c Emission of volatile and non-volatile ultra-fine particles from a combustion source during PartEmisndash J Aerosol Sci 34 (Suppl 2) Abstracts of the EuropeanAerosol Conference S1407ndashS1408

PUXBAUM H S VRCHOTICKY A PETZOLD R HITZEN-BERGER S NYEKI CW WILSON 2003 Chemical com-position of jet engine combustor particles during PartEmisndash J Aerosol Sci 34 (Suppl 2) Abstracts of the EuropeanAerosol Conference S1413ndashS1414

SCHMID H L LASKUS H-J ABRAHAM U BAL-TENSPERGER V LAVANCHY M BIZJAK P BURBA HCACHIER D CROW J CHOW T GNAUK A EVENHM TEN BRINK K-P GIESEN R HITZENBERGER CHUEGLIN W MAENHAUT C PIOA CARVALHO J-PPUTAUD D TOOM-SAUNTRY H PUXBAUM 2001 Re-sults of the ldquocarbon conferencerdquo international aerosol car-bon round robin test stage I ndash Atmos Environ 35 2111ndash2121

SCHROumlDER F C BROCK R BAUMANN A PETZOLD RBUSEN P SCHULTE M FIEBIG 2000 In-situ studies onvolatile jet exhaust particle emissions Impact of fuel sul-fur content and thermodynamic conditions on nuclei-modeaerosols ndash J Geophys Res 105 19941ndash19954

SCHUMANN U 1996 Contrail formation from aircraft ex-hausts ndash Meteorol Z NF 5 4ndash23

SCHUMANN U J STROumlM 2001 Aviation impact on atmo-spheric composition and climate ndash In European research inthe stratosphere 1996ndash2000 European Commission EUR19867 Brussels

SCHUMANN U F ARNOLD R BUSEN J CURTIUS BKAumlRCHER A KIENDLER A PETZOLD H SCHLAGERF SCHROumlDER K-H WOHLFROM 2002 Influence of fu-els sulphur on the composition of aircraft exhaust plumesThe experiments SULFUR 1-7 ndash J Geophys Res 107(1010292001JD000813) AAC 2-1 ndash AAC 2-27

SOROKIN A P MIRABEL F ARNOLD 2003 On the In-fluence of fuel fulfur induced stable negative ion formationon the total concentration of ions emitted by an aircraft gasturbine engine Comparison of model and experiment ndash At-mos Chem Phys 3 6001ndash6018

VANCASSEL X A SOROKIN P MIRABEL A PETZOLDCW WILSON 2004 Volatile particles formation duringPartEmis a modelling study ndash Atmos Chem Phys 4 439ndash447

WILHELM S H HAVERKAMP A SOROKIN F ARNOLD2004 Detection of very large ions in aircraft gas turbineengine combustor exhaustcharged small soot particles ndashAtmos Environ 38 4561ndash4569

WILSON CW A PETZOLD S NYEKI U SCHUMANNR ZELLNER 2004 Measurement and Prediction of Emis-sions of Aerosols and Gaseous Precursors from Gas Tur-bine Engines (PartEmis) An Overview ndash Aerosp SciTechnol 8 131ndash143

Page 10: Particle emissions from aircraft engines a survey of the ... · 466 A. Petzold et al.: Particle emissions from aircraft engines Meteorol. Z., 14 , 2005 1 Introduction The role of

474 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

(i) The combustor behaves like a typical aircraft en-gine combustor with respect to thermodynamic dataand main emissions It may thus be expected that thePartEmis database can be applied as a first order approx-imation to contemporary aircraft engines

(ii) The BC concentration is nearly independent ofFSC level

(iii) The number size distribution of primary particlesin the diameter range 10 lt Dlt 550 nm is almost unin-fluenced by the FSC level or combustor conditions Athigh FSC levels the formation of volatile condensationparticles with diameters Dlt 10 nm occurs

(iv) The volatile volume fraction of particles gt 50 nmincreases from 2 for low FSC to approx 5 for highFSC The volatile volume fraction of particles lt 30 nmis le 15

(v) Volatility measurements indicate an internallymixed aerosol for particles with D gt 15 nm conden-sation particles are below 15 nm in diameter

42 Ultrafine particles

Refer to PETZOLD et al (2003a) and VANCASSEL et al(2004)

(vi) Formation of condensation particles of size D=4minus 7 nm is observed in the Combustor tests for highFSC at old and modern conditions formation occurs ifthe surface area of the combustion aerosol falls below acertain threshold limit the observed number of formedparticles is confirmed by kinetic model studies

(vii) Formation of condensation particles is observedin the HES tests for high FSCmodern conditionslowpressure section only the ldquoformation strengthrdquo in termsof number of condensation particles per number of com-bustion particle is twice as high as for the Combustorcase

(viii) The model studies indicate that condensationparticles are not emitted from the combustor but form inthe sampling line at the sample dilution point

(ix) Ultrafine carbonaceous particles of size 7 nmlt D lt 20 nm are observed independently of the fuelsulphur content a volatility analysis indicates that theseparticles are nonvolatile and may be composed of car-bonaceous compounds

43 Sulphate-containing species chemi-ionsand OH

Refer to KATRAGKOU et al (2004) and HAVERKAMP etal (2004)

(x) S(VI) was measured in the form of sulphuric acidIncreased FSC results in higher S(VI) concentrationsThe measured conversion efficiency for fuel sulphur toS(IV) at the combustor exit varies from 035ndash14 andis independent of the investigated combustor operationconditions and FSC levels

(xi) 10ndash30 of sulphate is chemisorbed on primaryparticles the coverage of primary particles is below 01monolayers of sulphuric acid at 150C and approxi-mately one monolayer at 25C

(xii) Sulphur partitioning at 150C 97 SO2le 27 gaseous H2SO4 lt 03 chemisorbed

(xiii) Positive and negative gaseous ions were de-tected in the exhaust of a combustor with mass numbersup to 1500 The total ion emission index at the HES hpstage is as high as 54times 1017 positive and negative ionsper kg fuel burnt Large ions with mass numbers exceed-ing 4500 contribute about 10 to the total ion numberconcentration

(xiv) OH concentration in the exhaust gas after exit-ing the HES is lt 1 ppbv

44 Hygroscopic growth and CCN activation

Refer to GYSEL et al (2003a) and HITZENBERGER etal (2003a)

(xv) Small differences in particle hygroscopicity areobserved between old and modern cruise conditions

(xvi) Strong effect of the fuel sulphur content is ob-served particle hygroscopicity increases distinctly withincreasing FSC

(xvii) No significant effect of the HES at low andmid FSC level is observed compared to the combustorexit an increased particle hygroscopicity downstreamthe HES compared to the combustor is observed at highFSC indicating continuing transformation of S(IV) toS(VI) during the transfer through the HES

(xviii) CCN activation of combustion particles in-creases with increasing FSC as a result of larger par-ticle hygroscopicity additional effects of adsorbed or-ganic matter cannot be excluded

45 Gaseous organic fraction

(xix) More than 100 different NMVOCs (aliphatic andaromatic hydrocarbons carbonyls and acids) were iden-tified and quantified they account for 91 wt of thetotal detected compounds

(xx) A decrease in NMHC emissions from old tomodern conditions at fixed HES pressure stage and fuelsulphur content (FSC) is found

(xxi) A decrease of NMHC emissions with decreas-ing pressure in the HES at fixed combustor operationconditions and FSC is observed

(xxii) No clear influence on the NMHC emissionswith changing FSC at the same combustor operationcondition is observed

(xxiii) High EIrsquos for organic acids are found Pos-sible explanations are that organic compounds fixed atthe sampling line walls are oxidised by the hot exhaustgases or that the acids are fixed at the surface of thesoot particles which are also trapped in the samplingcartridges

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 475

46 Emission properties

Refer to WILSON et al (2004) and PETZOLD et al(2003a)

(xxiv) Emission properties regarding combustionaerosol mass number and size are determined by thecombustor the influence of the Hot End Simulator isweak

(xxv) The processing of exhaust gas and parti-cles during the transfer through the Hot End Simula-tor increases the particle hygroscopicity and formationstrength of condensation particles both effects are re-lated to gaseous sulphuric acid

The experimental overview of the EU PartEmisproject presents the first comprehensive aerosol and gasmeasurements to be conducted on an aircraft enginecombustor test-rig The database will be used for fur-ther analyses as well as for plume modelling studies ofthe effect of aviation-related particle emissions on theglobal atmosphere and climate

Acknowledgements

The PartEmis project was funded by the European Com-mission under contract no G4RD-CT-2000-00207 Theauthors are very grateful to the test rig operation crew atQinetiQ for their very strong support during the experi-ments

References

ANDERSON BE WR COFER DR BAGWELL JWBARRICK CH HUDGINS KE BRUNKE 1998 Air-borne observations of aircraft aerosol emissions I Totalnonvolatile particle emission indices ndash Geophys Res Lett25 1689ndash1692

BOUCHER O 1999 Influence of air traffic on cirrus occur-rence Nature ndash 397 30ndash31

BROCK CA F SCHROumlDER B KAumlRCHER A PETZOLDR BUSEN M FIEBIG 2000 Ultrafine particle size distri-butions measured in aircraft exhaust plumes ndash J GeophysRes 105 26555ndash26567

CACHIER H MP BREMOND P BIAT-MEacuteNARD 1989Determination of atmospheric soot carbon with a simplethermal method ndash Tellus 41B 379ndash390

DEMOTT PJ Y CHEN SM KREIDENWEIS DCROGERS DE SHERMAN 1999 Ice formation by blackcarbon particles ndash Geophys Res Lett 26 2429ndash2432

GIEBL H A BERNER G REISCHL H PUXBAUM AKASPER-GIEBL R HITZENBERGER 2002 CCN activa-tion of oxalic and malonic acid test aerosols with the Uni-versity of Vienna cloud condensation nuclei counter ndash JAerosol Sci 33 1623ndash1634

GYSEL M E WEINGARTNER U BALTENSPERGER2002 Hygroscopicity of aerosol particles at low tempera-tures 2 Theoretical and experimental hygroscopic proper-ties of laboratory generated aerosols ndash Environ Sci Tech-nol 36 63ndash68

GYSEL M S NYEKI E WEINGARTNER U BAL-TENSPERGER H GIEBL R HITZENBERGER A PET-ZOLD CW WILSON 2003a Properties of jet engine com-bustor particles during the PartEmis experiment Hygro-scopicity at subsaturated conditions ndash Geophys Res Lett30 1566 DOI1010292003GL016896

GYSEL M S NYEKI E WEINGARTNER U BAL-TENSPERGER H GIEBL R HITZENBERGER A PET-ZOLD CW WILSON 2003b Jet engine combustion par-ticle hygroscopicity under subsaturated conditions duringPartEmis ndash J Aerosol Sci 34 (Suppl 2) Abstracts of theEuropean Aerosol Conference S1415ndashS1416

HAGEN DE PD WHITEFIELD H SCHLAGER 1996Particle emissions in the exhaust plume from commercialjet aircraft under cruise conditions ndash J Geophys Res 10119551ndash19557

HAGEN DE PD WHITEFIELD J PALADINO M TRUE-BLOOD H LILENFELD 1998 Particle sizing and emissionindices for a jet engine exhaust sampled at cruise ndash Geo-phys Res Lett 25 1681ndash1684

HAGEN DE PD WHITEFIELD JD PALADINO OSCHMID 2001 Comparative study of jet engine com-bustion emissions and engine operating conditions ndashIn SCHUMANN U G AMANATIDIS (Eds) Aviationaerosols contrails and cirrus clouds Air pollution researchreport 74 106ndash110 European Commission Brussels

HAVERKAMP H S WILHELM A SOROKIN F ARNOLD2004 Positive and negative ion measurements in jet aircraftengine exhaust concentrations sizes and implications foraerosol formation ndash Atmos Environ 38 2879ndash2884

HITZENBERGER R H GIEBL A PETZOLD M GYSELS NYEKI E WEINGARTNER U BALTENSPERGERCW WILSON 2003a Properties of jet engine combus-tor particles during the PartEmis experiment Hygroscopicproperties at supersaturated conditions ndash Geophys ResLett 30 1779 DOI1010292003GL017294

HITZENBERGER R H GIEBL A PETZOILD M GYSELS NYEKI E WEINGARTNER U BALTENSPERGERCW WILSON 2003b CCN activation of jet engine com-bustor particles during PartEmis ndash J Aerosol Sci 34(Suppl 2) Abstracts of the European Aerosol ConferenceS1409ndashS1410

HURLEY CD 1996 AEROTRACE Measurements of Par-ticulates from an Engine Combustor ndash Proceedings Impactof Aircraft Emissions upon the Atmosphere Vol 1 113 In-ternational Colloquium Paris October 1996

KATRAGKOU E S WILHELM F ARNOLD CW WIL-SON 2004 First gaseous Sulfur (VI) measurements in thesimulated internal flow of an aircraft gas turbine engineduring project PartEmis ndash Geophys Res Lett 31 2117DOI1010292003GL018231

NYEKI S M GYSEL E WEINGARTNER U BAL-TENSPERGER R HITZENBERGER A PETZOLD CWWILSON 2003 Volatile properties of jet engine combus-tor particles during PartEmis ndash J Aerosol Sci 34 (Suppl2) Abstracts of the European Aerosol Conference S1411ndashS1412

PETZOLD A FP SCHROumlDER 1998 Jet engine exhaustaerosol characterization ndash Aerosol Sci Technol 28 62ndash76

PETZOLD A A DOumlPELHEUER CA BROCK FPSCHROumlDER 1999 In situ observations and model calcula-

476 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

tions of black carbon emission by aircraft at cruise altitudendash J Geophys Res 104 22171ndash22181

PETZOLD A C STEIN S NYEKI M GYSEL EWEINGARTNER U BALTENSPERGER H GIEBL RHITZENBERGER A DOumlPELHEUER S VRCHOTICKY HPUXBAUM M JOHNSON CD HURLEY R MARSHCW WILSON 2003a Properties of jet engine combus-tor particles during the PartEmis experiment Microphysi-cal and chemical properties ndash Geophys Res Lett 30 1719DOI1010292003GL017283

PETZOLD A CW WILSON U BALTENSPERGERM FIEBIG L FRITZSCHE H GIEBL M GYSELR HITZENBERGER CD HURLEY S NYEKI HPUXBAUM U SCHUMANN C STEIN S VROCHTICKY2003b Particle emissions from aircraft engines ndash anoverview of the European project PartEmis ndash J AerosolSci 34 (Suppl 2) Abstracts of the European Aerosol Con-ference S1405ndashS1406

PETZOLD A AE VRTALA M FIEBIG F FRITZSCHEPH MIRABEL S NYEKI X VANCASSEL CW WIL-SON 2003c Emission of volatile and non-volatile ultra-fine particles from a combustion source during PartEmisndash J Aerosol Sci 34 (Suppl 2) Abstracts of the EuropeanAerosol Conference S1407ndashS1408

PUXBAUM H S VRCHOTICKY A PETZOLD R HITZEN-BERGER S NYEKI CW WILSON 2003 Chemical com-position of jet engine combustor particles during PartEmisndash J Aerosol Sci 34 (Suppl 2) Abstracts of the EuropeanAerosol Conference S1413ndashS1414

SCHMID H L LASKUS H-J ABRAHAM U BAL-TENSPERGER V LAVANCHY M BIZJAK P BURBA HCACHIER D CROW J CHOW T GNAUK A EVENHM TEN BRINK K-P GIESEN R HITZENBERGER CHUEGLIN W MAENHAUT C PIOA CARVALHO J-PPUTAUD D TOOM-SAUNTRY H PUXBAUM 2001 Re-sults of the ldquocarbon conferencerdquo international aerosol car-bon round robin test stage I ndash Atmos Environ 35 2111ndash2121

SCHROumlDER F C BROCK R BAUMANN A PETZOLD RBUSEN P SCHULTE M FIEBIG 2000 In-situ studies onvolatile jet exhaust particle emissions Impact of fuel sul-fur content and thermodynamic conditions on nuclei-modeaerosols ndash J Geophys Res 105 19941ndash19954

SCHUMANN U 1996 Contrail formation from aircraft ex-hausts ndash Meteorol Z NF 5 4ndash23

SCHUMANN U J STROumlM 2001 Aviation impact on atmo-spheric composition and climate ndash In European research inthe stratosphere 1996ndash2000 European Commission EUR19867 Brussels

SCHUMANN U F ARNOLD R BUSEN J CURTIUS BKAumlRCHER A KIENDLER A PETZOLD H SCHLAGERF SCHROumlDER K-H WOHLFROM 2002 Influence of fu-els sulphur on the composition of aircraft exhaust plumesThe experiments SULFUR 1-7 ndash J Geophys Res 107(1010292001JD000813) AAC 2-1 ndash AAC 2-27

SOROKIN A P MIRABEL F ARNOLD 2003 On the In-fluence of fuel fulfur induced stable negative ion formationon the total concentration of ions emitted by an aircraft gasturbine engine Comparison of model and experiment ndash At-mos Chem Phys 3 6001ndash6018

VANCASSEL X A SOROKIN P MIRABEL A PETZOLDCW WILSON 2004 Volatile particles formation duringPartEmis a modelling study ndash Atmos Chem Phys 4 439ndash447

WILHELM S H HAVERKAMP A SOROKIN F ARNOLD2004 Detection of very large ions in aircraft gas turbineengine combustor exhaustcharged small soot particles ndashAtmos Environ 38 4561ndash4569

WILSON CW A PETZOLD S NYEKI U SCHUMANNR ZELLNER 2004 Measurement and Prediction of Emis-sions of Aerosols and Gaseous Precursors from Gas Tur-bine Engines (PartEmis) An Overview ndash Aerosp SciTechnol 8 131ndash143

Page 11: Particle emissions from aircraft engines a survey of the ... · 466 A. Petzold et al.: Particle emissions from aircraft engines Meteorol. Z., 14 , 2005 1 Introduction The role of

Meteorol Z 14 2005 A Petzold et al Particle emissions from aircraft engines 475

46 Emission properties

Refer to WILSON et al (2004) and PETZOLD et al(2003a)

(xxiv) Emission properties regarding combustionaerosol mass number and size are determined by thecombustor the influence of the Hot End Simulator isweak

(xxv) The processing of exhaust gas and parti-cles during the transfer through the Hot End Simula-tor increases the particle hygroscopicity and formationstrength of condensation particles both effects are re-lated to gaseous sulphuric acid

The experimental overview of the EU PartEmisproject presents the first comprehensive aerosol and gasmeasurements to be conducted on an aircraft enginecombustor test-rig The database will be used for fur-ther analyses as well as for plume modelling studies ofthe effect of aviation-related particle emissions on theglobal atmosphere and climate

Acknowledgements

The PartEmis project was funded by the European Com-mission under contract no G4RD-CT-2000-00207 Theauthors are very grateful to the test rig operation crew atQinetiQ for their very strong support during the experi-ments

References

ANDERSON BE WR COFER DR BAGWELL JWBARRICK CH HUDGINS KE BRUNKE 1998 Air-borne observations of aircraft aerosol emissions I Totalnonvolatile particle emission indices ndash Geophys Res Lett25 1689ndash1692

BOUCHER O 1999 Influence of air traffic on cirrus occur-rence Nature ndash 397 30ndash31

BROCK CA F SCHROumlDER B KAumlRCHER A PETZOLDR BUSEN M FIEBIG 2000 Ultrafine particle size distri-butions measured in aircraft exhaust plumes ndash J GeophysRes 105 26555ndash26567

CACHIER H MP BREMOND P BIAT-MEacuteNARD 1989Determination of atmospheric soot carbon with a simplethermal method ndash Tellus 41B 379ndash390

DEMOTT PJ Y CHEN SM KREIDENWEIS DCROGERS DE SHERMAN 1999 Ice formation by blackcarbon particles ndash Geophys Res Lett 26 2429ndash2432

GIEBL H A BERNER G REISCHL H PUXBAUM AKASPER-GIEBL R HITZENBERGER 2002 CCN activa-tion of oxalic and malonic acid test aerosols with the Uni-versity of Vienna cloud condensation nuclei counter ndash JAerosol Sci 33 1623ndash1634

GYSEL M E WEINGARTNER U BALTENSPERGER2002 Hygroscopicity of aerosol particles at low tempera-tures 2 Theoretical and experimental hygroscopic proper-ties of laboratory generated aerosols ndash Environ Sci Tech-nol 36 63ndash68

GYSEL M S NYEKI E WEINGARTNER U BAL-TENSPERGER H GIEBL R HITZENBERGER A PET-ZOLD CW WILSON 2003a Properties of jet engine com-bustor particles during the PartEmis experiment Hygro-scopicity at subsaturated conditions ndash Geophys Res Lett30 1566 DOI1010292003GL016896

GYSEL M S NYEKI E WEINGARTNER U BAL-TENSPERGER H GIEBL R HITZENBERGER A PET-ZOLD CW WILSON 2003b Jet engine combustion par-ticle hygroscopicity under subsaturated conditions duringPartEmis ndash J Aerosol Sci 34 (Suppl 2) Abstracts of theEuropean Aerosol Conference S1415ndashS1416

HAGEN DE PD WHITEFIELD H SCHLAGER 1996Particle emissions in the exhaust plume from commercialjet aircraft under cruise conditions ndash J Geophys Res 10119551ndash19557

HAGEN DE PD WHITEFIELD J PALADINO M TRUE-BLOOD H LILENFELD 1998 Particle sizing and emissionindices for a jet engine exhaust sampled at cruise ndash Geo-phys Res Lett 25 1681ndash1684

HAGEN DE PD WHITEFIELD JD PALADINO OSCHMID 2001 Comparative study of jet engine com-bustion emissions and engine operating conditions ndashIn SCHUMANN U G AMANATIDIS (Eds) Aviationaerosols contrails and cirrus clouds Air pollution researchreport 74 106ndash110 European Commission Brussels

HAVERKAMP H S WILHELM A SOROKIN F ARNOLD2004 Positive and negative ion measurements in jet aircraftengine exhaust concentrations sizes and implications foraerosol formation ndash Atmos Environ 38 2879ndash2884

HITZENBERGER R H GIEBL A PETZOLD M GYSELS NYEKI E WEINGARTNER U BALTENSPERGERCW WILSON 2003a Properties of jet engine combus-tor particles during the PartEmis experiment Hygroscopicproperties at supersaturated conditions ndash Geophys ResLett 30 1779 DOI1010292003GL017294

HITZENBERGER R H GIEBL A PETZOILD M GYSELS NYEKI E WEINGARTNER U BALTENSPERGERCW WILSON 2003b CCN activation of jet engine com-bustor particles during PartEmis ndash J Aerosol Sci 34(Suppl 2) Abstracts of the European Aerosol ConferenceS1409ndashS1410

HURLEY CD 1996 AEROTRACE Measurements of Par-ticulates from an Engine Combustor ndash Proceedings Impactof Aircraft Emissions upon the Atmosphere Vol 1 113 In-ternational Colloquium Paris October 1996

KATRAGKOU E S WILHELM F ARNOLD CW WIL-SON 2004 First gaseous Sulfur (VI) measurements in thesimulated internal flow of an aircraft gas turbine engineduring project PartEmis ndash Geophys Res Lett 31 2117DOI1010292003GL018231

NYEKI S M GYSEL E WEINGARTNER U BAL-TENSPERGER R HITZENBERGER A PETZOLD CWWILSON 2003 Volatile properties of jet engine combus-tor particles during PartEmis ndash J Aerosol Sci 34 (Suppl2) Abstracts of the European Aerosol Conference S1411ndashS1412

PETZOLD A FP SCHROumlDER 1998 Jet engine exhaustaerosol characterization ndash Aerosol Sci Technol 28 62ndash76

PETZOLD A A DOumlPELHEUER CA BROCK FPSCHROumlDER 1999 In situ observations and model calcula-

476 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

tions of black carbon emission by aircraft at cruise altitudendash J Geophys Res 104 22171ndash22181

PETZOLD A C STEIN S NYEKI M GYSEL EWEINGARTNER U BALTENSPERGER H GIEBL RHITZENBERGER A DOumlPELHEUER S VRCHOTICKY HPUXBAUM M JOHNSON CD HURLEY R MARSHCW WILSON 2003a Properties of jet engine combus-tor particles during the PartEmis experiment Microphysi-cal and chemical properties ndash Geophys Res Lett 30 1719DOI1010292003GL017283

PETZOLD A CW WILSON U BALTENSPERGERM FIEBIG L FRITZSCHE H GIEBL M GYSELR HITZENBERGER CD HURLEY S NYEKI HPUXBAUM U SCHUMANN C STEIN S VROCHTICKY2003b Particle emissions from aircraft engines ndash anoverview of the European project PartEmis ndash J AerosolSci 34 (Suppl 2) Abstracts of the European Aerosol Con-ference S1405ndashS1406

PETZOLD A AE VRTALA M FIEBIG F FRITZSCHEPH MIRABEL S NYEKI X VANCASSEL CW WIL-SON 2003c Emission of volatile and non-volatile ultra-fine particles from a combustion source during PartEmisndash J Aerosol Sci 34 (Suppl 2) Abstracts of the EuropeanAerosol Conference S1407ndashS1408

PUXBAUM H S VRCHOTICKY A PETZOLD R HITZEN-BERGER S NYEKI CW WILSON 2003 Chemical com-position of jet engine combustor particles during PartEmisndash J Aerosol Sci 34 (Suppl 2) Abstracts of the EuropeanAerosol Conference S1413ndashS1414

SCHMID H L LASKUS H-J ABRAHAM U BAL-TENSPERGER V LAVANCHY M BIZJAK P BURBA HCACHIER D CROW J CHOW T GNAUK A EVENHM TEN BRINK K-P GIESEN R HITZENBERGER CHUEGLIN W MAENHAUT C PIOA CARVALHO J-PPUTAUD D TOOM-SAUNTRY H PUXBAUM 2001 Re-sults of the ldquocarbon conferencerdquo international aerosol car-bon round robin test stage I ndash Atmos Environ 35 2111ndash2121

SCHROumlDER F C BROCK R BAUMANN A PETZOLD RBUSEN P SCHULTE M FIEBIG 2000 In-situ studies onvolatile jet exhaust particle emissions Impact of fuel sul-fur content and thermodynamic conditions on nuclei-modeaerosols ndash J Geophys Res 105 19941ndash19954

SCHUMANN U 1996 Contrail formation from aircraft ex-hausts ndash Meteorol Z NF 5 4ndash23

SCHUMANN U J STROumlM 2001 Aviation impact on atmo-spheric composition and climate ndash In European research inthe stratosphere 1996ndash2000 European Commission EUR19867 Brussels

SCHUMANN U F ARNOLD R BUSEN J CURTIUS BKAumlRCHER A KIENDLER A PETZOLD H SCHLAGERF SCHROumlDER K-H WOHLFROM 2002 Influence of fu-els sulphur on the composition of aircraft exhaust plumesThe experiments SULFUR 1-7 ndash J Geophys Res 107(1010292001JD000813) AAC 2-1 ndash AAC 2-27

SOROKIN A P MIRABEL F ARNOLD 2003 On the In-fluence of fuel fulfur induced stable negative ion formationon the total concentration of ions emitted by an aircraft gasturbine engine Comparison of model and experiment ndash At-mos Chem Phys 3 6001ndash6018

VANCASSEL X A SOROKIN P MIRABEL A PETZOLDCW WILSON 2004 Volatile particles formation duringPartEmis a modelling study ndash Atmos Chem Phys 4 439ndash447

WILHELM S H HAVERKAMP A SOROKIN F ARNOLD2004 Detection of very large ions in aircraft gas turbineengine combustor exhaustcharged small soot particles ndashAtmos Environ 38 4561ndash4569

WILSON CW A PETZOLD S NYEKI U SCHUMANNR ZELLNER 2004 Measurement and Prediction of Emis-sions of Aerosols and Gaseous Precursors from Gas Tur-bine Engines (PartEmis) An Overview ndash Aerosp SciTechnol 8 131ndash143

Page 12: Particle emissions from aircraft engines a survey of the ... · 466 A. Petzold et al.: Particle emissions from aircraft engines Meteorol. Z., 14 , 2005 1 Introduction The role of

476 A Petzold et al Particle emissions from aircraft engines Meteorol Z 14 2005

tions of black carbon emission by aircraft at cruise altitudendash J Geophys Res 104 22171ndash22181

PETZOLD A C STEIN S NYEKI M GYSEL EWEINGARTNER U BALTENSPERGER H GIEBL RHITZENBERGER A DOumlPELHEUER S VRCHOTICKY HPUXBAUM M JOHNSON CD HURLEY R MARSHCW WILSON 2003a Properties of jet engine combus-tor particles during the PartEmis experiment Microphysi-cal and chemical properties ndash Geophys Res Lett 30 1719DOI1010292003GL017283

PETZOLD A CW WILSON U BALTENSPERGERM FIEBIG L FRITZSCHE H GIEBL M GYSELR HITZENBERGER CD HURLEY S NYEKI HPUXBAUM U SCHUMANN C STEIN S VROCHTICKY2003b Particle emissions from aircraft engines ndash anoverview of the European project PartEmis ndash J AerosolSci 34 (Suppl 2) Abstracts of the European Aerosol Con-ference S1405ndashS1406

PETZOLD A AE VRTALA M FIEBIG F FRITZSCHEPH MIRABEL S NYEKI X VANCASSEL CW WIL-SON 2003c Emission of volatile and non-volatile ultra-fine particles from a combustion source during PartEmisndash J Aerosol Sci 34 (Suppl 2) Abstracts of the EuropeanAerosol Conference S1407ndashS1408

PUXBAUM H S VRCHOTICKY A PETZOLD R HITZEN-BERGER S NYEKI CW WILSON 2003 Chemical com-position of jet engine combustor particles during PartEmisndash J Aerosol Sci 34 (Suppl 2) Abstracts of the EuropeanAerosol Conference S1413ndashS1414

SCHMID H L LASKUS H-J ABRAHAM U BAL-TENSPERGER V LAVANCHY M BIZJAK P BURBA HCACHIER D CROW J CHOW T GNAUK A EVENHM TEN BRINK K-P GIESEN R HITZENBERGER CHUEGLIN W MAENHAUT C PIOA CARVALHO J-PPUTAUD D TOOM-SAUNTRY H PUXBAUM 2001 Re-sults of the ldquocarbon conferencerdquo international aerosol car-bon round robin test stage I ndash Atmos Environ 35 2111ndash2121

SCHROumlDER F C BROCK R BAUMANN A PETZOLD RBUSEN P SCHULTE M FIEBIG 2000 In-situ studies onvolatile jet exhaust particle emissions Impact of fuel sul-fur content and thermodynamic conditions on nuclei-modeaerosols ndash J Geophys Res 105 19941ndash19954

SCHUMANN U 1996 Contrail formation from aircraft ex-hausts ndash Meteorol Z NF 5 4ndash23

SCHUMANN U J STROumlM 2001 Aviation impact on atmo-spheric composition and climate ndash In European research inthe stratosphere 1996ndash2000 European Commission EUR19867 Brussels

SCHUMANN U F ARNOLD R BUSEN J CURTIUS BKAumlRCHER A KIENDLER A PETZOLD H SCHLAGERF SCHROumlDER K-H WOHLFROM 2002 Influence of fu-els sulphur on the composition of aircraft exhaust plumesThe experiments SULFUR 1-7 ndash J Geophys Res 107(1010292001JD000813) AAC 2-1 ndash AAC 2-27

SOROKIN A P MIRABEL F ARNOLD 2003 On the In-fluence of fuel fulfur induced stable negative ion formationon the total concentration of ions emitted by an aircraft gasturbine engine Comparison of model and experiment ndash At-mos Chem Phys 3 6001ndash6018

VANCASSEL X A SOROKIN P MIRABEL A PETZOLDCW WILSON 2004 Volatile particles formation duringPartEmis a modelling study ndash Atmos Chem Phys 4 439ndash447

WILHELM S H HAVERKAMP A SOROKIN F ARNOLD2004 Detection of very large ions in aircraft gas turbineengine combustor exhaustcharged small soot particles ndashAtmos Environ 38 4561ndash4569

WILSON CW A PETZOLD S NYEKI U SCHUMANNR ZELLNER 2004 Measurement and Prediction of Emis-sions of Aerosols and Gaseous Precursors from Gas Tur-bine Engines (PartEmis) An Overview ndash Aerosp SciTechnol 8 131ndash143