,~ Aerosol &t Vot. 31, Suppl. I, pp. S170-S171, 2000

Pergamon

www.elsevier.com/locate/j aerosci

Session 4A -Atmospheric aerosols: Particle formation and growth

NEW PARTICLE FORMATION IN AIRCRAFT EXHAUST PLUMES

B. K,~RCHER 1, R.E TURCO 2, E YU 2, M.Y. DANILIN 3, D.K. WEISENSTEIN 3, R.C. MIAKE-LYE 4, and R. BUSEN 1

1 DLR Institut for Physik der Atmosph~ire, D-82234 Wessling, Germany.

2 UCLA Department of Atmospheric Sciences, Los Angeles, CA 90095, U.S.A.

3 Atmospherl.'C and Environmental Research, Inc., Cambridge, MA 02139, U.S.A.

4 Aerodyne Research, Inc.,Billerica, MA 01821, U.S.A.

Keywords: aerosol nucleation and molecular clusters, ions, field campaigns, aerosol chemical composition, aerosol dynamics and transport, global effects.

INTRODUCTION

In addition to emitting nonvolatile soot particles formed within the engine combustors, jet aircraft generate copious volatile particles within their near-field exhaust plumes: Models and observations show that the total background aerosol number and surface area densities can be significantly enhanced by aircraft emissions. The resulting tendencies of aircraft sulphate and soot aerosols to influence cirrus clouds and to alter chemical processes are as yet poorly quantified, limiting our ability to assess climate and ozone impacts. Indeed, in- situ observations of particles in fresh aircraft plumes show a considerable scatter in the apparent volatile particle emission index, even for a fixed fuel sulphur (S) emission index El(S). The factors that contribute to this variability have not been indentified.

RESULTS AND DISCUSSION

We have developed an analytical model to explain observed ultrafine particle emissions that consolidates several new concepts concerning the physico-chemical mechanisms responsible for the nucleation and growth of aviation plume aerosols (Kfircher et al., 2000). Following Yu et al. (1999), we assume that par- ticulate organic matter (POM) and sulphuric acid (H2SO4) initially condense with associated water vapour (H20) to form both ionised and electrically neutral aerosol modes. The effects of electrical charge on the particle collection kernels lead to a preferential growth of an ion aerosol mode, whereas the neutral mode particles maintain smaller cluster sizes and are not detectable by condensation nuclei counters (CNCs). Scavenging of the neutral clusters by ion mode particles constitutes the major growth mechanism of the latter during the first minutes of plume evolution.

Our model yields the temporal evolution of the size distribution of the detectable volatile particles. To facilitate comparison with observations, we discuss the apparent particle emission indices (AEIs) as the cumulative number of particles per kg of burned fuel that are larger than a given size. The calculated AEIs account for estimated detection efficiencies of the particle counters used in the field. The model resfilts demonstrate that the use of CNCs with relatively large cut-off diameters (Dth ~5 nm) precludes the detection of all of the new aerosol particles in very young plumes (plume ages t below 10 s), especially foi" low EI(S) levels. The AEIs at low El(S) levels are determined by the condensation of organics, in which case the AEIs are very sensitive to the prescribed EI(POM) value. The number of detectable volatiles depends strongly on plume age, because they grow across the detection threshold Dth of the ultrafine CNCs over the course of time. Finally, the model also reflects the generally observed increase in AEIs with rising El(S).

S170

Abstracts of the 2000 European Aerosol Conference SI71

The complex evolution of the particulates can be forecast using only three parameters associated with aircraft

engines: the amount of emitted chemi-ions, condensable organic species, and fully oxidised sulphur gases at emission. Among these parameters, EI(POM) is not well characterised observationally, while the conversion

fraction q of fuel S to H2SO4 exhibits a range of 0.5-15% and the chemi-ion emission index Ni varies between (1 -4)× 101Z/kg-fuel. Much of the scatter in AEIs noted in field measurements can be explained consistently by variations of t,, EI(POM), and Dth. The sensitivity of AEIs to the plume dilution rate and environmental conditions is much less pronounced. Observed AEIs from recent field missions are plotted in the figure (panel a) and are compared to the more unified data set (panel b). The latter was obtained by normalising the original data with our model using comon values for t (3 s), EI(POM) (20 mg/kg-fuel), and D~h (5 nm). Comparing panels a and b reveals the common trends that underly the various measurements made behind different aircraft and under quite different experimental conditions.

10 ~ 1 0

t=3s b

D th=5nm T

El(POM)=2Omg/kg fuel , l j . . . ~ - - - - ~ / . ~ S : . ' "

-- ~ t~ 1 * r Y l =5 Y. ~1 / _A / t

4 , 0 ' . • q=0.5%

' v a r y E I (POM)

10 210_ 3 10 210 , ' . . . . '1'0' 2 . . . . . '1'0 '-~ ' . . . . . 1"0 ° ' " ' 1 0 '

S em iss ion index, g / k g fue l

Figure 1 : AEIs of detectable volatile particles versus EI(S). Panel a: Data from several European and US missions plotted irrespective of t, Dth, and EI(POM). Legend notes aircraft types and values for Dth. Panel b: Same data, but normalized to t = 3 s Dth = 5 rim, and 20mg POM/kg-fuel using our new analytical model. The curves emphasizing the trend of observed AEIs with increasing EI(S), and bracketing most of the data, are model results at the same values for t, Dth, and EI(POM), with N~ = 2 x 1017/kg-fuel and 71 = 5% (solid curve) or 1 / = 0.5c~ (dashed curve)

• 3;5nrn ATFAS 7 " a • 4 Ohm s737 | • • 4 Onto: DC-8 V90nm Concorde • 12 nm 8747 • • • Ib12 nm DC-10 A •

• m40nm: F-16 •

10-; 10 ̀ ~ 100 10 ~

S em iss ion index, g / k g fue l

SUMMARY

We have developed a unique representation of the properties of volatile particles produced by jet aircraft. Our new parameterisation is based upon detailed microphysical mechanisms that incorporate the latest phe- nomenology observed in aircraft plumes, including the roles of chemi-ions and organic vapours. By com- paring particle measurements taken under a wide range of flight conditions, and normalising these using our model, we have reduced dramatically the large residual variability in the volatile particle emission indices inferred from field observations. By coupling our model to a global tracer model, we have been able to nar- row considerably the range of predicted global enhancements in background aerosol surface area densities associated with commercial aircraft operations (not discussed here). Our novel approach should prove to be useful in future evaluations of the atmospheric perturbations associated with aviation.

ACKNOWLEDGEMENTS

We acknowledge support by a NATO Collaborative Research Grant and by grants from NASA and NSF.

REFERENCES

Kfircher, B., R.E Turco, E Yu, M.Y. Danilin, D.K. Weisenstein, R.C. Miake-Lye, and R. Busen (2000). A quantitative understanding of new particle formation in aircraft exhaust plumes. Submitted manuscript.

Yu, E, R.E Turco, and B. Kiircher (1999). The possible role of organics in the formation and evolution of ultrafine aircraft particles. J. Geophys. Res., 104, 4079.