Physical Properties, Chemical Composition, and Cloud Forming

The potential of particles to form cloud droplets (cloud condensation nuclei, ...... Spackman , Steven C. Wofsy , Doug Worsnop , Bin Xiang , and Eric ...
0 downloads 0 Views 855KB Size
Environ. Sci. Technol. 2010, 44, 3800–3805

Physical Properties, Chemical Composition, and Cloud Forming Potential of Particulate Emissions from a Marine Diesel Engine at Various Load Conditions A . P E T Z O L D , * ,† E . W E I N G A R T N E R , ‡ J. HASSELBACH,† P. LAUER,§ C. KUROK,| AND F. FLEISCHER§ Deutsches Zentrum fu ¨ r Luft- und Raumfahrt, Institut fu ¨r Physik der Atmospha¨re, Oberpfaffenhofen, 82234 Wessling, Germany, Laboratory of Atmospheric Chemistry, Paul Scherrer Institute, CH-5232 Villigen PSI, Switzerland, MAN Diesel SE, Stadtbachstrasse 1, 86135 Augsburg, Germany, and Germanischer Lloyd AG, Vorsetzen 35, 20459 Hamburg, Germany

Received December 4, 2009. Revised manuscript received March 22, 2010. Accepted April 4, 2010.

Particulate matter (PM) emissions from one serial 4-stroke medium-speed marine diesel engine were measured for load conditions from 10% to 110% in test rig studies using heavy fuel oil (HFO). Testing the engine across its entire load range permitted the scaling of exhaust PM properties with load. Emission factors for particle number, particle mass, and chemical compounds were determined. The potential of particles to form cloud droplets (cloud condensation nuclei, CCN) was calculated from chemical composition and particle size. Number emission factors are (3.43 ( 1.26) × 1016 (kg fuel)-1 at 85-110% load and (1.06 ( 0.10) × 1016 (kg fuel)-1 at 10% load. CCN emission factors of 1-6 × 1014 (kg fuel)-1 are at the lower bound of data reported in the literature. From combined thermal and optical methods, black carbon (BC) emission factors of 40-60 mg/(kg fuel) were determined for 85-100% loadand370mg/(kgfuel)for10%load.Theengineloaddependence of the conversion efficiency for fuel sulfur into sulfate of (1.08 ( 0.15)% at engine idle to (3.85 ( 0.41)% at cruise may serve as input to global emission calculations for various load conditions.

1. Introduction The continuing growth in global shipping gives rise to increasing concern about potential effects on global climate (1-3) and on air quality in coastal areas (4, 5). Currently, about 10% of global transportation CO2 emissions can be attributed to shipping with the number still increasing (6). By far the largest nongreenhouse gas contribution of shipping to global climate forcing is associated with the direct and indirect effects of sulfate particles forming from emitted SO2 (1, 2). In coastal Southern California shipping contributes * Corresponding author e-mail: [email protected]. † DLR Institut fu ¨ r Physik der Atmospha¨re. ‡ Paul Scherrer Institute. § MAN Diesel SE. | Germanischer Lloyd AG. 3800

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 10, 2010

10-44% of PM non-sea salt sulfate (7) and up to 9% of total PM2.5 (5). Direct aerosol forcing of sulfate particles from shipping ranges between -47 mW/m2 and -12 mW/m2 (1). Light-absorbing BC contributes only (2.0 ( 0.9) mW/m2 (3) which is negligible compared to the sulfate-associated effects. Radiative forcing from indirect effects range from -39 mW/m2 to -660 mW/m2 (1), indicating that the overall radiative forcing from shipping today is negative due to the strong cooling caused by the aerosol-cloud effects. In past years, several studies on particles emissions from shipping and their effects on marine stratus clouds were conducted. Approaches include test rig measurements or exhaust stack sampling (8-14), airborne and ship-borne measurements in ship plumes (15-18), combinations of exhaust studies and plume studies using airborne platforms (19, 20), and finally the investigation of the CCN fraction of particles which potentially form cloud droplets (16, 19-21). Our study focuses on the impact of load conditions of a marine diesel engine on physical and chemical properties including the CCN potential of PM emitted from shipping.

2. Experimental Methods 2.1. Test Rig Studies. In the EU Project HERCULES (High Efficiency R&D on Combustion with Ultra Low Emissions for Ships), particle emissions from two four-stroke mediumspeed marine diesel engines were studied. The engines under investigation were one multicylinder serial engine with 10 MW power which is one of the largest four-stroke diesel engines produced by the manufacturer and one singlecylinder test engine with 400 kW power. The serial engine is a common type operated on passenger vessels as QE 2, carriers, and feeder vessels. The bulk of results were obtained from tests with the serial engine operating at 100%, 110%, 85%, 50%, 25%, 10%, and again 100% load, with each condition lasting for 1 h. Heavy fuel oil (HFO) with a fuel sulfur content (FSC) of 2.21 weightpercent (wt %) was used. In a prestudy, the single-cylinder test engine was operated at extreme conditions of 10% and 100% load using HFO with an FSC of 2.32 wt %. The FSC of both fuels was close to the current global average value of 2.4 wt % (1), with an ash content of 0.03 wt %. One steadystate test was performed running the test engine from cold conditions for 5 h and monitoring the evolution of emissions. 2.2. Instrumentation. The instrumentation used for the measurement of aerosol microphysical properties consisted of Condensation Particle Counters (CPC, TSI 3010/3760A) with lower cutoff diameters of 0.01 µm, 0.03 µm, and 0.08 µm achieved by diffusion screen separators, one Differential Mobility Analyzer (DMA, TSI 3071), and a Multi-Angle Absorption Photometer MAAP (Thermo Instruments Model 5012) for BC measurement (BC measurement uncertainty 12% (22)). Separation of volatile and nonvolatile aerosol compounds was achieved by a thermodenuder operated at 350 °C. The samples for online aerosol microphysics measurements were diluted by a factor of 1000 for the DMA and the MAAP and by an additional factor of 100 for the CPC, using multiple isokinetic dilution stages (Model VKL-10; Palas, Germany). The measurement setup for the testing of the serial engine presented in ref 19 where further details are given. Filter stack samples were taken with an AVL 472 Smart Sampler modular dilution system (AVL, Graz Austria). Teflon filters were analyzed by gravimetry for total mass at the laboratories of Germanischer Lloyd (GL) and MAN Diesel SE (MBW). Preconditioned quartz fiber filters were analyzed for carbon species by multistep combustion and for sulfate by 10.1021/es903681z

 2010 American Chemical Society

Published on Web 04/19/2010

TABLE 1. Emission Factors of Aerosol Number and Mass Fractions Emitted from a Serial MAN B&W Four-Stroke Marine Diesel Engine Operating on HFO with 2.21 wt % Sulfur and 0.03 wt % Ash total engine, load serial engine, 100%a serial engine, 110% serial engine, 85% serial engine, 50% serial engine, 25% serial engine, 10% serial engine, 100%b serial engine, 85-110% test bed, stack test engine, 100% 2-stroke, cruise, 70%c (12) 2-stroke, cruse, 57%c (20) 2-stroke cruise, 84% (14) airborne plume slow-speed, 40-100%d (18) med.-speed, 30-70%d (18) high-speed, 40-100%d (18) 2-stroke cruise, 57% (20) 2-stroke, cruise (16) 2-stroke cruise, 85% (19) 2-stroke, at dock (17)

number emission factor nonvolatile N0.1-3 µm

1016 (kg fuel)-1 3.85 ( 0.30 1.38 ( 0.21 4.44 ( 0.11 1.39 ( 0.13 0.62 × 10-3 3.85 ( 0.17 1.37 ( 0.11 0.44 × 10-3 2.33 ( 0.18 1.28 ( 0.23 0.56 × 10-3 2.12 ( 0.09 1.33 ( 0.25 1.55 × 10-3 1.06 ( 0.10 0.70 ( 0.01 7.97 × 10-3 1.59 ( 0.05 0.91 ( 0.23 1.04 × 10-3 3.43 ( 1.26 1.26 ( 0.23 0.7 × 10-3

EC

BC

349 197 180 150 341 515 160 179 ( 18

75 70 57 72 204 367 58 62 ( 7

mass emission factor OM SO42mg (kg fuel)-1 1395 2403 1205 2146 1600 3156 1117 1689 1400 1567 1932 819 1555 2842 1450 ( 215 2710 ( 515

111 80 34 130 2.71 ( 1.08 1.08 ( 0.68 1.30 ( 0.20 4.0-6.2 1.36 ( 0.24

0.06 ( 0.01 0.11-0.51 0.23 ( 0.07

0.88 ( 0.10

410 ( 270 970 ( 660 360 ( 230 40 ( 10 180 ( 20 174 ( 43 520 ( 280

1470

3218

1580

760

1570 ( 830 650 ( 440 750 ( 220 1180 ( 40

1550 ( 1100 790 ( 700 530 ( 460 4300 ( 500

PM 6349 5544 7740 4586 4842 4202 7111 6800 ( 1130

8300

3530 ( 2200 2410 ( 1800 1640 ( 910 12900 ( 100

a Load condition measured at cold start. b Load condition measured at the end of the test for warm-up engine. are calculated from [g/kWh] using a value of 3.107 kg CO2/(kg fuel). d PM ) OM + EC + SO4.

ion chromatography (GL internal method validation: 5% measurement uncertainty). Multistep combustion ((23) and method VDI-2 in ref 24) analyzes organic carbon (OC) by sequentially heating the sample under helium to 350 °C, 620 °C, and 400 °C. Subsequently, elemental carbon (EC) is analyzed by switching to oxygen and heating the sample to 700 °C (EC, OC measurement uncertainty 20% (24)). Prior to thermal analysis, the extractable organic carbon fraction was removed in a solvent mixture of 50:50% toluene and isopropyl alcohol. As QA/QC check for mass conservation gravimetric data were compared to the sum of chemical components EC + OM + ash + hydrated sulfate. GL internal procedure based on heavy duty diesel engine data as well as data from ref 25 suggest OM = 1.2 × OC. Again from GL procedure hydrated sulfate was calculated as (SO4 · 4.5H2O). As a second QA/QC approach OM ) OC and hydrated sulfate ) (SO4 · 6.5 H2O) were used (12) and applied to our samples as well. 2.3. Potential Cloud Condensation Nuclei Activation. The theoretical hygroscopic growth factors at high relative humidity (RH) were calculated for an internally mixed aerosol from the measured chemical composition of the bulk aerosol, using the Zdanovskii-Stokes-Robinson (ZSR) relation which is reviewed by Gysel et al. (26) GFmixed(aw) )

(∑ i

)

εi · GFi(aw)3

1/ 3

with GF(aw)i )

(

1 + ki

aw 1 - aw

)

1/

3

(1)

GFmixed is the hygroscopic diameter growth factor of the mixed particles defined as GF ) D(RH)/D0 where D0 is the particle dry diameter, and D(RH) is its diameter at a specific RH. The factor ki depends on the water-soluble substance and is obtained by eq 1 from experimental growth factor data for pure substances (27). GFmixed is calculated from the growth factors of the individually measured chemical compounds and their corresponding volume fractions εi. Ash and EC were considered hydrophobic with ki ) 0.0 and GFi ) 1.00. For OM we used kOM ) 0.0809 and GFOM ) 1.36 which corresponds to GFOM ) 1.2 at aw ) 0.90 for the organic carbon

c

Values

fraction of aged polluted aerosols (26). For H2SO4 we used the values kH2SO4 ) 0.75106 and GFH2SO4 ) 2.48 from a semiempirical model prediction (ADDEM) (28, 29). Properties of mixed particles were calculated from the pure substances according to their volume fractions. This approximation ignores the effect of solute-solute interactions, which is expected to be fulfilled in sufficiently diluted solutions. The estimated hygroscopic growth factor GFmixed at aw ) 0.95 was used to calculate the size dependent critical supersaturation (Scrit ) RHmax - 1) by using modified Ko¨hler theory (30). Supersaturations of 0.1% and 0.3% were used as reference values for the calculation of activation diameters Dcrit. These values reflect the range of conditions in lowlevel, warm stratiform marine clouds (31). Additionally, a supersaturation of 0.3% is used in CCN counters (16), i.e., values for 0.3% supersaturation can be compared directly to data from literature. Assuming a homogeneously internally mixed aerosol, CCN number concentrations were calculated from NCCN )





Dcrit

dN d log D d log D

(2)

where Dcrit is the critical dry diameter which corresponds to the respective critical supersaturation. The method of calculating activation diameters from measured size distributions was evaluated against Koehler theory in emission studies on sulfur-coated combustion particles and provided agreement within 10% for saturation ratios of 0.6% and respective small Dcrit values (32). The calculation of emission factors from number concentrations is described in detail in ref 19.

3. Results 3.1. Bulk Particle Properties. The series of performed tests delivered an extensive set of physicochemical properties of particles emitted from a medium-speed marine diesel engine at load conditions varying from 10% to 110%. Table 1 summarizes the average physical and chemical properties of emitted PM which are reported as emission factors, i.e., in terms of quantities emitted per kg of burned fuel. Emission factors are converted into quantities emitted per generated VOL. 44, NO. 10, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3801

FIGURE 1. Particulate mass determined from the sum of chemical components (Σ) and from gravimetry (Γ) related to particle volume calculated from DMA particle size distributions with the assumption of particle spherical shape; data are shown for the single-cylinder test engine and for the serial engine. The dashed line represents the relationship between volume and mass expected for an effective particle density of 1385 kg m-3 which corresponds to the average chemical composition of the particles. power by multiplication with specific fuel oil consumptions of 0.212 kg/kWh for average load conditions of 85-110% and 0.531 kg/kWh for 10% load. Trends in emissions are different for the considered compounds. PM and sulfate emissions increase with load, while OM and EC decrease with load. Emission factors for PM at 10% load are 4.20 g/(kg fuel) for the serial engine and (4.59-4.74) g/(kg fuel) for the test engine. Respective values for high load are (6.80 ( 1.13) g/(kg fuel) (serial: 85-110%) and 5.86 g/(kg fuel) (test: 100%). Emission factors for PM agree for both engines, while OM emissions are lower for the test engine. Total PM mass was measured by gravimetry and calculated from the sum of EC, OM, ash, and hydrated sulfate. Assuming densities of 1500, 1400, 2600, 1840, and 1000 kg/m3 for EC, OM, ash, H2SO4, and H2O, respectively, the PM chemical composition of hydrated particles yields an effective particle density of (1385 ( 7) kg/m3. Applying this value, particle size distributions were converted into mass concentrations, assuming particle sphericity. PM mass obtained from size distributions, gravimetry, and chemical analyses differ by (5% maximum; see Figure 1. PM mass calculated from OM ) OC and hydrated sulfate ) (SO4 · 6.5H2O) (12) differs from our approach by +10% and thus falls at the upper limit of uncertainty spanned by the different approaches for PM mass determination. The QA/QC check on PM mass proves good data quality. However, it does not allow distinguishing between the different approaches for considering PM contributions from hydrated sulfate and OM instead of OC. Emission factors for light-absorbing BC and EC are frequently used synonymously although the properties describe different fractions of carbonaceous matter. Comparing EC determined by multistep combustion and BC determined by light absorption we found opposite trends for OC/TC and BC/EC, respectively, with load; see Figure 2. We have added information from the AVL 415 filter smoke number (FSN), which is an optical filter reflectance method measuring BC but sampled from another line as the MAAP. Both FSN and BC are highly correlated with r2 ) 0.984. EC values exceed BC values by a factor of 3 for load conditions g85% and by a factor of 1.4 for a load condition of 10%. It is known from round-robin tests (24) that for samples with a higher OC/TC ratios pyrolytic conversion of OC to EC during thermal analysis may cause erroneously high EC values. Based on our analysis, EC and BC emission 3802

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 10, 2010

FIGURE 2. Mass emission factor for carbon-containing compounds in the raw exhaust gas; black carbon (BC) and filter smoke number (FSN) refer to optical methods for the measurement of light-absorbing carbon, while elemental carbon (EC) is determined by a thermal method. The ratio of organic to total carbon is a measure for the importance of pyrolytic conversion of OC to EC by the thermal method.

FIGURE 3. Test bed measurements of aerosol size distributions by a differential mobility analyzer. Size distributions refer to the indicated load conditions of a serial MAN marine diesel engine operating on heavy fuel oil. factors vary by a factor of up to 3, depending on the applied measurement method. During the steady-state run of the single-cylinder test engine it required 2 h of operation until the engine was at its steady-state emission level. From engine cold-start to steady-state conditions emissions of nonvolatile PM by number decreased from 6.5 × 1016 (kg fuel)-1 to 1.8 × 1016 (kg fuel)-1, while EC emissions decreased from 390 g/(kg fuel) to 111 g/(kg fuel). 3.2. Particle Size. Particle size distributions are shown in Figure 3. Total aerosol is characterized by a nucleation mode centered at Dp ) 0.015 µm, a combustion aerosol mode centered at 0.04-0.06 µm, and a third but weak mode at 0.15 µm. Virtually no particles are observed with Dp > 0.25 µm. The emission factor for particles with Dp > 0.1 µm is 4 orders of magnitude lower than the total aerosol emission factor but increases with decreasing load, which corresponds to the simultaneous increase in BC emissions. The combustion particle mode centered at Dp ) 0.04-0.06 µm agrees well with test rig data from four-stroke engines (8) as well as from two-stroke engines (11). Airborne measurements in ship plumes report consistently larger modal diameters of 0.06 to 0.1 µm (19-21, 33), whereas no nucleation mode at Dp e 0.02 µm is observed in the ship plumes. One likely explanation for this systematic deviation is the aging of particles during plume evolution after leaving the stack (15, 19, 34).

FIGURE 4. Efficiency for converting fuel sulfur to particulate matter sulfate at various engine loads and for fuels with different sulfur contents given in wt %; the dashed line represents a linear relationship between ε and engine load. 3.3. Particle Number Emission. Number emission factors are (3.43 ( 1.26) × 1016 (kg fuel)-1 at 85-110% load and (1.06 ( 0.10) × 1016 (kg fuel)-1 at 10%; see Table 1. The thermodenuder removes 2/3 of the particles at high load and 1/3 at low load. The observation that light-absorbing BC emissions remain unaffected by engine load for load conditions >40% agrees well with the result that the number emission factor for combustion particles containing a nonvolatile core of EC and ash remains almost constant for load conditions g20%. The increase in total particle emissions with load by a factor of 3 is almost entirely attributed to sulfuric acid-water droplets. Observed total particle number emission factors are larger by a factor of 3 compared to recent data from plume studies (18-20). Interestingly, the emission factor for nonvolatile combustion particles from this emission study agrees well with the emission factors for total particles determined in these plume studies which supports the assumption that nonvolatile PM from emission dominates PM in the plume, while the volatile sulfuric-acid water droplets vanish quickly by coagulation in the early plume phase (15, 19). 3.4. Sulfur Conversion Efficiency. The increase in sulfate emissions with engine load (Table 1) by a factor of 3 coincides with the increased emission of total particles by a similar factor of 3. Furthermore, the build-up of an externally mixed volatile particle mode centered at 0.015 µm in diameter is observed (Figure 3). The production of PM sulfate from fuel sulfur is characterized by the conversion efficiency ε which is calculated from measured sulfate emissions and the FSC. As for the emission of sulfuric acid-water droplets, a significant dependence of ε on engine load is found (Figure 4). Average values calculated from the combined data sets of test and serial engines are (1.08 ( 0.15)% at 10% load and (3.85 ( 0.41)% at 85-110% load. Literature data for HFO with FSC of 2.05 wt % (12) and 2.85 wt % (13) indicate conversion efficiencies of 1.4-1.9% at 25% load to (3.9-5.0)% at 75% to 85% load. Data from a 40 MW 2-stroke marine diesel engine report conversion efficiencies of 1.8% at 20% load and 3.4% at 80% load (C. Kurok, unpublished data). The HFO data set shown in Figure 4 can be represented by the relation ε ) 1.0 + 0.035 × [engine load in %]. Results for MGO with a FSC of 0.155 wt % (11) indicate a lower conversion of 1.4% for 100% load. However, available data

do not allow the quantification of the impact of FSC on the conversion efficiency. 3.5. Cloud Forming Potential. The emission of potential CCN is calculated from the chemical composition of dry particles at RH ) 0%. Table 2 summarizes volume fractions of major chemical compounds. At high load (85-110%), the average fractional composition in % of dry particle volume is (6.0 ( 2.5)% EC, (37.3 ( 0.5)% OM, (4.0 ( 2.0)% ash, and (52.7 ( 3.2)% sulfate. Corresponding values at 10% load are 15.0% EC, 60.4% OM, 4.7% ash, and 19.9% sulfate. Focusing on the sulfate volume fraction for average cruising conditions, our value of (52.7 ( 3.2)% for a four-stroke engine is of similar magnitude as 61.6% for 2-stroke test rig (12) and 50.3% for slow speed diesel engines (18). Values for medium speed (18) and high speed (18) as well as for a cruising 2-stroke slow speed diesel engine (20) differ significantly. Growth factors were calculated for the respective chemical compositions using eq 1. Results are listed in Table 2. Calculated growth factors at aw ) 0.95 factors vary from 1.86 to 2.28 depending on the sulfate fraction. For sulfate fractions of 50-60% corresponding to our results GF values vary from 2.06 to 2.16. Using GF ) 1.0 for OM as recommended in ref 20 GF factors are reduced by 2%. An intercomparison of calculated and measured GF is available only from one study (20) where the authors report an overprediction of GF by the volume mixing rule against measurements values of about 15%. The calculated critical supersaturation Scrit as a function of dry particle size, assuming a surface tension of pure water at T ) 293 K, was used to determine critical particle diameters for CCN activation for all engine load conditions. The results are compiled in Table 3 for selected supersaturations of 0.1% and 0.3%. Supersaturations required for the CCN activation of a dry particle of Dp ) 0.1 µm are 0.1% for sea salt particles, 0.19% for particles emitted from a marine diesel engine operating on HFO at high load, 0.95% for combustion particles coated with 3.0 vol-% of sulfate (32), and 1.33% for almost sulfur-free diesel particles (35). For constant Scrit, ship exhaust particles are characterized by 1.5-2 times larger critical diameters compared to inorganic salts. If sulfur-free fuel is burned, emitted particles require supersaturations larger than 1% which do not occur in marine stratiform clouds. It has to be noted that the derived Scrit values represent an upper limit because we assumed surface tension of pure water in the Ko¨hler model extrapolation. The Scrit values can be significantly reduced if surface active organic compounds are present on the surface of the solution droplets. In addition, it is expected that aging processes in the plume will further increase the hygroscopicity of the freshly emitted soot particles. 3.6. CCN Number Emission Index. CCN number concentrations and respective emission factors are estimated for supersaturation ratios of 0.1% and 0.3% from size distributions by eq 2, using activation diameters from Table 3. An analysis (19, 21) of processed particle size distributions in the MBL yields that particles larger than 0.1 µm are potential CCN. The results for the emission factors for potential CCN are added to Table 3. Calculations of potential CCN at 0.3% supersaturation from fresh exhaust yield emission factors being smaller by a factor of 2-10 compared to literature data (16, 20), though all particles with Dp > 0.1 µm will be activated. Potential reasons for this difference are ongoing chemical transformation and particle growth processes in the expanding ship plume, and the fact that for plume studies the fraction of particles larger than 0.1 µm is 100 times larger than for the fresh exhaust. Indicators for particle growth from fresh exhaust to aged plume conditions are the decreasing number concentration with plume age due to particle coagulation and the observation that the combustion particle mode in the plume is centered at VOL. 44, NO. 10, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3803

TABLE 2. Chemical Composition of Dry Particles in Volume Fractions of Individual Compounds for Particle Emissions from the Serial Engine and for Data from the Literature (12, 18, 20)a GF at aw ) 0.95

dry volume fraction, % source

4-stroke, this study

2-stroke, test rig (12) slow speed (18) medium speed (18) high speed (18) 2 stroke slow speed (20) particles in plume (20) a

load

EC

OM

ash

SO42-

GFOM ) 1.36

GFOM ) 1.00

100% 110% 85% 50% 25% 10% 100% 70% cruise cruise cruise cruise plume

8.7 5.7 3.8 5.1 10.3 15.0 3.7 1.3 10.9 38.1 21.1 0.5 N/A

37.3 37.6 36.6 41.1 45.4 60.4 38.3 21.4 38.8 23.8 40.9 13.2 28-31

4.0 4.7 3.5 5.5 4.9 4.7 3.7 15.7 N/A N/A N/A 10.8 N/A

49.9 52.0 56.1 48.2 39.4 19.9 54.3 61.6 50.3 38.1 38.0 75.5 67-70

2.06 2.08 2.12 2.08 2.04 1.94 1.68 2.16 2.06 1.90 1.92 2.29 2.23

2.01 2.03 2.08 2.03 1.99 1.88 1.57 2.14 2.02 1.86 1.86 2.28 2.20

GF factors are calculated from eq 1.

TABLE 3. Key Properties of Potential CCN Emitted from the Marine Diesel Enginec activation diameter CCN Scrit ) engine load 110% 100%a 85% 50% 25% 10% cruise (16) cruise (18) cruise (20)

number concentration

Dcrit 0.1%

0.3%

0.1%

EFCCN 0.3%

N>5 nm 6

nm 152 154 149 155 165 195

emission factor

NCCN

73 72 72 75 79 94

0.027 0.025 0.022 0.029 0.091 0.25

-3

10 cm 5.4 3.9 1.9 2.6 3.5 8.3

STP 3207 381 2294 727 509 284

N0.1-3 µm

0.3% 10 (kg fuel)-1 0.65 (64%) 1.08 (70%) 0.37 (75%) 0.80 (64%) 1.58 (40%) 6.15 (27%) 8-11 16 ( 20 31 14

0.52 0.38 0.23 0.18 0.34 1.07

a

Measured after engine warm-up. b EICCN for 0.1% supersaturation can be scaled according to NCCN (0.1%)/NCCN (0.3%). CCN activation diameter Dcrit, number concentrations of potential CCN at 0.1% and 0.3% supersaturation as well as of particles with Dp > 0.1 µm, and the CCN (0.3%) emission factor; the relative uncertainty of EFCCN is given in brackets.

c

0.1 µm, while for fresh emissions the peak is found at 0.05-0.06 µm (18, 19).

4. Implications In the course of the HERCULES project, ship exhaust particulate matter was characterized with respect to chemical composition and microphysical properties for engine load conditions from 10% to 110%. The data served as input for the calculation of black carbon emission indices, sulfur conversion efficiency, and cloud forming potential of emitted PM. Table 1 compiles BC and EC emission factors from literature and respective data from this study. Test rig studies or stack measurements (12, 14, 19, 20) report emission factors of 60-180 mg/(kg fuel) which vary by a factor of 3. Studies measuring particles in ship plumes using airborne platforms (16-20) report emission factors of 40-960 mg/(kg fuel) exhibiting a much stronger data variability. The best statistically evaluated data set emerged from refs 17 and 18 where more than 100 vessels were probed. Although data from single engine test bed studies or from stack sampling of cruising vessels do fall into the standard deviation of these data, it is obvious that they define the lower limit of the overall data set on BC emissions from shipping. While test bed studies may be conducted with well-maintained engines, this is not the case for stack-sampling on ocean-goings vessels. An explanation for this widespread of BC emission factors is still missing. 3804

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 10, 2010

CCN emission factor values calculated from size distributions for a supersaturation of 0.3% are at the lower bound of values reported in the literature (16, 20). Combining current knowledge, CCN emission factors of 1-30 × 1014 (kg fuel)-1 are reported. The influence of particle processing in the expanding ship exhaust plume may partially explain the differences in emission factors between exhaust studies and plume measurements. Taking diesel particles and particles emitted from a gas turbine with a thin sulfate coating as proxies for future emissions from shipping burning low-sulfur fuels or even sulfur-free biofuels, these particles require activation diameters above 300 nm at supersaturations of 0.2% which, however, are not observed in the fresh exhaust to a significant level. Under conditions in natural marine stratus clouds, particles emitted from burning low-sulfur fuel are therefore expected to have only a minor impact on clouds, and ship tracks will form only very rarely.

Acknowledgments This work was supported by the EU Integrated Projects HERCULES (Contract No. TIP3-CT-2003-506676) and QUANTIFY (Contract No. 003893 GOCE). We are grateful to Andreas Minikin and Hans Ru ¨ ba (DLR) for their valuable support during the extensive measurements and to Veronika Eyring (DLR) for fruitful discussions. We appreciate the comments of one anonymous reviewer who helped to improve the manuscript considerably. One of the authors

(A. Petzold) gratefully acknowledges the hospitality of Manchester Metropolitan University and the support by the UK OMEGA programme for a research stay.

Literature Cited (1) Eyring, V.; Isaksen, I. S. A.; Berntsen, T.; Collins, W. J.; Corbett, J. J.; Endresen, O.; Grainger, R. G.; Moldanova, J.; Schlager, H.; Stevenson, D. S. Transport impacts on atmosphere and climate: Shipping. Atmos. Environ. 2009, 43, in press. (2) Lauer, A.; Eyring, V.; Hendricks, J.; Jo¨ckel, P.; Lohmann, U. Global model simulations of the impact of ocean-going ships on aerosols, clouds, and the radiation budget. Atmos. Chem. Phys. 2007, 7, 5061–5079. (3) Fuglestvedt, J.; Berntsen, T.; Myhre, G.; Rypdal, K.; Skeie, R. B. Climate forcing from the transport sectors. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 454–458. (4) Corbett, J. J.; Winebrake, J. J.; Green, E. H.; Kasibhatla, P.; Eyring, V.; Lauer, A. Mortality from Ship Emissions: A Global Assessment. Environ. Sci. Technol. 2007, 41, 8512–8518. (5) Agrawal, H.; Eden, R.; Zhang, X.; Fine, P. M.; Katzenstein, A.; Miller, J. W.; Ospital, J.; Teffera, S.; III, D.; B, C. Primary particulate matter from ocean-going engines in the Southern California air basin. Environ. Sci. Technol. 2009, 43, 5398–5402. (6) Eyring, V.; Ko¨hler, H. W.; Aardenne, J. v.; Lauer, A. Emissions from international shipping: 1. The last 50 years. J. Geophys. Res. 2005, 110, D17305. (7) Dominguez, G.; Jackson, T.; Brothers, L.; Barnett, B.; Nguyen, B.; Thiemens, M. H. Discovery and measurement of an isotopically distinct source of sulfate in Earth’s atmosphere. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 12769–12773. (8) Lyyra¨nen, J.; Jokiniemi, J.; Kauppinen, E. I.; Joutsensaari, J. Aerosol characterisation in medium-speed diesel engines operating with heavy fuel oils. J. Aerosol Sci. 1999, 30, 771–784. (9) Cooper, D. A. Exhaust emissions from ships at berth. Atmos. Environ. 2003, 37, 3817–3830. (10) Petzold, A.; Feldpausch, P.; Fritzsche, L.; Minikin, A.; Lauer, P.; Bauer, H. Particle emissions from ship engines. European Aerosol Conference, Budapest, 2004. J. Aerosol Sci. 2004, 35, S1095–S1096. (11) Kasper, A.; Aufdenblatten, S.; Forss, A.; Mohr, M.; Burtscher, H. Particulate emissions from a low-speed marine diesel engine. Aerosol Sci. Technol. 2007, 41, 24–32. (12) Agrawal, H.; Malloy, Q. G. J.; Welch, W. A.; Wayne Miller, J.; Cocker, D. R., III. In-use gaseous and particulate matter emissions from a modern ocean going container vessel. Atmos. Environ. 2008, 42, 5504–5510. (13) Agrawal, H.; Welch, W. A.; Miller, J. W.; Cocker III, D. R. Emission measurements from a crude oil tanker at sea. Environ. Sci. Technol. 2008, 42, 7098–7103. (14) Moldanova´, J.; Fridell, E.; Popovicheva, O.; Demirdjian, B.; Tishkova, V.; Faccinetto, A.; Focsa, C. Characterisation of particulate matter and gaseous emissions from a large ship diesel engine. Atmos. Environ. 2009, 43, 2632–2641. (15) Chen, G.; Huey, L. G.; Trainer, M.; Nicks, D.; Corbett, J.; Ryerson, T.; Parrish, D.; Neuman, J. A.; Nowak, J.; Tanner, D.; et al. An investigation of the chemistry of ship emission plumes during ITCT 2002. J. Geophys. Res. 2005, 110, D10S90. (16) Sinha, P.; Hobbs, P. V.; Yokelson, R. J.; Christian, T. J.; Kirchstetter, T. W.; Bruintjes, R. Emissions of trace gases and particles from two ships in the southern Atlantic Ocean. Atmos. Environ. 2003, 37, 2139–2148. (17) Lack, D.; Lerner, B.; Granier, C.; Baynard, T.; Lovejoy, E.; Massoli, P.; Ravishankara, A. R.; Williams, E. Light absorbing carbon emissions from commercial shipping. Geophys. Res. Lett. 2008, 35, L13815. (18) Lack, D. A.; Corbett, J. J.; Onasch, T.; Lerner, B.; Massoli, P.; Quinn, P. K.; Bates, T. S.; Covert, D. S.; Coffman, D.; Sierau, B.; et al. Particulate emissions from commercial shipping: Chemical, physical, and optical properties. J. Geophys. Res. 2009, 114, D00F04.

(19) Petzold, A.; Hasselbach, J.; Lauer, P.; Baumann, R.; Franke, K.; Gurk, C.; Schlager, H.; Weingartner, E. Experimental studies on particle emissions from cruising ship, their characteristic properties, transformation and atmospheric lifetime in the marine boundary layer. Atmos. Chem. Phys. 2008, 8, 2387–2403. (20) Murphy, S. M.; Agrawal, H.; Sorooshian, A.; Padro´, L. T.; Gates, H.; Hersey, S.; Welch, W. A.; Jung, H.; Miller, J. W.; Cocker, D. R., III; et al. Comprehensive simultaneous shipboard and airborne characterization of exhaust from a modern container ship at sea. Environ. Sci. Technol. 2009, 43, 4626–4640. (21) Hobbs, P. V.; Garrett, T. J.; Ferek, R. J.; Strader, S. R.; Hegg, D. A.; Frick, G. M.; Hoppel, W. A.; Gasparovic, R.; Russell, L. M.; Johnson, D. W.; et al. Emissions from ships with respect to their effects on clouds. J. Atmos. Sci. 2000, 57, 2570–2590. (22) Petzold, A.; Scho¨nlinner, M. Multi-angle absorption photometry - A new method for the measurement of aerosol light absorption and atmospheric black carbon. J. Aerosol Sci. 2004, 35, 421–441. (23) Ulrich, E.; Beckmann, C.; Israe¨l, G. The characterization of carbon species in particulate matter by successive thermal desorption. J. Aerosol Sci. 1990, 21, S609–S612. (24) Schmid, H.; Laskus, L.; Abraham, H. J.; Baltensperger, U.; Lavanchy, V.; Bizjak, M.; Burba, P.; Cachier, H.; Crow, D.; Chow, J.; et al. Results of the carbon conference international aerosol carbon round robin test stage I. Atmos. Environ. 2001, 35, 2111– 2121. (25) Aiken, A. C.; DeCarlo, P. F.; Kroll, J. H.; Worsnop, D. R.; Huffman, J. A.; Docherty, K. S.; Ulbrich, I. M.; Mohr, C.; Kimmel, J. R.; Sueper, D.; et al. O/C and OM/OC ratios of primary, secondary, and ambient organic aerosols with high-resolution time-offlight aerosol mass spectrometry. Environ. Sci. Technol. 2008, 42, 4478–4485. (26) Gysel, M.; Crosier, J.; Topping, D. O.; Whitehead, J. D.; Bower, K. N.; Cubison, M. J.; Williams, P. I.; Flynn, M. J.; McFiggans, G. B.; Coe, H. Closure study between chemical composition and hygroscopic growth of aerosol particles during TORCH2. Atmos. Chem. Phys. 2007, 7, 6131–6144. (27) Gysel, M.; Nyeki, S.; Paulsen, D.; Weingartner, E.; Baltensperger, U.; Galambos, I.; Kiss, G. Hygroscopic properties of water-soluble matter and humic-like organics in atmospheric fine aerosol. Atmos. Chem. Phys. 2004, 4, 35–50. (28) Topping, D. O.; McFiggans, G. B.; Coe, H. A curved multicomponent aerosol hygroscopicity model framework: Part 1 Inorganic compounds. Atmos. Chem. Phys. 2005, 5, 1205–1222. (29) Topping, D. O.; McFiggans, G. B.; Coe, H. A curved multicomponent aerosol hygroscopicity model framework: Part 2 Including organic compounds. Atmos. Chem. Phys. 2005, 5, 1223–1242. (30) Kreidenweis, S. M.; Koehler, K.; DeMott, P. J.; Prenni, A. J.; Carrico, C.; Ervens, B. Water activity and activation diameters from hygroscopicity data - Part I: Theory and application to inorganic salts. Atmos. Chem. Phys. 2005, 5, 1357–1370. (31) Zhang, L.; Michelangeli, D. V.; Taylor, P. A. Influence of aerosol concentration on precipitation formation in low-level, warm stratiform clouds. J. Aerosol Sci. 2006, 37, 203–217. (32) Petzold, A.; Gysel, M.; Vancassel, X.; Hitzenberger, R.; Puxbaum, H.; Vrchoticky, S.; Weingartner, E.; Baltensperger, U.; Mirabel, P. On the effects of organic matter and sulphur-containing compounds on the CCN activation of combustion particles. Atmos. Chem. Phys. 2005, 5, 3187–3203. (33) Frick, G. M.; Hoppel, W. A. Airship measurements of ship’s exhaust plumes and their effect on marine boundary layer clouds. J. Atmos. Sci. 2000, 57, 2625–2648. (34) Osborne, S. R.; Johnson, D. W.; Bower, K. N.; Wood, R. Modificatin of the aerosol size distribution within exhaust plumes produced by diesel-powered ships. J. Geophys. Res. 2001, 106, 9827–9842. (35) Weingartner, E.; Burtscher, H.; Baltensperger, U. Hygroscopic properties of carbon and diesel soot particles. J. Aerosol Sci. 1997, 31, 2311–2327.

ES903681Z

VOL. 44, NO. 10, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3805