Anal. Chem. 2010, 82, 7943–7951
In Situ Characterization of Cloud Condensation Nuclei, Interstitial, and Background Particles Using the Single Particle Mass Spectrometer, SPLAT II† Alla Zelenyuk,*,‡ Dan Imre,§ Michael Earle,| Richard Easter,‡ Alexei Korolev,| Richard Leaitch,| Peter Liu,| Anne Marie Macdonald,| Mikhail Ovchinnikov,‡ and Walter Strapp| Pacific Northwest National Laboratory, Richland, Washington 99354, Imre Consulting, Richland, Washington 99352, and Environment Canada, Toronto, Ontario, M3H 5T4, Canada The aerosol indirect effect remains the most uncertain aspect of climate change modeling, calling for characterization of individual particles sizes and compositions with high spatial and temporal resolution. We present the first deployment of our single particle mass spectrometer (SPLAT II) operated in dual data acquisition mode to simultaneously measure particle number concentrations, density, asphericity, and individual particle size and quantitative composition, with temporal resolution better than 60 s, thus yielding all the required properties to definitively characterize the aerosol-cloud interaction in this exemplary case. We find that particles are composed of oxygenated organics, many mixed with sulfates, biomass burning particles, some with sulfates, and processed sea-salt. Cloud residuals are found to contain more sulfates than background particles, explaining their higher efficiency to serve as cloud condensation nuclei (CCN). Additionally, CCN sulfate content increased with time due to in-cloud droplet processing. A comparison between the size distributions of background, CCN, and interstitial particles shows that while nearly all CCN particles are larger than 100 nm, over 80% of interstitial particles are smaller than 100 nm. We conclude that for this cloud, particle size is the controlling factor on aerosol activation into cloud-droplets, with higher sulfate content playing a secondary role. Atmospheric aerosols affect climate by scattering and absorbing solar radiation and in their role as cloud condensation nuclei (CCN) and ice nuclei (IN) by determining cloud properties. The relationship between the properties of aerosol particles and clouds, i.e., aerosol indirect effect, is among the most uncertain aspects in our current understanding of climate change.1 The efficacy with which particles activate to form cloud droplets depends on their interaction with the atmospheric water vapor, † Part of the special issue “Atmospheric Analysis as Related to Climate Change”. * To whom correspondence should be addressed. E-mail: alla.zelenyuk@ pnl.gov. ‡ Pacific Northwest National Laboratory. § Imre Consulting. | Environment Canada.
10.1021/ac1013892 2010 American Chemical Society Published on Web 08/18/2010
which is strongly dependent on the particles size and composition, of which the soluble hygroscopic fraction is most important.2 The number of particles that become cloud droplets determines the cloud droplet number concentrations and size distributions, which relates directly to the cloud optical properties and precipitation. Higher particle concentrations may result in higher cloud droplets concentrations, smaller drop sizes, and increased cloud brightness.3 Smaller drop sizes can also decrease precipitation leading to more sustained cloud liquid water contents and increased cloud lifetime4-7 or have the opposite effect and increase precipitation.8 In order to properly include the aerosol indirect effect in climate models, we must be able to sufficiently represent the information on the size and mixing state (its internal chemical composition) of individual aerosol particles.2 Such measurements represent a significant challenge because they require high instrument sensitivity and temporal resolution,9,10 particularly for measurements that take place onboard aircraft. An aircraft is the most appropriate platform for CCN characterization, since below and in-cloud measurements have the potential to provide the most direct information about the differences between the physical and chemical properties of CCN and unactivated interstitial particles. Thus far, field campaigns aimed at characterizing aerosol cloud interactions have failed to produce sufficiently detailed information on the size and compositions of individual particles and their (1) IPCC Intergovernmental Panel on Climate Change (IPCC) Climate Change 2007. The Physical Science Basis: Contribution of Working Group I to the Fourth Assessment Report of the IPCC; Cambridge University Press: Cambridge, U.K., 2007. (2) McFiggans, G.; Artaxo, P.; Baltensperger, U.; Coe, H.; Facchini, M. C.; Feingold, G.; Fuzzi, S.; Gysel, M.; Laaksonen, A.; Lohmann, U.; Mentel, T. F.; Murphy, D. M.; O’Dowd, C. D.; Snider, J. R.; Weingartner, E. Atmos. Chem. Phys. 2006, 6, 2593–2649. (3) Leaitch, W. R.; Lohmann, U.; Russell, L. M.; Garrett, T.; Shantz, N. C.; ToomSauntry, D.; Strapp, J. W.; Hayden, K. L.; Marshall., J.; Wolde, M.; Worsnop, D.; Jayne, J. Atmos. Chem. Phys. Discuss. 2010, 10, 2131–2168. (4) Twomey, S. Atmos. Environ. 1991, 25, 2435–2442. (5) Albrecht, B. A. Science 1989, 245, 1227–1230. (6) Lubin, D.; Vogelmann, A. M. Geophys. Res. Lett. 2007, 34, L11801. (7) Lubin, D.; Vogelmann, A. M. Nature 2006, 439, 453–456. (8) Rosenfeld, D.; Lohmann, U.; Raga, G. B.; O’Dowd, C. D.; Kulmala, M.; Fuzzi, S.; Reissell, A.; Andreae, M. O. Science 2008, 321, 1309–1313. (9) Zelenyuk, A.; Imre, D. Int. Rev. Phys. Chem. 2009, 28, 309–358. (10) Zelenyuk, A.; Yang, J.; Choi, E.; Imre, D. Aerosol Sci. Technol. 2009, 43, 411–424.
Analytical Chemistry, Vol. 82, No. 19, October 1, 2010
7943
relation to CCN activity. A recent publication11 describes a groundbased field study that took place at high elevation and deployed two single particle mass spectrometers to characterize the size and composition of mixed-phase cloud nuclei. Since one of the instruments characterized only 162 CCN and the other 1094 CCN over the entire field study, the data had to be averaged over the duration of the field campaign, which included air masses from different regions, making it impossible to establish a direct relationship between the properties of individual particles and their CCN and IN activity. Moreover, the two instruments produced different views of what were the average compositions and size distributions of the background aerosol.11 Aside from this recent study, available data are limited to the information on bulk chemical composition of the assembly of aerosol particles.12,13 Moreover, the rapidly changing aerosol compositions encountered during airborne campaigns has thus far made it impossible to obtain the size-resolved chemical composition with reasonable time scales even for high (1 000-10 000 cm-3) CCN concentrations.14 There has been a number of CCN closure studies14-19 that used measured particle size distributions and bulk chemical composition to calculate the number of CCN and compared the results to real-time CCN measurements. In these studies, the absence of knowledge of individual particle size and composition necessitates assumptions regarding the particles mixing states. For example, Lance et al. showed that depending on the assumptions made about aerosol external and internal mixing states, the overprediction of the calculated CCN was on average between 3% and 36%.14 However, the scatter was significantly larger (200-300%) and difficult to account for. The authors hypothesized that the large scatter in CCN closure is mainly due to the externally mixed fraction of the aerosol, while the overprediction bias is largely controlled by internally mixed aerosol fraction, both of which were not measured. Unfortunately, a test of the model concludes that the data quality is insufficient to actually test the model. Quinn et al.20 and Furutani et al.21 concluded that the observed changes in the average aerosol compositions lead to significant (11) Kamphus, M.; Ettner-Mahl, M.; Drewnick, F.; Keller, L.; Cziczo, D. J.; Mertes, S.; Borrmann, S.; Curtius, J. Atmos. Chem. Phys. Discuss. 2009, 9, 15375–15421. (12) Drewnick, F.; Hings, S. S.; DeCarlo, P.; Jayne, J. T.; Gonin, M.; Fuhrer, K.; Weimer, S.; Jimenez, J. L.; Demerjian, K. L.; Borrmann, S.; Worsnop, D. R. Aerosol Sci. Technol. 2005, 39, 637–658. (13) Jayne, J. T.; Leard, D. C.; Zhang, X. F.; Davidovits, P.; Smith, K. A.; Kolb, C. E.; Worsnop, D. R. Aerosol Sci. Technol. 2000, 33, 49–70. (14) Lance, S.; Nenes, A.; Mazzoleni, C.; Dubey, M. K.; Gates, H.; Varutbangkul, V.; Rissman, T. A.; Murphy, S. M.; Sorooshian, A.; Flagan, R. C.; Seinfeld, J. H.; Feingold, G.; Jonsson, H. H. J. Geophys. Res. 2009, 114, D00f15. (15) Cantrell, W.; Shaw, G.; Cass, G. R.; Chowdhury, Z.; Hughes, L. S.; Prather, K. A.; Guazzotti, S. A.; Coffee, K. R. J. Geophys. Res. 2001, 106, 28711– 28718. (16) Medina, J.; Nenes, A.; Sotiropoulou, R. E. P.; Cottrell, L. D.; Ziemba, L. D.; Beckman, P. J.; Griffin, R. J. J. Geophys. Res. 2007, 112, D10s31. (17) Broekhuizen, K.; Chang, R. Y. W.; Leaitch, W. R.; Li, S. M.; Abbatt, J. P. D. Atmos. Chem. Phys. 2006, 6, 2513–2524. (18) Ervens, B.; Cubison, M.; Andrews, E.; Feingold, G.; Ogren, J. A.; Jimenez, J. L.; DeCarlo, P.; Nenes, A. J. Geophys. Res. 2007, 112, D10s32. (19) Cubison, M. J.; Ervens, B.; Feingold, G.; Docherty, K. S.; Ulbrich, I. M.; Shields, L.; Prather, K.; Hering, S.; Jimenez, J. L. Atmos. Chem. Phys. 2008, 8, 5649–5667. (20) Quinn, P. K.; Bates, T. S.; Coffman, D. J.; Covert, D. S. Atmos. Chem. Phys. 2008, 8, 1029–1042. (21) Furutani, H.; Dall’osto, M.; Roberts, G. C.; Prather, K. A. Atmos. Environ. 2008, 42, 3130–3142.
7944
Analytical Chemistry, Vol. 82, No. 19, October 1, 2010
changes in aerosol CCN activity. Somewhat in contrast, Dusek et al. concluded that CCN concentrations are mainly determined by the particle size distributions, with particle chemical composition playing only a secondary role.22 This very brief survey of studies of aerosol-cloud interactions illustrates the fact that absent knowledge of the sizes and compositions of individual particles that served as CCN and those that did not precludes firm model testing and leaves the question whether nearly opposing conclusions represent differing conditions or problems with measurements. In this paper, we describe the application of SPLAT II,9,10 our single particle mass spectrometer, to characterize properties of cloud condensation nuclei, interstitial, and background particles. SPLAT II operated for the first time in a dual data acquisition mode23,24 to simultaneously measure particle number concentrations, density, aerosol asphericity, and individual particle size and quantitative composition. The data presented here were acquired by SPLAT II on Flight 31 (F31) that took place on April 26, 2008 as part of the Indirect and Semi-Direct Aerosol Campaign (ISDAC), a month-long field campaign at the North Slope of Alaska.25 The main scientific objective of ISDAC was to improve our understanding of how changes in the size, composition, and concentration of aerosols particles influence cloud properties and their associated radiative forcing. A detailed overview of the campaign and the measurements that were carried out is given elsewhere.25 SPLAT II was deployed on the National Research Council of Canada Convair580 aircraft that was equipped with a suite of instruments for aerosol and cloud characterization designed to provide a detailed picture of the particle number concentrations, size distributions, concentrations of CCN and IN, along with cloud microphysics measurements. Two different particle inlets were deployed: a counterflow virtual impactor (CVI)26 inlet that was used for incloud sampling to characterize droplet residuals that served as seeds for the cloud droplets and ice crystals and an aerosol inlet (AI), to sample the overall clear sky aerosol population, particles above, and below clouds. In addition, sampling through the AI during in-cloud flight segments provides information about interstitial particles. During this and other flights, SPLAT II characterized particles with a wide range of compositions, both internally and externally mixed. Many of the particle types were observed in horizontal and vertical filamentous layers that appear as highly variable particle number concentrations and compositions throughout the flight. Proper characterization of these rapidly varying atmospheric air masses requires instrumentation with high temporal resolution. (22) Dusek, U.; Frank, G. P.; Hildebrandt, L.; Curtius, J.; Schneider, J.; Walter, S.; Chand, D.; Drewnick, F.; Hings, S.; Jung, D.; Borrmann, S.; Andreae, M. O. Science 2006, 312, 1375–1378. (23) Vaden, T. D.; Imre, D.; Zelenyuk, A. Aerosol Sci. Technol. 2010, accepted for publication. (24) Vaden, T. D.; Imre, D.; Zelenyuk, A. Aerosol Sci. Technol. 2010, accepted for publication. (25) McFarquhar, G. M.; Ghan, S.; Verlinde, J.; Korolev, A. J.; Strapp, W.; Schmid, B.; Tomlinson, J. M.; Wolde, M.; Brooks, S. D.; Cziczo, D.; Dubey, M. K.; Fan, J.; Flynn, C.; Gultepe, I.; Hubbe, J.; Gilles, M. K.; Laskin, A.; Lawson, P.; Leaitch, R.; Liu, P.; Liu, X.; Lubin, D.; Mazzoleni, C.; Macdonald, A.-M.; Moffet, R. C.; Morrison, H.; Ovchinnikov, M.; Shupe, M. D.; Turner, D. D.; Xie, S.; Zelenyuk, A.; Bae, K.; Freer, M.; Glen, A. Bull. Am. Meteorol. Soc. 2010, in press. (26) Hayden, K. L.; Macdonald, A. M.; Gong, W.; Toom-Sauntry, D.; Anlauf, K. G.; Leithead, A.; Li, S. M.; Leaitch, W. R.; Noone, K. J. Geophys. Res. 2008, 113, D18201.
During F31, a single-layer mixed stratocumulus cloud was encountered over mostly ice-covered ocean, about 80 miles north-northeast of the Department of Energy Atmospheric Radiation Measurement (DOE ARM) site at Barrow, Alaska. The goal of this flight was to identify which of the background particles acted as CCN and find the differences between CCN and interstitial particles. Note that while the sampled cloud was a mixed-phase stratocumulus, the number of ice crystals was more than 5 orders of magnitude smaller than that of the liquid cloud droplets, making it impossible to characterize the particles that served as INs. The properties of IN measured by SPLAT II during ISDAC will be presented in separate publications. Here we demonstrate how SPLAT II measurements of background, CCN, and interstitial aerosol are used to determine what separates particles that served as CCN from background aerosols. The most direct approach is to characterize background particles by performing measurements above and below the cloud using the AI and then sample droplet residual particles through the CVI inlet while flying in the cloud. A comparison between activated and background particles should, in principle, provides the desired information. However, we show that in-cloud processing can change the droplets compositions and thus the droplet residuals, making it impossible to unambiguously determine the CCN composition. An alternative approach relies on sampling above, below, and in-cloud through the AI, since as noted above, when sampling in the cloud through the AI only interstitial particles, i.e., particles that did not get activated are transmitted. In other words, instead of looking for the difference between droplet residuals and background aerosol, this sampling scheme helps determine the properties of particles that render them more difficult to activate. EXPERIMENTAL SECTION The Convair-580 aircraft was equipped with 42 instruments for aerosol and cloud characterization. A detailed overview of the measurements is given elsewhere. Here we focus on the measurements performed by SPLAT II and a few complementary measurements by other instruments. During ISDAC, SPLAT II successfully participated in all 27 flights that altogether lasted slightly over 100 h, without the need for realignment or adjustments throughout the entire field campaign. It measured the size of tens of millions of particles and characterized the composition of ∼3 million of them. SPLAT II is described in detail elsewhere. In brief, SPLAT II uses an aerodynamic lens to form a low divergence particle beam and transport the particles into the vacuum system with extremely high efficiencies.10 The aerodynamic lens imparts each particle with a velocity that is a narrow function of the particle’s vacuum aerodynamic diameter (dva).10,27 To measure individual particle size and composition, each particle is detected by light scattering at two optical detection stages that are spaced 10.5 cm apart. Particle time-of-flight (PTOF) between the two stages is used to determine particle velocity and thus its dva. A particle detection event and its PTOF are used to generate size-dependent triggers that fire the CO2 laser, used to evaporate semivolatile components in the particle, and the excimer laser that ionizes the evaporated plume and ablates the nonvolatile particle (27) Zelenyuk, A.; Imre, D. Aerosol Sci. Technol. 2005, 39, 554–568.
fraction.10,28,29 Individual particle mass spectra are acquired using an angular reflectron time-of-flight mass spectrometer (TOF-MS) and digitized by a 14 bit A/D converter. During the entire campaign SPLAT II operated in two simultaneous data acquisition modes: one characterizes the size and composition of individual particles at an operator determined rate of up to 100 particles/s. The second mode records particle number concentrations, size distributions, densities, and aerosol asphericity at a rate determined by the ambient particle number concentrations and instrument detection efficiency.23,24 Because of the high cabin temperature during F31, SPLAT II was set to acquire a maximum of 20 individual particle mass spectra per second to prevent the laser from overheating. Particle number concentrations are determined by SPLAT II with 1 s time resolution based on the measured particle detection rate in the first optical detection stage23 and found to be in excellent agreement with the measurements by the Passive Cavity Aerosol Spectrometer Probe (PCASP, PMS PCASP-100X) and the on-board condensation particle counter (CPC, TSI 7610). Particle asphericity is measured with 1 s temporal resolution by calculating the ratio of the particle detection rates in the two optical detection stages.23 This ratio is sensitive to particle beam divergence, which is affected by particle shape. For spherical particles the ratio is between 1 and 1.1 and for aspherical particles it increases to 2-4, depending on particle shape.23 Because spherical particles smaller than 100 nm also form more diverged aerosol beam, it is advisible to consider particle size distribution when determining particle asphericity. Particle dva size distributions are calculated from the measured PTOF with a temporal resolution on the order of ∼60 s, depending on the particle concentrations and the breadth of the size distributions.23 Particle dva distributions are also used to calculate particle density based on the instrument characteristic drop-off detection efficiency for small particles.24 This approach is especially useful when applied to the size distributions of composition resolved particles, which are obtained after the mass spectral data are classified.30 These densities together with the mass spectra are used to obtain quantitative information on particle composition and determine relative amounts of components in internally mixed aerosol particles.24,31 Detailed description of the methods and precision, with which particle number concentrations, asphericity, size distributions, and densities are calculated, are described in separate publications.23,24 RESULTS AND DISCUSSION During F31, the measurements were conducted above, below, and within the single layer stratocumulus cloud, yielding a complete data set that is ideally suited for a process-oriented study aimed at elucidating cloud-aerosol interactions and the aerosol affect on the microphysical and radiative properties of Arctic clouds, ISDAC’s main scientific objective. (28) Zelenyuk, A.; Yang, J.; Imre, D.; Choi, E. Aerosol Sci. Technol. 2009, 43, 305–310. (29) Zelenyuk, A.; Yang, J.; Imre, D. Int. J. Mass Spectrom. 2009, 282, 6–12. (30) Zelenyuk, A.; Imre, D.; Cai, Y.; Mueller, K.; Han, Y. P.; Imrich, P. Int. J. Mass Spectrom. 2006, 258, 58–73. (31) Zelenyuk, A.; Imre, D.; Han, J. H.; Oatis, S. Anal. Chem. 2008, 80, 1401– 1407.
Analytical Chemistry, Vol. 82, No. 19, October 1, 2010
7945
Figure 1a shows the layout of the complete F31 flight-altitude track as function of time, in which the region where particles were sampled through the CVI is indicated in cream color and in-cloud segments are marked by the blue areas. For ease of presentation, a number of key flight segments are numbered and labeled. The flight profile involved legs above (A1 and A2) and below (B1 and B2) the cloud to characterize background aerosol. Segments 1, 2, 3, and 4 correspond to in-cloud sampling, with 1, 2, and 4 sampled through the CVI inlet and 3 through the AI. In-cloud sampling through the CVI (1, 2, and 4) was carried out at nearly constant altitude (2, 4) and by porpoising through the cloud (1). In-cloud porpoising while sampling through the AI during segment 3 yields information about the interstitial aerosol. During the porpoising segments, 1 and 3, the aircraft altitude varied to sample above, below, and in the cloud a number of times. Below we will show that in-cloud sampling through both inlets, CVI and AI provide unambiguous information about the differences between activated and interstitial particles. The segment labeled 5 at higher altitude provides information about particles above the boundary layer. Figure 1b is plot of particle number concentrations measured by SPLAT II with 1 s temporal resolution. These number concentrations were independently calculated, have not been scaled to fit any other observable, and can directly be compared with the number concentrations measured by the CPC and PCASP shown in Figure 1c,d, while taking into account the fact that each of these instruments has a different small particle detection limit. The CPC detects all particles with diameters, d, above 14 nm and samples from the same inlets as SPLAT II. The PCASP detects particles with d > 125 nm, is mounted outside of the aircraft, and does not sample through the CVI. SPLAT II measures the number concentrations of particles larger than ∼100 nm. When SPLAT II samples through the AI, its measurements are in nearly perfect agreement with those by PCASP.23 A comparison with the CPC shows that when sampling through the AI, the CPC often detects more particles and the difference relates directly to the number of small particles (
