Anal. Chem. 2008, 80, 1401-1407
Articles
Simultaneous Measurements of Individual Ambient Particle Size, Composition, Effective Density, and Hygroscopicity Alla Zelenyuk,*,† Dan Imre,‡ Jeong-Ho Han,§ and Susan Oatis⊥
Pacific Northwest National Laboratory, Richland, Washington 99354, Imre Consulting, Richland, Washington 99352, Boston College, Chestnut Hill, Massachusetts 02467, and State University of New York, Stony Brook, New York 11794
The interaction between atmospheric particles and water vapor impacts directly and significantly the effect that these particles exert on the atmosphere. The hygroscopicity of individual particles, which is a quantitative measure of their response to changes in relative humidity, is related to their internal compositions. To properly include atmospheric aerosols in any model requires knowledge of the relationship between particle size, composition, and hygroscopicity. Here we demonstrate the capability to conduct in real time the simultaneous measurements of individual ambient particle hygroscopic growth factors, densities, and compositions using a humidified tandem differential mobility analyzer that is coupled to an ultrasensitive single-particle mass spectrometer. We use as an example the class of particles that are composed of sulfate mixed with oxygenated organics to illustrate how multidimensional single-particle characterization can be extended to yield in addition quantitative information about the composition of individual particles. We show that the data provide the relative concentrations of organics and sulfates, the density of the two fractions, and particle hygroscopicity. Atmospheric aerosols exert a significant effect on the global climate by scattering and absorbing solar radiation and in their role as cloud condensation nuclei (CCN). In addition, because of their public health hazard atmospheric aerosol loadings are regulated by the U.S. Environmental Protection Agency. The vast majority of atmospheric particles contain hygroscopic salts and acids that strongly interact with the atmospheric relative humidity (RH). These substances absorb water as the RH increases, resulting in changes in particle size, optical properties, and lung deposition efficiency. Decreasing RH brings about decrease in particle size and, if RH reaches values below the efflorescence points, phase transitions and crystals formation. At RH slightly above 100% aerosol act as CCN. Which of the particles get * Corresponding author. E-mail:
[email protected]. † Pacific Northwest National Laboratory. ‡ Imre Consulting. § Boston College. ⊥ State University of New York. 10.1021/ac701723v CCC: $40.75 Published on Web 02/07/2008
© 2008 American Chemical Society
activated to form cloud droplets and under what conditions strongly depends on the particles’ size and hygroscopic properties. In turn, the hygroscopic properties are directly related to the particle composition, and since most atmospheric particles are internally mixed, their hygroscopic properties depend not only on which of the hygroscopic substances they are composed of but also on what else is in the same particle and how much of it is present. Particle Hygroscopicity. The hygroscopicity of ambient particles is most commonly expressed in terms of their hygroscopic growth factor (GF), which is measured using a humidified tandem differential mobility analyzer (HTDMA)1,2 and is defined as the ratio of the particle diameter at high RH to the diameter of the same particle at near-zero RH. Note that by this definition GFs are dependent on the RHs at which they were measured and even on whether the RH was increased or decreased during the measurement. The observed hygroscopicity, or growth factors (GFs), are often interpreted to indicate that different particle compositions are present and are even used to deduce very approximate particle compositions. A solid fundamental understanding of the observed interaction between ambient particles and water vapor can best be developed on the basis of coupled measurements of particle compositions and hygroscopicities. A number of laboratory studies using HTDMA systems were conducted to quantify hygroscopic properties of laboratory-generated pure and mixed particles made of inorganic salts, organics, elemental carbon, and mineral dust. 3,4 Several studies on ambient aerosol, in which HTDMA systems were deployed in parallel with aerosol sampling by impactor5 or the aerodyne aerosol mass spectrometer6 observed some correlations between ambient particle hygroscopicity and bulk aerosol chemical composition. Since none of these measurements pro(1) Liu, B. Y. H.; Pui, D. Y. H.; Whitby, K. T.; Kittelson, D. B.; Kousaka, Y.; Mckenzie, R. L. Atmos. Environ. 1978, 12, 99-104. (2) Rader, D. J.; Mcmurry, P. H. J. Aerosol Sci. 1986, 17, 771-787. (3) Chan, M. N.; Chan, C. K. Atmos. Chem. Phys. 2005, 5, 2703-2712. (4) Vlasenko, A.; Sjogren, S.; Weingartner, E.; Gaggeler, H. W.; Ammann, M. Aerosol Sci. Technol. 2005, 39, 452-460. (5) Pitchford, M. L.; Mcmurry, P. H. Atmos. Environ. 1994, 28, 827-839. (6) 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. Atmos. Chem. Phys. Discuss. 2006, 6, 12503-12548.
Analytical Chemistry, Vol. 80, No. 5, March 1, 2008 1401
vided the composition of individual particles, they could not distinguish between internally mixed particles that are composed of a mixture of compounds and externally mixed particles composed of different single components. As a result they could not quantify the effect of complex internal aerosol mixtures on particle hygroscopicity. To address this deficiency McMurry et al.7 used the HTDMA to select the aerosol by both size and hygroscopicity for analysis of individual particle shape and composition by electron microscopy (EM). The authors observed systematic differences in particle composition and shape that correlated with GFs. Although the coupled HTDMA/EM system can in many cases provide the critically needed information that connects individual particle composition and hygroscopicity, the utility of this off-line technique is limited because it is slow and labor intensive. Particle Density. Particle density is another important attribute of atmospheric particles that is used to convert particle size distributions into mass loading, to model particle aerodynamic properties, and is often correlated to particle optical properties.8 Monitoring changes in particle densities can also be used to follow aerosol transformations due to chemical reaction.9 Most often particle densities are measured in such a way that the observed values are impacted to some extent by the particle shape. It is for that reason the term particle effective density is commonly used to distinguish between the true particle material density and a measured quantity that is dependent on particle density and shape.10,11 When particles are spherical and contain no internal voids the two are identical.12 Since particle densities and effective densities are also a function of particle compositions the observed densities of ambient particles are also often used to construe approximate particle compositions. Particle Composition. Direct measurements of particle compositions are most commonly carried out on bulk aerosol samples and are not the focus of the present work. Because particles’ behavior is a strong function of the internal composition of individual particles we will limit our discussion to measurements of individual particle compositions, which are typically conducted by single-particle mass spectrometers.13-17 These instruments are designed to identify the various components in each particle, including inorganic acids and salts, elemental carbon, and dust constituents. They also identify organic constituents and provide a characterization of the classes that these organics belong to, but in most cases they do not provide detailed list of all the molecular compounds that are present in each particle. (7) McMurry, P. H.; Litchy, M.; Huang, P. F.; Cai, X. P.; Turpin, B. J.; Dick, W. D.; Hanson, A. Atmos. Environ. 1996, 30, 101-108. (8) Tang, I. N.; Munkelwitz, H. R. J. Geophys. Res. [Atmos.] 1994, 99, 1880118808. (9) Katrib, Y.; Martin, S. T.; Rudich, Y.; Davidovits, P.; Jayne, J. T.; Worsnop, D. R. Atmos. Chem. Phys. 2005, 5, 275-291. (10) Zelenyuk, A.; Cai, Y.; Imre, D. Aerosol Sci. Technol. 2006, 40, 197-217. (11) DeCarlo, P. F.; Slowik, J. G.; Worsnop, D. R.; Davidovits, P.; Jimenez, J. L. Aerosol Sci. Technol. 2004, 38, 1185-1205. (12) Zelenyuk, A.; Cai, Y.; Chieffo, L.; Imre, D. Aerosol Sci. Technol. 2005, 39, 972-986. (13) Sullivan, R. C.; Prather, K. A. Anal. Chem. 2005, 77, 3861-3885. (14) Noble, C. A.; Prather, K. A. Mass Spectrom. Rev. 2000, 19, 248-274. (15) Murphy, D. M. Mass Spectrom. Rev. 2007, 26, 150-165. (16) Johnston, M. V. J. Mass Spectrom. 2000, 35, 585-595. (17) Coe, H.; Allan, J. In Analytical Techniques for Atmospheric Measurement; Heard, D., Ed.; Blackwell Publishing: Oxford, U.K., 2006.
1402
Analytical Chemistry, Vol. 80, No. 5, March 1, 2008
Despite this limitation single-particle mass spectrometers are suited to provide information whose content has sufficient detail to allow one to determine all other aerosol properties, like toxicity, index of refraction, density, and hygroscopicity. However, at present they cannot be used alone to unambiguously determine the relative concentrations of all the components. In the present work we are particularly interested in obtaining, for each particle, the relative concentrations of the organic and hygroscopic fractions that are very often found to reside in the same particles. Multidimensional Characterization. Ambient particle hygroscopicities, densities, indices of refractions, size distributions, and compositions are typically measured in field programs in parallel, on the same air mass, using a variety of instruments but not on the same particles. This approach leaves the question of which of the attributes is associated with what particle composition ambiguous. The most direct approach to eliminate many of these ambiguities is to simultaneously measure, for each individual particle, its size, internal composition, density, and hygroscopicity. Such a multidimensional data set has the potential to yield solid correlations between internal composition, density, and hygroscopicity for ambient particles. However, because under most conditions, the number concentration of particles that make it through the HTDMA is very low, to carry out such a study requires a single-particle mass spectrometer with extremely high sensitivity.18,19 Here we will describe an experimental setup that consists of an HTDMA system coupled with SPLATsour high-sensitivity single-particle mass spectrometer, which is capable of measuring simultaneously and in real time the size, composition, hygroscopicity, and effective density of individual particles. We will show how the combined measurements of all these properties on the same particles are used to obtain quantitative information on each of these important particle attributes. We will illustrate how these data can be used to quantify the organic mass fraction of ambient sulfate particles that are internally mixed with organics. EXPERIMENTAL SECTION SPLAT. SPLAT is a single-particle mass spectrometer that has previously been described in some detail.20 Here we provide only the relevant performance parameters. SPLAT uses an aerodynamic lens as an inlet to transport ambient particles from atmospheric pressure into the vacuum with extremely high efficiency. The aerodynamic lens forms a very narrow particle beam with very low divergence.21,22 In addition, the lens imparts on the particles a very narrow velocity distribution that is a function of the particle aerodynamic diameter. The relationship between the aerodynamic size and velocity is obtained from calibration runs using NIST certified size standards. We demonstrated that SPLAT can be used to measure particle aerodynamic diameters with better than 0.5% precision and accuracy,20 which in addition provides us the ability to measure the density of spherical particles, or the effective (18) Zelenyuk, A.; Imre, D.; Han, J.-H.; Oatis, S. EOS Trans. AGU 2003, 84 (46, Fall meeting Suppl.), Abstract A12B-0087. (19) Buzorius, G.; Zelenyuk, A.; Brechtel, F.; Imre, D. Geophys. Res. Lett. 2002, 29 (20), Article No. 1974. (20) Zelenyuk, A.; Imre, D. Aerosol Sci. Technol. 2005, 39, 554-568. (21) Liu, P.; Ziemann, P. J.; Kittelson, D. B.; Mcmurry, P. H. Aerosol Sci. Technol. 1995, 22, 293-313. (22) Liu, P.; Ziemann, P. J.; Kittelson, D. B.; Mcmurry, P. H. Aerosol Sci. Technol. 1995, 22, 314-324.
Figure 1. Measurements of particle hygroscopicity. The red histogram bars labeled “Dry” indicate a scan in which the RH of both DMAs was below 1%. The blue size distribution is for humidified 200 nm particles. The labels above the dashed lines indicate the positions where the second DMA was parked to classify particles to be characterized by SPLAT.
density of aspherical particles, with an ultimate precision of (0.5%.12 Particles are detected twice by light scattering of two green Nd:YAG lasers spaced by 16 cm apart. Particle velocity is determined from the measured time it takes the particle to travel between the two detection stages. The overall optical system leads to very high detection sensitivities. SPLAT detects and characterizes at least 1 particle/s, when the concentration of particles that are 200 nm or bigger are as low as 1 particle/cm3. Moreover SPLAT can detect and characterize particles as small as 50 nm with high efficiency. Overall, it is SPLAT’s extremely high sensitivity that makes the measurement scheme described in this paper possible. The two-stage detection and the calculated particle velocity are used to generate a trigger to fire an excimer laser synchronously with the arrival of the particle at the entrance to the reflectron time-of-flight mass spectrometer and acquire individual particles’ mass spectra. SPLAT can record ∼20 mass spectra/s. The individual particle mass spectra are classified, visualized, analyzed, and mined using ClusterSculptor and SpectraMiner, our dedicated data classification and mining software packages.23,24 In the present experiment 5860 particles were detected and characterized. The data were classified into 40 clusters the properties of some of which will be presented here. Complete analysis of the data will be presented in a separate publication. HTDMA. HTDMA setup was previously described elsewhere;19,25 here we provide a brief description of the salient points. In the present study prior to entering the first DMA ambient particles were dried with two diffusion dryers (TSI Inc., model 3062) connected in sequence. The RH in the first DMA was kept to be below 1%. The first DMA was set to select particles with (23) Zelenyuk, A.; Imre, D.; Cai, Y.; Mueller, K.; Han, Y. P.; Imrich, P. Int. J. Mass Spectrom. 2006, 258, 58-73. (24) Zelenyuk, A.; Imre, D.; Nam, E.-J.; Han, Y.; Mueller, K. Int. J. Mass Spectrom., submitted for publication, 2007. (25) Zelenyuk, A.; Imre, D.; Cuadra-Rodriguez, L. A.; Ellison, B. J. Aerosol Sci. 2007, 38, 903-923.
mobility diameters of 200 nm. The monodisperse aerosol was conditioned in the humidification chamber, where the RH was raised to 85% ( 2%. The size distributions of the humidified particles were measured by scanning the second DMA, in which the sheath flow was also conditioned to 85% ( 2% RH. All of the reported GFs in this study refer to 85% RH. The RH measurements were found to be reproducible to (2%, and the RH meter was checked and calibrated using ammonium sulfate particles. Figure 1 illustrates an example of the HTDMA measurements. The scan labeled “Dry” and marked by red histogram bars is a test run, in which the RH of both DMAs was kept below 1%. It was meant to test the system, define the resolution, and quantify the ambient particle number concentrations. The dry scan shows that the concentrations of particles with 200 nm mobility diameters were ∼20 particles/cm3. The blue histogram bars mark the result of a scan in which the first DMA was used to classify the dry (
