Geophysical Research Letters

Aerosol chemistry and climate: Laboratory studies of the carbonate component of mineral dust and its reaction products

Authors


Abstract

[1] The hygroscopicity, cloud condensation nucleation (CCN) activity and infrared optical extinction of CaCO3, a reactive component of mineral dust aerosol, and Ca(NO3)2, a product of atmospherically-aged CaCO3 through reaction with nitrogen oxides, have been measured. The hygroscopic growth and CCN activity of Ca(NO3)2 are orders of magnitude greater than CaCO3 and more similar to ammonium sulfate. Ca(NO3)2 particles also reflect a greater amount of near infrared radiation at higher relative humidity. These measurements provide the first quantitative assessment of the important changes in climate forcing that can occur as mineral dust aerosol is transported, reacted and aged in the atmosphere.

1. Introduction

[2] Mineral dust aerosol impacts the Earth's climate through the scattering and absorption of solar radiation [Sokolik and Toon, 1999] and the nucleation of clouds [DeMott et al., 2003; Rudich et al., 2002]. It can also impact the chemical balance of the Earth's atmosphere through heterogeneous chemistry [Bian and Zender, 2003; Tang et al., 2003]. One important issue when quantifying the impact of mineral dust on atmospheric chemistry and climate is in assessing how its physicochemical properties change as it is transported, reacted and aged in the atmosphere [Garrett et al., 2003; Ravishankara, 2005]. Although these issues are recognized as being important, they are poorly understood and little quantitative laboratory data are available on the link between mineral dust aerosol chemistry and climate.

[3] Mineral dust aerosol is a complex mixture of particles and particle aggregates of varying composition and mineralogy [Claquin et al., 1999]. The carbonate component of mineral dust (e.g., calcite (CaCO3) and dolomite (CaMg(CO3)2)) is particularly reactive and can comprise as much as 30% of the total mineral dust aerosol, depending on the source region [Claquin et al., 1999]. It is well known from field studies that there is a strong correlation between Ca2+ and NO3 [Dentener et al., 1996; Song and Carmichael, 2001; Sullivan et al., 2006] and recently single particle analyses have shown that CaCO3 particles can be completely converted to calcium nitrate (CaNO3)2) in the troposphere [Laskin et al., 2005; Matsuki et al., 2005]. This conversion can occur through two known pathways:

equation image
equation image

[4] There are several important changes in the physicochemical properties of the particles upon conversion of carbonate minerals to nitrate salts, including increased hygroscopicity. The cloud condensation nuclei (CCN) activity of the particle will increase with increasing hygroscopicity. Enhanced water uptake by atmospherically aged CaCO3 can also increase the mineral dust single scattering albedo [Vlasenko et al., 2005], a key variable in assessing the impact of mineral dust on climate [Seinfeld et al., 2004]. This paper reports on a quantitative comparison of the hygroscopicity, optical properties and CCN activity of CaCO3 and Ca(NO3)2 aerosols, providing important insights into the link between aerosol chemistry and climate as freshly emitted CaCO3 mineral dust aerosol is entrained in the atmosphere and undergoes heterogeneous reactions with nitrogen oxides.

2. Experimental Methods

[5] Size distributions and infrared spectra were acquired with a Multi-Analysis Aerosol Flow Reactor System (MAARS) designed to measure (i) hygroscopic growth and phase transitions; (ii) CCN activity; and (iii) extinction spectra of aerosols along with size distributions. Aerosols are generated with a constant output atomizer (TSI, Inc. Model 3076) from either a water suspension of insoluble CaCO3 (OMYA) particles or a 1–5 wt% by volume solution of Ca(NO3)2 (Alfa Aesar, ACS, 99.0–103.0%). A fraction of the aerosol is dried to a relative humidity (RH) <10% and then either fed directly into a flow reactor/hydration chamber or size selected with a differential mobility analyzer (DMA-1; TSI, Inc. Model 3080). For size-selected hygroscopic growth and CCN measurements, particles with mobility diameters of ca. 100 nm are used.

[6] The aerosol is combined with humidified air in the conditioning tube of the chamber and then flows into an observation tube positioned along the path of an IR spectrometer (Thermo Nicolet, Nexus Model 670 with MCT-A detector). An external RH system provides the steady air flow that can be varied between 3 and 90% RH for size-selected experiments. For full distributions, the aerosol flow bypasses DMA-1 and the infrared cell is used to measure the IR extinction spectra in the 800 to 7000 cm−1 range. RH above 65% is difficult to reach when measuring full size distributions.

[7] From the flow reactor/hydration chamber, the aerosol flow can be directed to either a scanning mobility particle sizer (SMPS) (TSI, Inc. Model 3936) or a continuous flow streamwise thermal-gradient CCN counter (Droplet Measurement Technologies, Model CCN-2). For growth and IR extinction measurements, particle size distributions are measured by an SMPS, which consists of a second DMA (DMA-2) and a condensation particle counter (CPC). For CCN measurements, the aerosol flow is divided between the CCN counter and the CPC, bypassing DMA-2.

3. Results and Discussion

[8] The hygroscopic growth of carbonate and nitrate aerosol particles as a function of relative humidity can be measured with size-selected particles using a tandem differential mobility analyzer [Rader and McMurry, 1986]. CaCO3 particles show no increase in diameter as a function of RH, while Ca(NO3)2 particles grow significantly and reach a diameter of 160 nm at higher RH, as shown in Figure 1. These data are best represented in terms of growth curves, displayed in Figure 2a. At each RH, the total hygroscopic growth is expressed as the growth factor,

equation image

where Dp is the particle diameter after humidification and D0 is volume equivalent diameter of the solid particle. There is no observable growth of CaCO3 particles up to 85% RH (Figure 2a), and g(RH) has an average value of 1.00 ± 0.02 over the entire RH range. Previous FTIR and thermogravimetric analysis measurements of water adsorption on CaCO3 show that 1–3 monolayers of water can be formed on the carbonate surface with increasing RH [Al-Hosney and Grassian, 2005; Gustafsson et al., 2005]. Assuming that three monolayers correspond to between 0.5–1 nm at most, this would be equal to a growth factor near 1.01, in good agreement with our results.

Figure 1.

Hygroscopic growth of size-selected 100 nm (a) CaCO3 and (b) Ca(NO3)2 particles as a function of RH.

Figure 2.

(a) Growth factors measured for size-selected 100 nm CaCO3 (filled circles) and Ca(NO3)2 (open circles). Each data point represents an average of 3–6 measurements. The error associated with each point is the standard deviation of these multiple measurements, which are on the order of the size of the symbols. The dashed line represents a g(RH) of one and the solid line is the theoretical growth curve for Ca(NO3)2 calculated from Köhler theory. (b) The calculated ratio of H2O to Ca(NO3)2, by both mole (left axis) and mass (right axis). The ratios are determined from the Köhler curve shown in Figure 2a. For comparison, the molar WSR deliquescence and efflorescence (solid lines, left axis) curves for (NH4)2SO4 are also displayed.

[9] The measured growth factor of size-selected 100 nm Ca(NO3)2 particles is quite different. Ca(NO3)2 particles formed from saturated solution droplets are amorphous hydrates [Tang and Fung, 1997]. These amorphous particles are never completely “dry” and become liquid droplets near 10% RH. Thus, even at the lowest relative humidity, the particles maintain a spherical shape and a Water to Solute molar Ratio (WSR) near 1 [Tang and Fung, 1997]. Because particles are never completely “dry” upon formation, g(RH) will never be unity and the particles will be aqueous droplets at atmospherically relevant RH. The “dry” particle diameter used to calculate the growth factor must be determined from Köhler theory [Gysel et al., 2002; Kreidenweis et al., 2005] which describes the growth of aqueous solution droplets in humid air. For Ca(NO3)2 a “dry” diameter of 89 nm (based on the theoretical g(RH) at 15% RH) was used as D0 in further calculations. With a growth factor of 1.13 ± 0.01 at 10% RH and 1.77 ± 0.05 at 85% RH, Ca(NO3)2 nearly doubles in diameter (corresponding to an 8-fold increase in volume) creating not only a highly aqueous environment for reactions but also can lead to an increase in the single scattering albedo of mineral dust aerosol.

[10] The WSR of Ca(NO3)2 on a mole and mass basis are shown in Figure 2b and are in excellent agreement with previous experiments [Tang and Fung, 1997]. At 15 and 70% RH a molar WSR of 2.94 ± 0.03 and 11.06 ± 0.23 is determined, corresponding to mass ratios of 0.32 ± 0.003 and 1.21 ± 0.03, respectively. To better understand the significance of the calculated WSR, the water content of Ca(NO3)2 can be compared with that of ammonium sulfate ((NH4)2SO4) and montmorillonite. (NH4)2SO4 is known to be important in the troposphere and its hygroscopic properties have been well-characterized [Tang and Munkelwitz, 1994]. Montmorillonite, another component of mineral dust, is a swellable clay mineral that can readily take up water. Below its efflorescence RH of 37%, (NH4)2SO4 is dry while Ca(NO3)2 contains as much as 39% water by mass. The water associated with Ca(NO3)2 is as much as 8 times greater than the water content in montmorillonite (∼5% water by mass at similar RH) [Frinak et al., 2005]. Between 37 and 80% RH, (NH4)2SO4 may contain some amount of water, but only after it has deliquesced at 80% RH. Near 80% RH, the molar WSR of Ca(NO3)2 measured in this work is greater than that of deliquesced (NH4)2SO4 by a factor of approximately 1.5. These results indicate that under typical atmospheric conditions, Ca(NO3)2 is highly hygroscopic and may provide a more aqueous environment for heterogeneous chemistry than (NH4)2SO4 and montmorillonite.

[11] Based on the increased hygroscopicity of Ca(NO3)2, it is expected that the CCN activity of Ca(NO3)2 will be markedly different from CaCO3. The CCN activity of 100 nm CaCO3, Ca(NO3)2 amorphous hydrate and (NH4)2SO4 aerosol particles were measured and are compared in Figure 3. The critical supersaturation, where 50% of the particles are activated, for Ca(NO3)2 is 0.11 ± 0.01% SS, the same as that found for (NH4)2SO4 [Seinfeld and Pandis, 1998], which represents a large component of CCN in the atmosphere [Ishizaka and Adhikari, 2003; Twomey, 1971]. In contrast, at 0.11% SS none of the CaCO3 aerosol is activated. This difference is further illustrated at 0.2% SS, where Ca(NO3)2 is 90% activated and CaCO3 is <2% activated. Based on these results, it can be concluded that as CaCO3 undergoes processing by HNO3 and N2O5 in the troposphere to form Ca(NO3)2, its ability to nucleate clouds increases significantly and is nearly two orders of magnitude greater between 0.075 and 0.2% SS compared to freshly emitted carbonate aerosol, and is similar to (NH4)2SO4.

Figure 3.

CCN activity for size-selected 100 nm Ca(NO3)2 amorphous hydrate (open circles), CaCO3 (filled circles), and (NH4)2SO4 (filled diamonds). The data are plotted as the fraction of CCN active particles (#CCN/#CN) as a function of percent water supersaturation (% SS). The inset shows an expanded view of the region from 0.075 to 0.225% SS. The dashed line at 0.5% SS indicates the critical supersaturation.

[12] Additional experiments examining the infrared extinction of CaCO3 and Ca(NO3)2 support size-selected hygroscopic growth and CCN results and can be used to further characterize water content of the particles and changes in light scattering, both as a function of RH. Adsorbed water is not present in the CaCO3 spectra (Figure 4a), consistent with the size-selected particle results. In contrast, Figure 4b shows a large peak in the O–H stretching, ν(H2O), and bending regions, δ(H2O), for Ca(NO3)2 at both low and high RH, indicative of water associated with the Ca(NO3)2 particles. This supports the conclusion that even at low RH the Ca(NO3)2 particles are amorphous hydrates. For Ca(NO3)2, the integrated absorbance of the O-H stretching region increases with RH (shown in Figure 4c) and has a shape similar to that of the size selected growth curve in Figure 2a.

Figure 4.

IR extinction spectra from 800 to 7000 cm−1 for (a) CaCO3, at 3 and 63% RH and (b) Ca(NO3)2 at 8 and 63% RH. Spectra have been offset for clarity and the zero baseline (dashed lines) for each is shown. In Figure 4a absorptions due to carbonate are seen at 879 and 1464 cm−1 and in Figure 4b absorptions due to the nitrate ion are seen at 1050 cm−1 and as a doublet between 1450 and 1350 cm−1. (c) Integrated absorbance of the O–H stretching region (2950 to 3700 cm−1), scaled to total aerosol volume.

[13] There is also an increase in the scatter region beyond 4000 cm−1 in the Ca(NO3)2 spectra of Figure 4b, signifying the presence of more larger particles. The relative amount of scattered light from 4000 to 7000 cm−1 is a factor of approximately three greater at the higher RH, determined from the increase of the integrated area of this region at high RH relative to low RH. There is only a slight increase in the infrared scatter region beyond 4000 cm−1 at higher RH for CaCO3. It has been estimated that highly hygroscopic particles reflect light 2–3 times more than dry particles in areas of high relative humidity [Day et al., 2000; McInnes et al., 1998]. These measurements agree well with this estimate, implying that an aged dust particle composed of Ca(NO3)2 would scatter appreciably more light than the original CaCO3 particle at higher RH.

4. Conclusions and Atmospheric Implications

[14] The effect an aerosol has on climate through direct and indirect radiative forcing is strongly dependent on its size, shape, and hygroscopic properties, all of which can change as a particle undergoes heterogeneous chemistry and atmospheric aging. We have quantified the differences in the physicochemical properties of carbonate minerals compared to the nitrate salt and have related these properties to changes in direct and indirect climate forcing. The increased hygroscopicity of Ca(NO3)2 relative to CaCO3 will result in larger, more spherical particles. CCN activity and scatter in the IR region will be enhanced as well. The observed increase the hygroscopicity and CCN activity of the mineral dust aerosol will also result in greater cloud reflectivity and a decrease in precipitation efficiency [Lohmann and Feichter, 2005]. The data presented in this paper is an important step in trying to better understand the impact that mineral dust aerosol chemistry with trace atmospheric gases has on climate. Ultimately, quantifying these effects on a global scale will require the use of detailed cloud microphysical and radiative forcing models, which would take into account and include heterogeneous chemistry, the presence of background particles, and meteorological conditions.

Acknowledgments

[15] This material is based upon work supported by the National Science Foundation under Grant No. CHE-0503854. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation. The authors would like to thank Professor Gregory Carmichael. Professor Paul Kleiber, Professor Mark Young, Professor Kimberly Prather, and Dr. Anthony Prenni for helpful discussions.

Ancillary