Aerosol Processes In The Stratosphere



It is now well established that chemical reactions involving aerosol particles play a key role in stratospheric ozone depletion.1-3 Some of these reactions take place on the surface of solid particles, while others occur inside liquid particles; both are commonly referred to as heterogeneous processes because they involve both the gas and the condensed phase.

The aerosol layer is located in the lower stratosphere and consists predominantly of aqueous sulfuric acid droplets, commonly labeled SSAs (sulfate stratospheric aerosols). At mid and low latitudes their concentration is 70 to 80% by weight H2S04, corresponding to mole fractions between ~0.3 and 0.5; at high latitudes and in the winter and spring months the SSAs may grow significantly in size, becoming polar stratospheric clouds (PSCs). As they cool, they absorb water vapor and also nitric acid vapor but remain in liquid form, becoming type la PSCs. If they freeze, they are labeled type lb PSCs, and at sufficiently low temperatures they become ice crystals (type II PSCs). The mechanism of conversion between the various stratospheric aerosol types is not well established and is discussed in Section 5 below. Typical particle sizes are ~0.1, 1, and 10 |im diameter for SSAs and PSCs type I and type II, respectively; their abundance is ~ 1 to 10 particles/cm3.


Stratospheric trace species include sources, free radicals (species with an unpaired electron), and temporary reservoirs. Source species are produced at Earth's surface

Handbook of Weather, Climate, and Water: Atmospheric Chemistry, Hydrology, and Societal Impacts, Edited by Thomas D. Potter and Bradley R. Colman. ISBN 0-471-21489-2 © 2003 John Wiley & Sons, Inc.

and are stable enough to eventually reach the stratosphere. Temporary reservoirs are generated in the stratosphere, but they are ultimately transported downwards into the troposphere, where they are rapidly removed by rainout or washout. Both sets of species decompose in the stratosphere producing free radicals—either by photolysis or by reaction with another radical. Free radicals can participate in ozone destruction cycles but can also react with each other to produce stable reservoirs. Thus photochemistry is a source of radicals, while gas-phase reactions interconvert radicals into different forms or else destroy free radicals by producing stable reservoirs. Practically all gas-phase chemical reactions involve free radicals; reactions in the gas phase between nonradical (saturated) species are usually too slow to matter at atmospheric temperatures. However, aerosols provide a pathway for such reactions to take place.

The two most important sets of heterogeneous reactions in the stratosphere are chlorine activation and nitrogen deactivation. Chlorine activation reactions transform temporary chlorine reservoirs to a form that is photolytically active; the most important ones are the following:

HCl + CIONO2 ->- HNO3 + Cl2 HCl + HOC1 H20 + Cl2

The species Cl2 is photolytically very active; it readily absorbs near-ultraviolet (UV) and visible light to produce free CI atoms, which in turn react rapidly with ozone:

The most important nitrogen deactivation reaction is N205 + H20 2HN03 The N205 species is produced in turn from nitrogen oxides:

The net effect of reactions (5), (6), and (7) is to convert active forms of nitrogen (NOj) to the relatively stable temporary reservoir HN03, a species which, however, may regenerate radicals by solar photolysis:

On the other hand, a significant fraction of the gas-phase HN03 is incorporated at low temperatures into PSCs, where it is further stabilized and protected against solar photolysis. Yet another effect of this HN03 scavenging process is denitrification: if


some fraction of the PSC particles grow sufficiently large, they may settle to lower altitudes, thus removing the NOx source more permanently.

The combined effect of chlorine activation and nitrogen deactivation is accelerated ozone depletion: Chlorine free radicals destroy ozone more rapidly in the absence of nitrogen radicals because the two sets of radicals tend to react with each other to produce temporary reservoirs, slowing ozone depletion. The most important example of this process is the formation of C10N02 (chlorine nitrate) through the following radical recombination reaction:

Reaction (9) is termolecular, and M is the "third" body (mostly N2 or 02) required to stabilize the newly formed bond. In contrast to other radical recombination products such as C1NO (nitrosyl chloride), CONO (chlorine nitrite), and C1N02 (nitryl chloride), the species C10N02 is relatively stable toward solar photolysis.4 However, in the presence of HC1 it rapidly regenerates the chlorine radicals by the chlorine activation reaction discussed above [reaction (1)].


Rate constants for heterogeneous reaction rates are commonly expressed in terms of a reaction probability y, which is the probability per collision of a gas-phase reactant molecule with the aerosol surface that chemical reaction will occur. For reaction on a solid aerosol surface of a species with a mean gas-phase concentration [C] molecule/cm3), the overall rate may be approximated by the following expression, which is based on the "resistance" model3:

where t is time (s), k, is the effective overall first-order rate constant (cm/s) for surface reaction, S is the aerosol surface area per unit volume (cm2/cm3), 1 /km- is the resistance associated with gas-phase diffusion, 1 /kcou is the resistance associated with molecular collisions with the particle surface, Dg is the gas diffusion coefficient (cm2/s), Rp is the average particle radius (cm), and (v) is the mean molecular speed (cm/s) of the gas-phase reactant. For values of y of less than ~0.2, the expression y/(l — y/2) may be approximated by y. For certain conditions Eq. (11) can be further simplified, e.g., at low pressures and for small particles (large Dg and small Rp), the effect of gas-phase diffusion may be neglected (1 /km % 0); etc. On the other hand, for small particlcs with sizes approaching the gas-phase mean free path, additional correction factors are needed.5 For reaction on liquid particles, liquid-phase diffusion also needs to be taken into account, leading to additional resistance terms in Eq.

1 /*, = 1/Adiff + 1 Aeon = Rp/Dg + 4/[<c)y/(l - y/2)]

The rate of reaction (5) on sulfuric acid solutions is nearly independent of the acid concentration,3'4 hence the reaction occurs readily at low and midlatitudes. In contrast, the rates of reactions (1) and (2) are negligible at those latitudes: The reaction mechanism involves as a first step incorporation of HC1 vapor into the condensed phase, and HC1 is practically insoluble in concentrated H2S04 solutions. On the hand, as the sulfuric acid particles cool and become more dilute at higher latitudes, the solubility of HC1 increases sharply, and the reaction probability increases accordingly, reaching values larger than ~0.1 for temperatures below 200 K.

As the particles freeze at high latitudes, the reaction probabilities may be strongly affected: hydrolysis of N205 [reaction (5)] becomes very slow, while reactions (1) and (2) occur very efficiently on ice surfaces, requiring only a few collisions of the reactant C10N02 or HOC1 with the particles exposed to HC1 vapor.4 Thus nitrogen deactivation [reaction (5)] occurs predominantly at mid and low latitudes and also at high latitudes as long as the aerosol particles remain liquid. In contrast, chlorine activation [reactions (1) and (2)] occurs efficiently on both liquid and solid particles, but only at high latitudes where the temperature drops below a threshold value of about 195 K.

The mechanism of reactions (1) and (2) is ionic in nature: HC1 solvates in aqueous phase forming hydrochloric acid, and hence chloride anions. The chlorine atom in HOC1 and C10N02 is slightly electropositive; both of these species react very fast with the chloride anions to produce molecular chlorine, which is rapidly desorbed from the ice surface.

The first step in the mechanism of reactions (1) and (2) on ice particles involves incorporation of HC1 vapor into the surface layers. The high affinity of HC1 for the ice surface is a consequence of ion pair formation, which takes place because the surface layers of ice are not as ordered as the ice crystal itself; they form a "liquidlike" aqueous layer in the presence of trace amounts of HC1 (this species can depress the freezing point of water down to ~ 195 K). The amount of energy associated with physical adsorption involving only a hydrogen bond is too small to explain the experimental observations of HC1 uptake by ice6; hence, reaction mechanisms involving weak physical adsorption and resorting to conventional Langmuir-type adsorption isotherms are not suitable.

The second step in the reaction mechanism involves incorporation into the surface layers of HOC1 for reaction (2), or C10N02 for reaction (1). All the reactants have a high mobility on the surface: Once incorporated into the condensed phase, the HOC1 molecules almost always find chloride ions before returning to the gas phase. Similarly, the C10N02 molecules in the surface layers also find chloride ions before reacting with water. This explains the experimental observation of a lack of dependence of the reaction rate on the concentration of HC1 vapor: as long as HC1 is in excess, the overall reaction is nearly zero order in HC1 and first order in HOC1 (or C10N02), and is only very weakly dependent on temperature.

Reactions (1) and (2) also occur rapidly on nitric acid trihydrate (NAT) surfaces, with a mechanism similar to that on ice surfaces. However, there is an additional parameter that should be taken into account, namely the relative humidity. When


NAT is in equilibrium with ice, its H20 vapor pressure is the same as that of ice, and the reaction probability y for reactions (1) and (2) is practically the same as that on ice. As the relative humidity (and hence the H20 vapor pressure of NAT) decreases, the reaction probability y initially remains high, but it decreases for relative humidity values below ~50% (with respect to ice), and eventually it reaches values more than two orders of magnitude smaller. This behavior can be explained with a reaction mechanism involving the availability of water at the NAT surface to induce solvation and uptake of HC1 vapor: at very low relative humidities, there is excess HN03 on the surface and solvation is hindered.


To investigate the nature and chemical identity of stratospheric aerosols, it is useful to consider first the thermodynamic properties of the aerosols, and subsequently the rates of transformation between the various phases for the different chemical systems of interest. The primary thermodynamic properties of interest are the mole fractions of the various chemical components of the particles in the condensed phase and their vapor pressures, the partial pressures or concentrations of these components in the gas phase, and the temperature. The vapor pressures for low-volatility components (e.g., NaCl, and sometimes H2S04) need not be considered explicitly; furthermore, the effect of total pressure on thermodynamic properties is negligible for atmospheric conditions.

The thermodynamic properties that determine the stability and equilibrium composition of the various phases can be represented conveniently by phase diagrams. A comparison of the atmospheric partial pressures with the vapor pressures displayed in the phase diagrams provides a useful guideline to establish the identity of the various condensed phases that can exist under atmospheric conditions. A specific atmospheric condition or state can be represented by a point in a phase diagram, while an atmospheric process or trajectory is represented by a line.

To illustrate the use of phase diagrams, consider the H2S04/H20 system. Figure 1 shows the equilibrium freezing temperatures for this system as a function of composition, and Figure 2 is a logarithmic plot of the water vapor pressure versus inverse temperature. The dashed lines in Figure 2 represent the H20 vapor pressure of solutions with constant composition; it follows from the Clausius-Clapeyron equation that these lines are nearly straight, their slopes being equal to the partial molar enthalpy of evaporation of H20. The solid lines in Figures 1 and 2 represent conditions of coexistence of two condensed phases; the lines separate regions of stability for the different phases. Note also that the stability region for a particular hydrate is represented in Figure 2 by a surface, whereas in Figure 1 it is a vertical line, as the composition of the hydrate is fixed.

Phase diagrams represent properties at thermodynamic equilibrium. However, it often happens that a new phase does not form because of the presence of kinetic barriers to nucleation. This is the case for sulfuric acid solution droplets: They remain liquid throughout most of the stratosphere, even though their temperature

%Wt H2S04

%Wt H2S04

Mole Fraction H2SO4

Mole Fraction H2SO4

Figure 1 Temperature vs. composition phase diagram for the HsS04/H20 system. The solid line represents freezing temperatures (solid-liquid coexistence conditions). The thin vertical lines give the composition of the solids. The dotted line represents the equilibrium composition of metastable liquid particles in the stratosphere as a function of temperature, in an air parcel containing 3 ppmv of water vapor at ~16km altitude; TF indicates the ice frost point for this air parcel.

is below the freezing point; that is, they supercool very readily. Phase diagrams still provide useful representations of such metastable phases; for example, in Figure 2 the vapor pressures of supercooled solutions are given by extensions of the dashed lines into the solid stability regions. Consider, for example, a stratospheric air parcel around 16 km altitude containing 3 parts per million (ppm) of water vapor, and cooling between 220 and 190 K; the properties of liquid sulfuric acid aerosols in such a parcel are represented by the dotted lines in Figures 1 and 2. The droplets swell and become less concentrated as the temperature drops.

There are similar phase diagrams for the HN03/H20 system; however, because of the relatively high volatility of HN03 compared to H2S04, it is useful to consider an additional phase diagram consisting of a logarithmic plot of the HN03 vapor pressure versus inverse temperature in order to elucidate the nature of the phases that are stable in the stratosphere for this system. Yet another version of a phase diagram is shown in Figure 3: It is a logarithmic plot with the vapor pressure of one component (HN03) in one axis and the vapor pressure of the other component (H20) in the

Weight %

Figure 3 Nitric acid vs. water vapor pressure phase diagram for the HN03/H20 system. The thick solid lines represent coexistence conditions for two condensed phases; the thin lines are isotherms (labeled with T in Kelvin). The dashed lines represent vapor pressures of liquids with constant composition (labeled as wt % HN03 in the upper and right axis). The dashed region in the lower left corner represents typical conditions in the lower stratosphere over the poles.

H2O Vapor Pressure, Torr

Figure 3 Nitric acid vs. water vapor pressure phase diagram for the HN03/H20 system. The thick solid lines represent coexistence conditions for two condensed phases; the thin lines are isotherms (labeled with T in Kelvin). The dashed lines represent vapor pressures of liquids with constant composition (labeled as wt % HN03 in the upper and right axis). The dashed region in the lower left corner represents typical conditions in the lower stratosphere over the poles.

from laboratory experiments. However, the vapor pressures can also be determined indirectly by other means, e.g., from voltage measurements in electrochemical cells—accurate phase diagrams can be constructed from measurements of such voltages together with calorimetric measurements of the enthalpies and temperatures of the various phase transitions of interest. For ternary systems such as H2S04/HN03/H20 the phase diagrams are more complicated but are nevertheless just extensions of the binary diagrams to one more dimension. On the other hand, the vapor pressures for such multicomponent systems can be reliably estimated using semiempirical thermodynamic models.7


The source of sulfuric acid in the stratosphere is carbonyl sulfide (COS), which is of biological origin. Although emitted from the ground, it is sufficiently stable to reach


the stratosphere, where it oxidizes to form sulfur dioxide, S02. This species is further oxidized to produce H2S04 through the following mechanism:

The rate-determining step is reaction (12). Reaction (14) is second order in water vapor2; it is fast throughout the atmosphere, except in the upper stratosphere, where the water vapor concentration is relatively small. There is no net consumption of radicals with this mechanism in the atmospheric oxidation of S02; the overall effect is merely the conversion of OH into H02.

A second important sulfur source consists of volcanic eruptions, a few of which inject S02 directly into the stratosphere, such as El Chichon in Mexico, in 1982, and Mount Pinatubo in the Philippines, in 1991. Mount Pinatubo introduced enough S02 to increase the stratospheric H2S04 burden by a factor of ~30,8 inducing noticeable global cooling. Satellite observations indicate that the S02 oxidation process takes several weeks, and that the excess particles remain in the stratosphere a couple of years; the sulfuric acid haze formed is the origin of bright red sunsets.

As mentioned above, at high latitudes and in the winter months, the sulfuric acid/water droplets cool and grow to become PSCs, absorbing water and nitric acid vapor. Observations in the lower stratosphere of a rapid growth in the volume of these aerosol particles around 195 K were originally interpreted as resulting from the formation of nitric acid trihydrate (NAT); however, a more recent analysis of the field observations indicates that the particles often remain liquid, reaching compositions such as 30 wt % H2S04 and 30 wt % HN037, with the particle growth being a consequence of rapid H20 and HN03 uptake below a threshold temperature, which happens to approximately coincide with the temperature below which NAT becomes stable.

There is a large nucleation barrier for these supercooled liquid particles to freeze, and laboratory observations show that freezing does not occur until the temperature has dropped several degrees below the ice frost point, which is around 185 K in the lower stratosphere. Under such conditions water ice crystallizes first, leading to the formation of type II PSCs. Some atmospheric observations indicate the presence of solid particles at temperatures above the frost point; it is likely, however, that such particles had reached lower temperatures earlier and that water ice induced the formation of the acid hydrates. As the particles warm up, ice evaporates first and eventually the hydrates melt at the equilibrium phase transition temperatures expected from the phase diagrams (i.e., temperature TM in Fig. 1), since there is essentially no nucleation barrier for the melting process.

Many questions remain, however, regarding the nature and the rates of liquidsolid phase transformations in PSCs. For example, in the Arctic stratosphere temperatures fall below the frost point much less frequently than over Antarctica, and yet solid PSCs do form, perhaps as a consequence of mesoscale temperature

S02 + OH H0S02 H0S02 + 02 H02 + S03 S03 + 2H20 H2S04 + H20

fluctuations.9 There are also questions regarding denitrification, the PSC sedimentation process referred to above leads to the removal of nitric acid, and hence, of nitrogen oxides. The process is not sufficiently well understood to permit reliable predictions, for example, of H20, HN03, and NOx levels at high latitudes for scenarios involving emissions from proposed future supersonic transports that would fly in the lower stratosphere.


1. Molina, M. J., L. T. Molina, and D. M. Golden, Environmental chemistry (gas and gas-solid interactions): The role of physical chemistry, J. Phys. Chem., 100, 12888, 1996.

2. Molina, M. J., L. T. Molina, and C. E. Kolb, Gas phase and heterogeneous chemical kinetics of the troposphere, Ann. Rev. Phys. Chem., 47, 327, 1996.

3. Kolb, C. E., D. R. Worsnop, M. S. Zahniser, P. Davidovits, C. F. Keyser, M. T. Leu, M. J. Molina, D. R. Hanson, A. R. Ravishankara, L. R. Williams, and M. A. Tolbert, Laboratory studies of atmospheric heterogeneous chemistry, in J. R. Barker (Ed.), Advanced Series in Physical Chemistry: Progress and Problems in Atmospheric Chemistry, World Scientific Publishing, Singapore, 1995, pp. 771-875.

4. DeMore, W. B., S. P. Sander, D. M. Golden, R. F. Hampson, M. J. Kurylo, C. J. Howard, A. R. Ravishankara, C. E. Kolb, and M. J. Molina, Chemical Kinetics and Photochemical Data for Use in Stratospheric Modeling, Evaluation No. 12, JPL Publication 97-4, Jet Propulsion Laboratory, Pasadena, CA, 1997.

5. Seinfeld, J. H., and S. N. Pandis, Atmospheric Chemistry and Physics from Air Pollution to Climate Change, Wiley, New York, 1997.

6. Molina, M. J., The probable role of stratospheric "ice" clouds: Heterogeneous chemistry of the "ozone hole," in J. G. Calvert (Ed.), The Chemistry of the Atmosphere: Its Impact on Global Change, Blackwell, Oxford, 1994, pp. 27-38.

7. Peter, T., Microphysics and heterogeneous chemistry of polar stratospheric clouds. Ann. Rev. Phys. Chem., 48, 785, 1997.

8. McCormick, M. P., L. W. Thomason, and C. R. Trepte, Atmospheric effects of the Mt Pinatubo eruption, Nature, 373, 399, 1995.

9. Carslaw, K. S., M. Wirth, A. Tsias, B. P. Luo, A. Dornbrack, M. Leutecher, H. Volkert, W. Renger, J. T. Bacmeister, and T. Peter, Particle microphysics and chemistry in remotely observed mountain polar stratospheric clouds, J. Geophys. Res., 103, 5785-5796, 1998.


0 0

Post a comment