Aviation and the Global Atmosphere

Other reports in this collection

2.1.3. Effects of Aircraft Aerosol Emissions

There is strong laboratory and observational evidence for the importance of aerosols in atmospheric photochemical processes such as lower stratospheric ozone depletion (WMO-UNEP, 1995), tropospheric SO2 oxidation (Calvert et al., 1985), and tropospheric soluble trace gas removal. A large number of additional chemical processes involving aerosols has been postulated; however, a lack of information on the abundance, nature, and reactivity of the various aerosol types (e.g., dust, soot, sulfate, nitrate, organic) has hindered an accurate assessment. In this section, we summarize the connections between aerosol processes and ozone chemistry in the UT and LS. In addition, we assess the nature of aircraft-derived aerosol impacts on ozone abundances at aircraft cruising altitudes. Aerosols and Tropospheric Ozone Chemistry

Reactions occurring on the surface of solid and liquid aerosols (i.e., heterogeneous reactions) or inside aqueous aerosols (i.e., homogeneous bulk reactions) can lead to a decrease in the production of ozone (reactions 1-5) by catalyzing the removal of NOx and HOx. Generally, heterogeneous reactions counteract reactions involving free radical species. From the ozone production perspective, removal of active species can occur by irreversible deposition of HOx, NOx, or NOx and HOx source species or by conversion of more reactive nitrogen- and hydrogen-containing species into less reactive ones. Conversion of N2O5 into HNO3 on sulfate or ice particles is the best established example of the latter mechanism. Wet deposition of HNO3, hydrogen peroxide (H2O2), and other soluble acids is an example of the former mechanism. Because removal of active species can also occur by gas-phase mechanisms, the importance of this aerosol chemistry depends on the relative rates of the gas and aerosol processes. The rates of heterogeneous processes depend on the available aerosol surface area as well as on their size distribution. As summarized in Chapter 3, typical mid-latitude upper tropospheric soot and sulfate aerosol surface areas are 10 �m2 cm-3 (see also Pueschel et al., 1997). Much larger water-ice surface areas, on the order of 104 �m2 cm-3, are found inside young contrails (Petzold et al., 1997) and natural cirrus clouds (Dowling and Radke, 1990).

Aircraft impact studies have motivated laboratory investigations of heterogeneous soot reactions. Interest in reactions occurring on soot has also increased in response to suggestions, based on analysis of field observations (Hauglustaine et al., 1996; Jacob et al., 1996; Lary et al., 1997), that aerosol-assisted conversion of HNO3 to NOx may occur under some unknown conditions. Processes like the HNO3-to-NO2 conversion, which result in a reduction of the oxidation state of an atmospheric species, are of particular interest because they proceed counter to the overall tendency of the troposphere to act as an oxidizing medium. For the specific HNO3 example, the reduction mechanism operates at the expense of the oxidation of the soot substrate and would promote production of ozone by increasing levels of active NOx.

Some of the most important heterogeneous reactions on soot involving NOx and its reservoirs, which are of likely importance in the UT and have been studied in the laboratory, are presented below. In view of the fact that the number concentration of nonvolatile particles-the majority of which are presumed to be soot-in a young contrail is on the order of 104 particles cm-3 (Brasseur et al., 1998), we start our discussion with heterogeneous oxidation-reduction reactions that may occur on soot. Heterogeneous kinetic studies involving "soot" have been performed on a variety of substrates, encompassing materials as diverse as commercially available amorphous carbon, carbonaceous material ("active carbon" or "carbon black"), and soot from hydrocarbon diffusion flames that use fuels such as hexane, toluene, ethylene, and acetylene. Therefore, comparisons of results obtained on different substrates should be made with caution. The heterogeneous reaction of NO2 with soot may be represented by the following reactions (Tabor et al., 1993, 1994; Rogaski et al., 1997; Gerecke et al., 1998):

NO2 + soot NO + [soot . O] (19a)

followed by the thermal decomposition of the oxygen adduct [soot . O]:

[soot . O] CO, CO2 (19b)
NO2 + soot + H2O HONO + [soot . OH] (19c)
NO2 + H(ads) HONO (19d)

where [soot.O] and [soot.OH] represent surface sites on soot that have been oxidized, hence deactivated by the heterogeneous reaction. It is not clear yet if soot participates as a reducing agent (reaction 19c) or simply as a reservoir of hydrogen (H(ads), reaction 19d) in its reaction with NO2 to yield HONO.

Reactions 19a and 19c correspond to a reduction-oxidation reaction in which soot is the reducing agent. Soot is oxidized in the process and releases CO and CO2 upon heating, although the primary oxidation product [soot.O] has not been characterized on a molecular level. It is thought that the surface of soot will be modified in the oxidation process, resulting in the accumulation of multiple functional groups bearing oxygen (Chughtai et al., 1990).

The branching ratio between reactions 19a and 19c or 19d depends on the soot sampling location as well as the fuel used to produce the soot, albeit to a minor extent. In general, amorphous carbon does not generate HONO, whereas soot sampled early in its formation process may give rise to a HONO yield of up to 90% compared to the NO2 taken up (Gerecke et al., 1998). Measurements within exhaust plumes have established the presence of nitrous acid (HNO2) and HNO3 at concentrations well above background (Arnold et al., 1992, 1994). The data indicate that about 0.6% of the NOx is converted to HNO2 and HNO3. More recent data on exhaust plumes of five B-747s and one DC-10 increase this efficiency to 1-5% of the NOx (Brasseur et al., 1998). Therefore, care must be exercised when HONO concentrations measured in the plume are used to infer OH concentrations or emission indices for OH using known reaction parameters for OH + NO (Tremmel et al., 1998).

Another pair of heterogeneous reactions occurring on soot have been shown to be of importance in the troposphere (Brouwer et al., 1986):

N2O5 + H2O + [soot] 2HNO3 + [soot] (20a)
N2O5 + [soot] NO2, NO + [soot . O] (20b)
[soot.O] CO, CO2 (19b)

Reaction 20a is a hydrolysis reaction for which soot is the support material, whereas reaction 20b corresponds to a reduction-oxidation reaction in which soot is the reducing agent. The primary oxidation product [soot.O] of reaction 20b has not been identified, but it releases CO and CO2 upon heating (reaction 19b), similar to the oxidation of soot by NO2 in reaction 19a. The heterogeneous reaction of the NOx reservoir HNO3 on soot also corresponds to a reduction-oxidation reaction analogous to the examples listed above (Rogaski et al., 1997):

2HNO3 + 2[soot] NO2, NO + 2[soot . O] + H2O(ads) and/or [soot . OH] (21)

Finally, ozone reacts heterogeneously with soot in a redox reaction in which soot is the reducing agent (Stephens et al., 1986; Fendel et al., 1995; Rogaski et al., 1997):

O3 + [soot] O2 + [soot . O] (22)
[soot.O] CO, CO2 (19b)

It is important to note that reaction 22 does not correspond to a catalytic reaction because for every ozone molecule reacted, CO and CO2 are formed according to reaction 19b. Therefore, soot is consumed as shown in the mass balance involving carbon (Stephens et al., 1986). Thus, in all of the above reactions-with the possible exception of reaction 19d-soot appears to act as a reducing agent that is consumed in the course of the reduction-oxidation reaction.

Heterogeneous reactions of other potentially important reservoir and active compounds on soot are virtually unexplored. Examples may include reactions of HNO4, H2O2, HOx, XNO2, and XONO2, with X=Cl, Br. HNO2 does not, however, seem to undergo heterogeneous interaction on soot to any significant extent (Gerecke et al., 1998).

Ice as a substrate in heterogeneous reactions has not received as much attention in laboratory studies as reactions occurring on other stratospherically relevant substrates such as supercooled and frozen sulfuric acid (H2SO4) hydrates, as well as type Ia and Ib polar stratospheric cloud (PSC) substrates. N2O5 undergoes a reactive uptake on tropospherically important substrates such as H2SO4/H2O (Hanson and Ravishankara, 1991; Fried et al., 1994) and water ice (Quinlan et al., 1990; Hanson and Ravishankara, 1992), which results in HNO3 formation. HNO3 undergoes non-reactive uptake only on the aforementioned substrates (i.e., condensation) (Hanson, 1992; IUPAC, 1997a; JPL, 1997). N2O5 also undergoes a bimolecular reaction with hydrogen chloride (HCl) on ice, resulting in ClNO2 in competition with the well-documented hydrolysis reaction leading to HNO3 (IUPAC, 1997a; JPL, 1997; Seisel et al., 1998):

N2O5 + HCl ClNO2 + HNO3 (23)

Reaction 23 occurs faster on ice than on H2SO4 because of limitations of HCl solubility in H2SO4. The increase in the rate of heterogeneous reactions on ice compared to acidic surfaces generally occurs for all reactions involving hydrohalic acids (IUPAC, 1997a; JPL, 1997). Reaction 23 competes with heterogeneous hydrolysis of N2O5, similar to reaction 20a. At temperatures near 200 K and at the limit of high HCl concentrations, 65% of the reaction proceeds via ClNO2 formation; the rest proceeds through N2O5 hydrolysis.
NO, NO2, and ozone do not interact with H2SO4 at concentrations encountered in the UT, or with water ice. Similarly, HONO does not interact with water ice down to temperatures of 180 K (Fenter and Rossi, 1996). However, HONO is taken up by H2SO4/H2O binary mixtures (weight of H2SO4 > 65%) from ambient temperatures down to 180 K and eventually yields nitrosylsulfuric acid (NSA) after protonation (Becker et al., 1996; Fenter and Rossi, 1996; Zhang et al., 1996; Kleffmann et al., 1998; Longfellow et al., 1998):

HONO + H2SO4 NO+HSO4- + H2O (24)

NSA has the ability to activate chlorine by reacting with HCl (reaction 25). The resulting NOCl product rapidly photolyzes in the UT into NO and Cl.

NO+HSO4- + HCl NOCl + H2SO4 (25)

An important finding regarding halogen activation by reactions 24 and 25 is the fact that these reactions occur approximately 20 times faster on pure ice, bypassing NSA as an intermediate species altogether because of its instability on ice (Fenter and Rossi, 1996). The propensity of many bimolecular heterogeneous reactions to occur more efficiently on ice surfaces compared to sulfuric acid aerosols is considered particularly relevant with regard to chemical processes occurring on natural cirrus clouds in the UT and on those seeded by particles emitted from aircraft engines, owing to a change in the reaction mechanism. The relevant water-ice mechanism corresponds to a simple ionic displacement of the form

Cl- + HONO NOCl + OH- (26)

where NOCl may undergo photolysis to Cl and NO. Although reactions 24-26 may not be important under volcanically quiescent periods (Longfellow et al., 1998), they may nevertheless play an important role in young contrails and perhaps in aged plumes, in view of the large number concentrations of volatile particles whose densities may be on the order of a few 104 particles cm-3 (Brasseur et al., 1998).

The chemical transformation of NO to N2O on acidic surfaces is too slow to be of significance in the present context (Wiesen et al., 1994). However, an efficient heterogeneous reaction involving NO and SO2 on solid and liquid aerosols could lead to the formation of N2O within the plume on a time scale of tens of minutes (Pires et al., 1996). In this chemistry, the interaction of HONO and SO2 on humid surfaces, solid or liquid, may lead to N2O, which also has been observed as a result of other laboratory experiments on heterogeneous chemistry of NO2-soot interactions (Smith et al., 1988).

Many of the other potentially important reservoir species-such as HNO3, HNO4, and H2O2-absorb onto sulfuric acid aerosols and ice particles but do not react with them (IUPAC, 1997a; JPL, 1997). These heterogeneous interactions are central to the removal of NOx and HOx species during denitrification and sedimentation events in the UT and LS.

The particle dimension of cirrus cloud crystals is on the order of 10 to 15 �m, which is approximately 10 times larger than PSC type IIs and 50 to 100 times larger than PSC type Is (Petzold and Schr�der, 1998). This result, together with the fact that the pressure is higher in the UT than in the LS, necessitates a downward correction for heterogeneous rates in cases in which the uptake coefficient (g) is larger than 5x102 because molecular diffusion of the gas toward the surface of the particle becomes rate limiting. This correction may be especially important for the case of a heterogeneous reaction on ice with an uptake coefficient exceeding 0.1 and may thus increase the uncertainty when extrapolating laboratory rate parameters to atmospheric concentrations. At present, most published models assume that the interfacial reaction is rate limiting and prescribe a rate constant of the form k=gA/4, where g is the uptake coefficient, is the average molecular velocity, and A is the surface area of the atmospheric particle per unit volume. Aerosols and Stratospheric Ozone Chemistry

It is now well-established that aerosols-consisting mainly of water, sulfate, and nitrate-play a crucial role in defining lower stratospheric levels of active NOx and ClOx (WMO-UNEP, 1995). Heterogeneous reactions in the stratosphere convert the inactive species HCl and chlorine nitrate (ClONO2) into the photolytically labile Cl2 species and the more active N2O5 species into the less active HNO3 species. The net effect of these processes is to decrease NOx levels and increase stratospheric levels of ClOx and HOx, hence the importance of ClOx- and HOx-catalyzed ozone loss relative to NOx-catalyzed ozone loss (refer to Figure 2-2). Increases in the total surface area of lower stratospheric aerosol by additional aerosol sources such as aircraft further accentuate the importance of ClOx- and HOx-catalyzed chemistry and reduce the impact of aircraft NOx emissions on ozone depletion. The net effect on ozone depletion will depend on the magnitude of the aerosol perturbation relative to the natural background and on given background levels of ClOx and NOx radicals, as well as on possible changes in the frequency of denitrification events induced by aircraft aerosols, water vapor, and HNO3. Aircraft-induced aerosols will be imbedded in a highly variable natural stratospheric aerosol background. In particular, stratospheric sulfate surface areas-typically on the order of 1 �m2 in volcanically quiescent periods-are nearly an order of magnitude larger than corresponding soot surface areas. During PSC type II events, surface areas of water-ice can approach 10 �m2. Likewise, amplification of H2SO4 aerosol surface area can be up to a factor of 100 following volcanic eruptions.

A number of heterogeneous reactions are important to various degrees on the surfaces of ice particles (PSC type II), HNO3/H2O aerosols (PSC type Ia), and saturated ternary solutions of water, HNO3, and H2SO4 (PSC type Ib). Rate parameters for heterogeneous reactions depend complexly on temperature, pressure, and humidity conditions. Consequently, modeling of stratospheric reactions must explicitly address differences in conditions between poles and mid-latitudes and between the tropopause and the LS. The most important heterogeneous reaction in relation to ozone depletion on a global scale is the heterogeneous hydrolysis of N2O5 (reaction 20) (Hofmann and Solomon, 1989; Solomon et al., 1996; Kotamarthi et al., 1997). This and the following reactions are likely to be enhanced in the presence of increased atmospheric particulates such as contrails and aircraft-induced cirrus clouds, which correspond to PSC type II in the LS from a chemical point of view.

Chlorine and bromine activation on PSC type Ia and Ib as well as on PSC type II (ice) takes place according to the following reactions:

ClONO2 + H2O(s) HOCl + HNO3 (27)
ClONO2 + HCl(s) Cl2 + HNO3 (28)
HOCl + HCl(s) Cl2 + H2O (29)
BrONO2 + H2O(s) HOBr + HNO3 (30)
BrONO2 + HCl(s) BrCl + HNO3 (31)
HOBr + HCl(s) BrCl + H2O (32)

Reactions of N2O5 with HCl(s) and HONO with HCl(s) (reactions 23 and 26) are also possible (IUPAC, 1997a; JPL, 1997). Because of the strong decrease in the solubility of HCl with increasing acidity of H2SO4 solutions-thus with increasing stratospheric temperature-reactions 28, 29, 31, and 32 will be most important at high latitudes (Portmann et al., 1996). Accordingly, these bimolecular reactions will proceed fastest on type II PSC surfaces in the polar vortex or on cirrus cloud particles in the tropopause region (Borrmann et al., 1996; Solomon et al., 1997). In addition, the hydrolysis reaction of ClONO2 (reaction 27) is less efficient than the corresponding one involving BrONO2 (reaction 30), on account of thermochemical differences between the reactions (Allanic et al., 1997; Barone et al., 1997; Oppliger et al., 1997). One consequence of this difference is that reaction 27 is prone to surface saturation on solid substrates, whereas reaction 30 is a strong and continuous source of HOBr, thereby making BrONO2 a relatively labile reservoir compound. Aircraft Aerosols and Heterogeneous Reactivity

Heterogeneous reactions are important in defining ozone removal rates in the stratosphere and troposphere. Stratospheric and tropospheric aerosol particles differ markedly in composition, lifetime, and size distribution. In the stratosphere, the global background aerosol consists predominantly of sulfate; water-ice and carbonaceous particles are minor components on the global scale, although PSCs that contain water-ice can be important on smaller geographic scales. Tropospheric aerosols consist of a wide variety of constituents, including water, sulfuric acid, soot, mineral dust, sea salt, and organic particles.

Soot, sulfuric acid ("sulfate"), and water-ice particles are the main condensed-phase species found in the exhaust of jet aircraft. As mentioned above, all of these species are present in the background atmosphere, and all derive from natural and other non-aircraft anthropogenic sources. Consequently, the effect of aircraft aerosol emissions on atmospheric photochemistry, to a first approximation, is to increase the soot, sulfate, and water-ice surface areas available for heterogenous and multiphase processes discussed in the previous subsections. Large-scale aerosol loading from aircraft can be estimated roughly from knowledge of fleet emission rates and average residence time of particles deposited at given altitudes. A detailed discussion of aircraft particle loading appears in Chapter 3. For the purposes of this chapter, it suffices to note that aircraft soot and sulfur emissions are thought to be significant at cruise altitudes, whereas estimates of the perturbation from aircraft H2O remain highly uncertain.

Beyond the simple aircraft aerosol loading approximation, chemistry occurring inside aircraft plumes carries the potential to produce volatile aerosols of significantly different heterogeneous reactivity relative to typical background aerosols. For instance, the water content of H2SO4 particles will vary over a wide range in an aircraft plume and wake, with concomitant effects on the rates of reactions 27-32. In addition, under cold lower stratospheric conditions, aircraft plume production of HNO3-rich liquid aerosol may propogate to the larger scale and contribute to the formation of solid PSCs, hence enhance processing of inactive chlorine species (K�rcher, 1997). Finally, evidence is mounting that volatile aerosols formed under conditions of low fuel sulfur content may contain significant amounts of light fuel-bound organic constituents (K�rcher et al., 1998b). The heterogeneous reactivity of such organic-containing aerosols is unknown and must await further in situ aerosol characterization.

Significant differences may also exist between aircraft-derived and ambient background soot particles. As discussed in Section, soot surfaces may act as reaction catalysts or consumables, depending on their properties. In addition, the surface properties likely determine the extent to which reduction-oxidation reactions occur on a particular soot particle. At present there is virtually no information on the properties of aircraft-derived or ambient background carbonaceous particles upon which to base an evaluation. Net Effects on Ozone

Based on our current understanding of heterogeneous chemistry, aircraft sulfate and water-ice particles will lower ozone concentrations in the UT and LS relative to what they would be if aircraft emissions contained only NOx. This process occurs because the particles remove the ozone precursors HOx and NOx in the UT and liberate ozone-destroying ClOx in the LS. The heterogeneous chemistry occurring on soot is much less well understood, so its role in atmospheric chemistry is much harder to define. However, because particle abundances of sulfate in the LS are much greater than those of soot, we can conclude with some confidence that present aircraft soot emissions are having little impact on stratospheric ozone, provided that their primary impact is on partitioning between NOy and NOx.

The extent to which aircraft aerosols offset the effects of aircraft NOx emissions on atmospheric ozone depends on a variety of chemical and dynamical factors. To quantify this balance, we will need atmospheric models that combine representations of aerosol microphysics, gas and heterogeneous chemistry, and atmospheric dynamics.

Other reports in this collection