Tropospheric Reservoir Molecules

6.7.1 H202, CHjOOH, and HONO

In the CO oxidation cycle, hydrogen peroxide, H202, is a reservoir molecule for HOt:

In the CH4 oxidation cycle, methyl hydroperoxide, CH3OOH, is also a HO* reservoir:

Nitrous acid, HONO, which is formed by a heterogeneous reaction involving NO? and H20 (Calvert et al. 1994), is a reservoir for both HO, and NO,. HONO dissociates by photolysis to regenerate OH atid NO:

Formed overnight, HONO photo dissociates upon sunrise to inject a pulse of OH into the early-morning atmosphere.

6.7.2 Peroxyacyi Nitrates (PANs)

The class of compounds of general formula RC(0)00N02 called peroxyacyi nitrates (PANs) was first discovered in the early 1950s as components of photochemical smog. The first two compounds in the series are

CH302 + H02 —♦ CH3OOH + 02 CH3OOH + hv —> CH3O + OH CH3OOH + OH —* HCHO + OH + H20

CH3COONO2 peroxy acetyl nitrate

CH3CH2C00N02 peroxypropionyl nitrate

Peroxyacetyl nitrate is the first compound in the series of PANs, which itself is usually called PAN. One route of formation of PAN is the OH reaction with acetaldehyde:


Since the second reaction is very fast, it can be combined with reaction 1:

As do other peroxy radicals, the peroxyacetyl radical, CH3C(0)02, reacts with NO:

CH3C(0)02 + NO ^ N02 + CH3C(0)0 CH3C(0)0 + 02 —► CH302 + C02


Again, we can incorporate the fast reaction into its precursor to give

The peroxyacetyl radical also reacts with N02 to form PAN

and PAN thermally decomposes back to its reactants by reaction 4.

Initially thought to be of importance only in polluted urban atmospheres, PAN has been identified as one of the most abundant reactive nitrogen-containing species in the troposphere (Roberts 1990). PAN mixing ratios ~100ppt are present in the Northern Hemisphere free troposphere, although its abundance is highly variable (Singh et al. 1995). Near the tropics, mixing ratios around lOppt are often prevalent.

PAN acts as a reservoir species for both CH3C(0)02 radicals and NO*. Because of this, the atmospheric lifetime of PAN is important; if its lifetime is relatively long, PAN can act as an effective reservoir for NO*. Potential atmospheric removal processes for PAN include thermal decomposition (reaction 4 above), UV photolysis, and OH reaction. PAN is not highly water-soluble; it is more soluble than NO or N02 but considerably less soluble than HN03. Thus, wet deposition is a minor removal process. Dry deposition is also unimportant. The PAN-OH rate constant is <3 x 10 14 cm3 molecule-1 s"1, and OH reaction is not an effective removal process. PAN absorbs UV radiation up to 350 nm (Libuda and Zabel 1995; Talukdar et al. 1995). Thus, thermal decomposition and photolysis are the principal removal processes for PAN.

Figure 6.8 shows the first-order loss rate of PAN as a function of altitude by thermal decomposition and photodissociation. The thermal decomposition rate coefficient for PAN can be obtained from the forward rate coefficient for reaction 3, and the equilibrium constant for reactions 3 and 4, KiA (Sander et al. 2003),

Atmospheric Chemistry

10"" 10" 10° PAN Loss Rate, s"1

FIGURE 6.8 Atmospheric loss rate of PAN as a function of altitude.

10"" 10" 10° PAN Loss Rate, s"1

FIGURE 6.8 Atmospheric loss rate of PAN as a function of altitude.


k0 = 9.7 x 10"2y (77300) koo = 9.3 x 10~12 (7/300) F = 0.6

cm molecule

The PAN absorption cross section has been determined empirically to be CTpan = 4 x 10 8 exp(—0.102X) cm2 molecule"1

where A, is in nm. Typical actinic fluxes in summer are given in Table 4.3.

From Figure 6.8 we see that at the Earth's surface (288 K) the lifetime of PAN against thermal decomposition is about 3h, whereas that against photodissociation is about 13 days. Because the photolytic loss of PAN is approximately independent of altitude and the rate of thermal decomposition is strongly temperature dependent, a point is reached, at about 7 km, where the two rates become equal; above that altitude, photolysis is the more important loss process. At the temperature of the upper troposphere, PAN is an effective reservoir for NO*; because PAN is transported in the upper troposphere, this amounts to a mechanism for long-range transport of NO*.

j NO

FIGURE PAN* chemical family.


Temporary Storage of NO, by PAN PAN acts as a reservoir species for both CH3C(0)02 radicals and N02. In order to calculate the lifetime for storage of NO* by PAN, one can consider the sum of CHjC(0)02 and CH3C(0)02N02 as a chemical family PAN (Figure 6.9). Let us calculate the lifetime of this chemical family, as compared to that for the thermal decomposition of PAN:

When dealing with chemical families, we generally assume that equilibrium is established within the family. We may not be certain that this is the case here because of the significant perturbation to this equilibrium caused by the CH-,C(0)02 + NO reaction. Since CH3C(0)02 itself is a free radical, its concentration should obey a steady-state relation. Thus wc ohtain

The lifetime of PAN* can be compared with the thermal decomposition lifetime of PAN itself in Figure 6.8. The lifetime of the PAN family is about 5 times as long as that of PAN itself for these conditions.

Acetone The first measurement of HO, radicals in the upper troposphere in 1994— 1996 revealed a more photochemically active region than had been expected (Brune et al. 1998; Wennberg et al. 1998). Observed HO, levels tended to be 2-4 times higher than that based simply on Oj photolysis to O('O) followed by reaction of O('D) with H20. One explanation for these observations is transport of photochemically active species from the lower to the upper troposphere. In addition, injection of air from the surface, carrying high levels of HO, reservoir species through deep convection is another source of HO,. Such species can include CH?OOH, H202, and aldehydes (principally HCHO). Along with these reservoir species, acetone (CH-,C(0)CH,) is a source of HO, in the upper troposphere. The 1997 SONEX aircraft campaign over the North Atlantic provided the first simultaneous measurements of HO„ H203. CH^OOH, HCHO, 03, H20, and acetone and allowed an assessment of the extent to which HO, precursors influence the HO,v chemistry of the upper troposphere (Brune et al. 1999; Faloona et al. 2000). The primary source of HO,, besides 0(1D) + H20, was found to be acetone photolysis.

Acetone, CH3C(0)CH3, is an ubiquitous atmospheric species having a mixing ratio of about 1 ppb in rural sites in a variety of locations {Singh et al. 1994, 1995). Under extremely clean conditions, ground-level background mixing ratios of 550 ppt have been found throughout the NH troposphere. In the free troposphere, acetone mixing ratios on the order of 500 ppt are present at northern midlatimdes, declining to about 200ppt at southern latitudes {Singh et al, 1995). From atmospheric data and three-dimensional photochemical models, a global acetone source of 40-60 Tg yr_l has been estimated, composed of 51 % secondary formation from the atmospheric oxidation of precursor hydrocarbons (principally propane, i so butane, and isobutene), 26% direct emission from biomass burning, 21% direct biogenic emissions, and 3% primary anthropogenic emissions (Singh et al, 1994). Atmospheric removal of acetone is estimated to result from photolysis (64%), OH reaction (24%), and deposition (12%). The average lifetime of acetone in the atmosphere is estimated to be 16 days (Singh et al, 1995).

By virtue of its photooxidation chemistry (see Section 6.11), acetone is a source of HOt radicals in the upper troposphere. Under the dry conditions of the upper troposphere, where 0('D)-l-H20 is relatively slow, acetone makes an important additional contribution to HOx. Photolysis of acetone (A < 360 nm) yields two H02 and two HCHO molecules {when [NO]»!N02|), and 30% of the HCHO molecules photoiyze via the radical-forming branch to yield two more H02 molecules. Thus the HO* yield from the photolysis of acetone is ~ 3.2 (the result of 2 + 4 x 0.3), as compared to a yield of 2 from the reaction of O('D) + H^O. A simple photochemical model calculation Tor the upper troposphere at the equinox {40°N, 11 km, 50 ppb 0;„ 90 ppm H20, and 0.5 ppb acetone) produced 24-h average HO, production rates of 7 x 103 molecules cm-3 s_1 from O('D) + H20 and 9 x 103 molecules cm-3 s~' from photolysis of acetone (Singh et al. 1995).


The hydroxy! radical is the key reactive species in the chemistry of ozone formation. The VOC-OH reaction initiates the oxidation sequence. There is a competition between VOCs and NO, for the OH radical. At a high ratio of VOC to NO, concentration, OH will react mainly with VOCs; at a low ratio, the NOr reaction can predominate. Hydroxy! reacts with VOC and N02 at an equal rate when the VOC: N02 concentration ratio is a certain value; this value depends on the particular VOC or mix of VOCs present, as the OH rate constants of VOCs differ for each VOC species.

At ambient conditions the second-order rate constant for the OH + N02 reaction is, in mixing ratio units, approximately 1.7 x 104 ppm-1 min-1. Considering an average urban mix of VOCs, an average VOC-OH rate constant, expressed on a per carbon atom basis, is about 3.1 x 103 ppmC-1 min "1. Using this value for an average VOC-OH rate constant, the ratio of the 0H-N02 to OH-VOC rate constants is about 5.5. Thus, when the VOC; N02 concentration ratio is approximately 5.5:1, with the VOC concentration expressed on a carbon atom basis, the rates of reaction of VOC and N02 with OH are equal. If the VOC: N02 ratio is less than 5.5 :1, reaction of OH with N02 predominates over reaction of OH with VOCs. The OH-NQ2 reaction removes OH radicals from the active VOC oxidation cycle, retarding the further production of 03. On the other hand, when the ratio exceeds 5.5 :1, OH reacts preferentially with VOCs. At a minimum, no new radicals are produced or destroyed; however, in actuality, photolysis of intermediate products generated by the OH-VOC reactions generates new radicals, accelerating 03 production.

Imagine starting with a given mixture of VOCs and NO*. Because OH reacts about 5.5 times more rapidly with N02 than with VOCs, NO* tends to be removed from the system faster than VOCs.4 In the absence of fresh NO* emissions, as the system reacts, NO* is depleted more rapidly than VOCs, and the instantaneous VOC: N02 ratio will increase with time. Eventually the concentration of NO* becomes sufficiently low as a result of the continual removal of NO* by the 0H-N02 reaction that OH reacts preferentially with VOCs to keep the ozone-forming cycle going. At very low NO* concentrations, peroxy radical-peroxy radical reactions begin to become important.

The essential role of NO* in ozone formation is evident in the CO oxidation mechanism (Section 6.4). For example, in the low NO* limit (NO*-limited), the rate of 03 formation increases linearly as [NO] increases and the rate is independent of [CO]. In the high NO* limit (NO*-saturated), the rate of 03 formation increases with [CO], but decreases as [NO*] increases. The explanation for the behavior in the high NO* limit is that, with ample NO* available, as NO* increases, the rate of the OH + N02 termination reaction increases, removing both HO* and NO* from the system, limiting OH - H02 cycling, and thereby decreasing the rate of 03 formation.

At a given level of VOC, there exists a NO* concentration at which a maximum amount of ozone is produced, an optimum VOC: NO* ratio. For ratios less than this optimum ratio, NO* increases lead to ozone decreases; conversely, for ratios larger than this optimum ratio, NO* increases lead to ozone increases.

6.8.2 Ozone Isopleth Plot

The dependence of 03 production on the initial amounts of VOC and NO* is frequently represented by means of an ozone isopleth diagram. Such a diagram is a contour plot of maximum 03 concentrations achieved as a function of initial VOC and NO* concentrations. The diagram is generated by contour plotting the predicted ozone maxima obtained from a large number of simulations with an atmospheric VOC/NO* chemical mechanism with varying initial concentrations of VOC and NO* while all other variables are constant.

Figure 6.10 is an ozone isopleth plot for Atlanta. To generate this plot, ozone formation was simulated in a hypothetical well-mixed box of air from the ground to the mixing height that is transported over an emissions grid from the region of most intense source emissions, the center city, to the downwind location of maximum ozone concentration. The mixing height rise throughout the day increases the volume of the cell, leading to dilution of the cell's contents. VOC and 03 in the air above the cell are entrained into the cell as the mixing height rises. Chemical transformations in the cell are described by a chemical reaction mechanism appropriate for the VOC mixture. The cell for Atlanta initially contained 600ppbC of anthropogenic controllable VOCs, 38ppbC of background, uncontrollable VOCs, and 100 ppb of NO*. The air above the cell was assumed to

4 The crossover VOC: NO2 ratio of 5.5 that we have been using in our discussion applies, more or less, to an average urban VOC mix. Because individual VOC-OH rate constants vary significantly, this ratio will also vary significantly if only a single VOC is present.

Ozone Isopleth Diagram

FIGURE 6.10 Ozone isopleth plot based on simulations of chemistry along air trajectories in Atlanta (Jeffries and Crouse 1990). Each isopleth is lOppb higher in 03 as one moves upward and to the right.

Initial VOC, ppbC

FIGURE 6.10 Ozone isopleth plot based on simulations of chemistry along air trajectories in Atlanta (Jeffries and Crouse 1990). Each isopleth is lOppb higher in 03 as one moves upward and to the right.

contain 20 ppbC VOC and 40 ppb of 03. The cell begins moving from center city at 0800 LDT and moves to the suburbs over a 14-h period. The initial mixing height was 250 m and the maximum mixing height was 1500 m. The chemical mechanism used for the simulation was the carbon bond IV mechanism (Gery et al. 1989). The peak ozone concentration predicted for these conditions was 114.6 ppb. We describe moving box models of this type in Chapter 25.

The base case described above corresponds to the dot on the line at initial NO* equal to 100 ppb in Figure 6.10. The full plot is generated by systematically varying the initial VOC and NO* concentrations and running the same scenario. The ozone ridge in Figure 6.10 separates the regions of low VOC: NO* ratio, above the ridge line, and high VOC: NO* ratio, below the ridge line. In Figure 6.10, the maximum 03 concentration is not zero at 0.0 ppmC initial VOC because the scenario included 38ppbC surface background VOC and 20ppbC VOC aloft.

The ozone ridge line identifies the maximum 03 concentration that can be achieved at a given VOC level, allowing the NO* level to vary. If the NO* level is either increased or decreased relative to that at the ridge line, the maximum amount of 03 produced at that VOC level will decrease. The region of the ozone isopleth plot below the ridge line is what we have denoted as "NO,v-limited"; that above the ridge Sine as "NO* saturated." It is apparent that above the ridge line a reduction in NO* can lead to an increase in maximum O3, Also, below the ridge line at low NO* levels there is a region where large reductions in organics have almost no effect on maximum 03. The ridge line defines the combination of initial VOC and NO* at which all the NO* is converted into nitrogen-containing products by the end of the simulation so that there is no NO left to participate in NO-to-N02 conversions nor any N02 left to photolyze. From a point on the ridge line if more VOC is added to the initial mixture to move into the NO*-limited regime the consumption of NO* occurs sooner than at the ridge line. Because of the increasing predominance of peroxy radicals relative to NO*, termination reactions are not favored relative to propagation and the amount of Os increases.

To explain the NO*-saturated region imagine a vertical line at constant initial VOC of 600ppbC. The maximum 03 at this initial VOC is about 142ppb at SOppb NO*. At low initial NO*, the average number of cycles an OH makes before termination is low as there is insufficient NO to convert all the H02 into OH and radical-radical reactions lead to termination. As NO* increases to the ridge line and beyond, the average number of OH cycles first increases and then begins to decrease because of increased termination from OH + N02 rather than OH propagation occurring by OH + VOC. Peak OH cycling occurs just about at the ridge line. The amount of 03 produced per OH-VOC reaction increases as NO* increases, since the increased NO is more competitive for R02 radicals and therefore produces more N02. Above the ridge line, as initial NO* is increased, the absence of sufficient radicals, however, makes the system increasingly unable to recycle NO. Also, fresh OH production from 03 photolysis is reduced because 03 appears later in the process and at a lower concentration. Thus increased OH termination as initial NO* is increased above the ridge line and reduction in fresh OH production together result in less production of 03.

Relation of Oa to NOy From the definition of the ozone production efficiency, the signature of an NO* molecule lost is the appearance of a number of 03 molecules, the specific number depending on atmospheric conditions and the HC and NO* levels. Thus, the O3 concentration attained in an airmass should be correlated with the quantity [NOy]-[NOiV], which is the total concentration of products of NO* oxidation (HN03, PAN, etc.). That this correlation should exist in airmasses was first pointed out by Trainer (1991, 1993), and it has been subsequently pursued in numerous studies [e.g., Kleinman (1994, 1997), Carpenter et a!. (2000)]. To obtain a good correlation between [03] and [NO,.]-[NO*], 03 production must have occurred within a day or so in the airmass, before significant removal of NOv can take place, for example by wet and dry deposition of HNOj.

Kleinman et al. (1994) suggested the linear least-squares correlation

(all concentrations in ppb) for the southeastern United States. The intercept of 27 ppb can be interpreted as an eastern North American background 03 level so that the NOv associated with that 03 has since been removed, leaving behind the longer-lived 03. Equation (6.32) suggests that the OPE for air in the southeastern United States was, at the time period on which the correlation is based, about 11. In general, the more remote the airmass, the lower the NO* level, and the less

FIGURE 6.11 Comparison of ozone production efficiencies on high 03 days in Phoenix, Arizona and Houston, Texas, (Courtesy Larry J. Kleinman, Brookhaven National Laboratory.)

important the NO* removal reactions. In this case, each NO* molecule is a more effective producer of O3, giving a larger slope of the correlation between [03 J and [NO,] = [NOy] - [NO,-j.

The overall linear relation between [03] and [NO,,] - [NO*] has proved to be an effective way to compare the ozone-forming potential of different ai mi asses. Figure 6.11 shows [03] + [N02] versus [NO>]-[NO*] for Houston, Texas and Phoenix, Arizona on specific dates in 2000 and 1998, respectively, when high 03 concentrations were achieved in both cities. The observed maximum in 03 in Houston (194 ppb) was more than double that observed in Phoenix (93ppb). The slopes of the straight-line fits to the data indicate that the number of 03 molecules formed per molecule of NO* converted to oxidation products in Houston is more than twice that in Phoenix. (In each city about 20ppb of NO* was consumed in forming 03.) The higher 03 production efficiency in Houston is a result of both high humidity [high rate of O('D) + H20] and a high overall radical production rate, arising in part from HCHO produced from alkene oxidation.


The chemistry of the background troposphere is fueled largely by the oxidation ofCH4and CO. Still, even the most pristine regions of the troposphere contain a variety of other volatile organic compounds (VOCs), arising from both biogenic and human emissions. After this section we will consider the chemistry of individual classes of organics. The OH radical is the dominant species oxidizing VOCs, although for alkenes 03 is also an important oxidant.

By analogy to CH4, reaction of OH with many hydrocarbons (RH) leads to alkyl peroxy radicals (R02):

RH + OH —>R + HsO ) n, r + O2 + M-+RO2 + M RH + OH^RO2 + H2O

Fait J

Alkyl Peroxy Radical

FIGURE 6.11 Comparison of ozone production efficiencies on high 03 days in Phoenix, Arizona and Houston, Texas, (Courtesy Larry J. Kleinman, Brookhaven National Laboratory.)

The alkyl peroxy radical reacts with NO:


The second reaction increases in importance monotonically as the size of R increases. The alkoxy radical (RO) reacts rapidly with 02:


This reaction is the sole fate of small alkoxy radicals. Larger alkoxy radicals can undergo other reactions; these need not be considered at this point. The correspondence of the RO + 02 reaction to the CH3O + 02 reaction is evident, where HCHO is the carbonyl product of the latter reaction. The higher carbonyl R'CHO can, in general, continue to be oxidized. The H02 radical reacts with NO to regenerate OH

and the main HOr-HOt and HO,-NO, termination reactions are

H02 + HO2 —« H2O2 + 02 R02 + H02 —> ROOH + 02

The resulting generalized mechanism is summarized in Table 6.3, where we have given representative rate coefficients.

TABLE 6.3 Generalized VOC/NOx Mechanism


Rate Constant (298 K)




ro2 + h2o

26.3 x 10-|2a


R02 + no


N02 + R'CHO + H02

7.7 x 10-12fc


H02 + no


N02 + OH

8.1 x 10-12


OH + N02



1.1 x 10"" (at 1 atm)


ho2 + ho2

h2o2 + o2

2.9 x 10-12


ro2 + ho2

rooh + 02

5.2 x 10"12c


N02 + hv

no + o3

Depends on light intensity^


03 +NO


no2 + 02

1.9 x 10-14

"Rate coefficient for propene (Table B.4). Other reactions consider R equal to CH3. Propene is selected because it is a relatively important constituent of the urban atmosphere. Even though OH-propene reaction proceeds by

OH addition to the double bond of propene (Section 6.10.2), the net result after 02 attack on the initial radical formed is a peroxy radical.

fcRate coefficient for CH302 + NO.

'Rate coefficient for CH302 + H02.

^Typical photolysis rate coefficient for N02 is y'No2 = 0.015 s_1.

"Rate coefficient for propene (Table B.4). Other reactions consider R equal to CH3. Propene is selected because it is a relatively important constituent of the urban atmosphere. Even though OH-propene reaction proceeds by

OH addition to the double bond of propene (Section 6.10.2), the net result after 02 attack on the initial radical formed is a peroxy radical.

fcRate coefficient for CH302 + NO.

'Rate coefficient for CH302 + H02.

^Typical photolysis rate coefficient for N02 is y'No2 = 0.015 s_1.

Let us integrate the rate equations for the mechanism in Table 6.3 for a variety of initial RH and NO, concentrations for a 10-h period to see how the amount of 03 formed and its rate of formation depend on the amount of NO, and the initial ratio of RH to NO,. We assume that the source of OH is constant at Pho,. = 0.1 ppt s 1. If HO, = OH + HO2+ RO2, we can assume that HO, is in steady state, in which Pho, is balanced by reactions 4, 5, and 6. It can also be assumed that the RO2 steady state is the result of a balance between reactions 1 and 2, and that HO2 steady state is the result of a balance between reactions 2 and 3. We will assume that initially [N0]/[N02] = 2 and that the initial concentration of 03 is that from the photostationary state corresponding to the initial concentrations of NO and N02 and the value of j'No2 given in Table 6.3. With the aid of the steady-state expressions for [OH], [H02], and [R02], the concentrations of these radical species can be expressed as functions of [RH], [NO], and [NO2]. We assume that over the timescale of the simulation, reactions of the products, R'CHO, H202, and ROOH, can be neglected. Therefore, one needs to solve the four differential equations for [RH], [NO], [N02], and [03], (Ordinarily, one would use the photostationary state relation (6.6) to determine the 03 concentration. However, in a rapidly reacting system like this one where the HO2 + NO and RO2 + NO reactions are not necessarily small relative to 03 + NO, (6.6) does not hold exactly. As a result, one needs to integrate the 03 rate equation explicitly.)

Figure 6.12 shows isopleths of the maximum 03 mixing ratio achieved over a 10-h period generated from the mechanism in Table 6.3. We see that the simple mechanism in Table 6.3 is able to produce the characteristic features of the ozone isopleth plot in Figure 6.10, based on a much more complex mechanism.

Phox= 0.1 ppts"1

Phox= 0.1 ppts"1

Ozone Isopleths Plot

RH, ppb

FIGURE 6.12 Isopleths of maximum O3 mixing ratio achieved over a 10-h period by integrating the rate equations arising from the mechanism in Table 6.3.

RH, ppb

FIGURE 6.12 Isopleths of maximum O3 mixing ratio achieved over a 10-h period by integrating the rate equations arising from the mechanism in Table 6.3.

Was this article helpful?

0 0
How to Improve Your Memory

How to Improve Your Memory

Stop Forgetting and Start Remembering...Improve Your Memory In No Time! Don't waste your time and money on fancy tactics and overpriced

Get My Free Ebook


  • Rose
    What are reservoirs in atmospheric chemistry ?
    6 years ago
  • ricky
    What are reservoirs in tropospheric chemistry?
    6 years ago
  • sofia
    What are reservoir molecules ?
    6 years ago

Post a comment