Between Absorption And Vapor Pressure Of

Following the approach of Goss and Schwarzenbach (1998), for equilibrium between a species, i, in the gas phase at pressure Pt and dissolved in a liquid organic material at a mole fraction Xt om, the chemical potentials (ß) in the two phases must be equal, i.e.,

The chemical potential in the gas phase is given by

where P" is a standard-state pressure and , is the associated chemical potential of the standard state. Taking the standard state as the pure liquid, P° is the saturation vapor pressure over pure liquid (subcooled, where appropriate), i.e., P pL. Similarly, the chemical potential for the ith compound in the liquid organic layer is given by m M; RT(\nyiXl: ,,m), (GGG)

where the standard state is the pure liquid at 1 bar pressure. Combining Eqs. (EEE) through (GGG), one obtains

Goss and Schwarzenbach (1998) define a unitless gas-particle partitioning coefficient, Kn as the ratio of molar concentrations in the condensed, om, phase (Cj om) to that in the gas phase (c,- ). They show that using Eq. (HHH), Ki can be given by ci om RT

JíPlKT

where Vam is the molar volume of the liquid organic layer.

the slope becomes (1 s), where s reflects the simultaneous changes in In y, when pL changes. They conclude that the slope should lie between about 0.2 and 1.0 when the possible changes in the activity coefficients are taken into account. For example, the slope of In Kj defined by Eq. (Ill) as a function of In pL for the partitioning of a series of alkanes and alkylbenzenes between the gas phase and liquid octanol at 25°C is about 1; however, that for partitioning of chlorobenzenes into octanol is only 0.59. Goss and Schwarzenbach (1998) attribute this to the smaller attractive forces between the chlorobenzenes and octanol compared to those in the pure liquid chlorobenzenes.

Because the slopes of log Kp against log pL can be

1 for both absorption into a liquid layer and adsorption onto a solid, the slope alone cannot be used to differentiate the mechanism of gas-particle partitioning. However, a combination of the slope and the absolute values of Kp can be used to test the two mechanisms for cases where only van der Waals interactions occur, since can be calculated from Eq. (XX). For example, Goss and Schwarzenbach (1998) show that the calculated values of Kpds for a series of alkanes and PCBs partitioning into typical urban aerosols are much smaller than reported in the literature, indicating that absorption into a liquid phase must be important.

As is the case for adsorption, the trend in gas-solid partitioning coefficients for absorption can be used to probe for similarities and differences in interactions between the gaseous SOC and liquid on a molecular level. For example, Goss and Schwarzenbach (1998) point out that if two liquids and their interactions with SOC are the same, a plot of the gas-particle partitioning coefficients for a series of gases in one liquid, Ki t, against the analogous coefficients in the second liquid, Kt 2, should be a straight line with a slope of 1.0. If the interactions are different, however, this correlation will not hold. For example, such a log-log plot of the gas-solid partitioning coefficients for the uptake of a series of alkanes and ethanol in «-heptane and dibutyl ether shows that the correlation does not hold for ethanol. This indicates that the two liquids are not chemically similar in terms of their interactions with the SOC, which is not surprising in this case because of the strong interactions between ethanol and the ether. If the liquids were ambient air particles of unknown surface composition, such a plot would show that the

FIGURE 9.63 Plots of om-phase-normalized gas-particle partitioning constant log Kp om vs logarithm of the subcooled liquid vapor pressure, log pl , for a series of semivolatile PAHs partitioning on (•) dioctyl phthalate (DOP) or (a) secondary organic aerosol (SOA) from photooxidized gasoline vapor. PAHs are as follows: naphthalene, A; acenaphthalene, B; fluorene, C and C'; phenanthrene, D and D'; anthracene, E and E'; fluoranthene, F and F'; pyrene, G and G'; chrysene, H (adapted from Liang et al., 1997).

FIGURE 9.63 Plots of om-phase-normalized gas-particle partitioning constant log Kp om vs logarithm of the subcooled liquid vapor pressure, log pl , for a series of semivolatile PAHs partitioning on (•) dioctyl phthalate (DOP) or (a) secondary organic aerosol (SOA) from photooxidized gasoline vapor. PAHs are as follows: naphthalene, A; acenaphthalene, B; fluorene, C and C'; phenanthrene, D and D'; anthracene, E and E'; fluoranthene, F and F'; pyrene, G and G'; chrysene, H (adapted from Liang et al., 1997).

particles are quite different in their chemical properties and hence their effects on gas-particle partitioning.

In short, differentiating between adsorption on a solid and absorption into a liquid for partitioning of semivolatile compounds in the atmosphere is often difficult to do in an unambiguous manner. However, the use of a combination of approaches can help to differentiate these two mechanisms and, perhaps more important, give some insight into the mechanisms of interaction of the SOC with the condensed-phase material.

3. Octanol-Air Partitioning Coefficients

One problem with the use of pL as a key parameter in both adsorption and absorption is the difficulty in obtaining accurate values for pL for solid SOCs, since they are not experimentally accessible and must be estimated (e.g., see Finizio et al., f997, and references therein), in addition, as discussed in the preceding section with respect to absorption into a liquid phase, slopes of f for plots of log Kp against log pL are only expected if the activity coefficients, -y,, do not change along a series of compounds.

An alternate approach has been proposed that avoids the use of pL and introduces a ratio of activity coefficients rather than an absolute, single value. The advantage in terms of the activity coefficients is that the ratio is usually less sensitive to changes along a series of compounds than is a single value of the activity coefficient. in this approach, the parameter pL is replaced by a different descriptor of an SOC's volatility, its octanol-air partition coefficient, KOA (Finizio et al., 1997; Harner and Bidleman, 1998; Pankow, f998). This is defined for a given SOC partitioned between liquid octanol and air as

where C() and CA are the concentrations of the SOC in octanol and in air, respectively, both in units of mol m 3. A distinct advantage of the use of KOA compared to pL is that it can be directly measured (e.g., see Harner and Mackay, 1995; Harner and Bidleman, 1996) or can be estimated from knowledge of two well-established and widely used partition coefficients, Kim for octanol-water and KAW for air-water (e.g., see Mackay et al., 1992; and Baum, 1998).

The relationship between Kp and KOA can be readily developed based on thermodynamic principles (e.g., see Finizio et al., 1997); see Box 9.4.

Thus, the relationship between Kp and KOA by Finizio et al. (1997) leads to Eq. (PPP):

Was this article helpful?

0 0

Post a comment