GasParticle Partitioning

1
Gas/Particle Partitioning
2
Why is gas/particle partitioning important?
Dispersion of Pollutants Introduced into the
Atmosphere as Determined by Residence
Times (Husar and Patterson, 1980)
3
Washout by rain
Washout consists of capture of gas and particle
phases Gases equilibrate with rain via Henrys
Law (reversible air-water exchange) Particles are
irreversibly captured For PCBs Wg 200-400,
Wp 104 105 For PAHs Wg 10-500, Wp 103
106
4
Atmospheric reactions
Many (most?) organic species are reactive in the
atmosphere via direct or indirect
photolysis Half-lives of PAHs Gas phase
minutes-hours Particle phase hours-days How
does this contradict figure 1? Health
effects Particles can deposit in lungs. What
role does chemical makeup of these particles play?
5
Background of gas/particle distribution

• Parameters affecting G/P distribution
• Chemicals properties (liquid VP, Koa)
• particle characteristics size distribution,
  carbon content, composition
• Atmospheric conditions T, RH

6
How do we measure G/P partitioning?
Traditional hi-vol air sampler uses a filter to
capture particle phase and then an adsorbent to
capture gas phase Filters GFF, QFF,
PTFE AdsorbentsPUF, polymer resin, tenax
7
Denuder
under laminar flow, gases diffuse faster, so
adsorb to surface best for hi conc and lo MW
gases (NOx, etc)
8
Impactor
MOUDI micro-orifice uniform deposit
impactor nozzle size and thus air velocity
changes in each chamber 8-10 stages 0.2 0.3 um
cutoff filter gets the rest
9
Cyclone
10
sampling artifacts?
Sorption of gases to filters or particles during
sampling Volatile losses from filter, particles,
impactor stagesespecially at high T, high flow
(I.e. large pressure drop Effect on partition
coefficient?
11
Theory adsorption vs. absorption
12
Particle phase concentration dynamics?
Adsorption (Pankow, 1987)
Absorption (Pankow, 1994)
Kp mass normalized partition coefficient Cg, Cp
gas and particle phase concentrations NS of
adsorption sites Ql, Qv enthalpy of desorption
and vaporization aTSP aerosol surface area
TSP total suspended particle mass/volume fom
fraction of organic matter MWom molecular wt of
NOM ?om activity coeff. of SOC in OM
13
If absorption is important, Koa may be a better
predictor of Kp
those previous equations used p as the descriptor
variable. we could also use Koa
Koa octanol-air partition coefficient fom
fraction of organic matter MWom, MWoct
molecular wt of NOM or octanol ?om , ?oct
activity coeff. of SOC in NOM or octanol roct
density of octanol
14
Adsorption
Junge, 1977 (based on linear Langmuir isotherm) A
limited number of sites, S, of which S1 are
occupied and S0 S S1 are unoccupied Rate of
evaporation k1S1 Rate of condensation
k2CgS0 At eqbm k1S1 k2CgS0 k2Cg (S- S1) S1/S
? and b k2/k1
S1(Cs)
(Cg)
15
Surface area of particles
? amount sorbed/total
c energetics term (strength of interaction)
? (cm2/cm3)
Curved lines are constant VP (mm Hg)
16
(No Transcript)
17
Temperature dependence
Yamasaki (1982)
semi-volatile compounds have significant conc in
both the gas and particle phases
18
Adsorption Pankow 1987
Kp mass normalized partition coefficient Cg, Cp
gas and particle phase concentrations NS of
adsorption sites Ql, Qv enthalpy of desorption
and vaporization aTSP aerosol surface area
Linearize by taking log of both sides
Pankow suggests that at eqbm, m -1,
assuming Ns does not vary within a compound
class Ql-Qv does not vary within a compound
class
19
Absorption Pankow 1994
TSP total suspended particle mass/volume fom
fraction of organic matter MWom molecular wt of
NOM ?om activity coeff. of SOC in OM
Linearize by taking log of both sides
Pankow suggests that at eqbm, m -1,
assuming ?om does not vary within a compound
class THUS Pankow suggest the slope should be 1
regardless of whether absorption or adsorption
dominates
20
What is observed in the field?
21
Mader and Pankow, 2002
Kp alone often gives slopes of -1
accounting for OC and EC usually gives best R2,
slopes still -1
normalizing Kp to om, OC, or EC improves R2, but
slopes still -1
22
difference between PAHs and PCBs in same air mass
sampled twiceJuly 22-23, 1994
Slopes Chicago PAH -0.734 PCB
-0.517 LM PAH -0.811 PCB -0.634
Simcik et al, 1998
Despite a transport time of 3.5 hrs and
reduction in concentration to 1/3, Kp and slopes
have not changed. Slopes not necessarily 1 at
eqbm (?)