Escolar Documentos
Profissional Documentos
Cultura Documentos
1025-1037, 1986
Abstrsct--Forecasts of the impact of emissions changes on visibility use light scattering efficienc~es---the
change in the amount of light scattered with a change in mass of an aerosol constituent. This paper
demonstrates how fight scattering ef~ciencies depend on interactions between the aerosol constituents.
Calculations are presented for two model aerosols: NH,HSO4 droplets and a 'typical' urban aerosol. These
calculations demonstrate that the traditional means of predicting visibility impairment for increases in
atmospheric emissions is not generally appropriate. Different light scattering ef~ciencies are needed for
distinct meteorological conditions and aerosol compositions thus for typical high and low visibility events.
Key word index: Aerosols, light scattering, atmospheric optics, sulfate, visibility.
!. I N T R O D U C T I O N
(1)
is the usual starting point for statistical fitting procedures. Here, M i represents the fine particle mass (per
volume of air) of chemical species i. aj is the fitting
constant which is usually interpreted as the light
scattering efficiency of chemical species i. Indeed, if the
a i were true constants then a change in the mass of
species i, AM i, would produce a Change in Bsp given by
ABsp ffi aiAM i, so that a i ffi ABsp/AM i would correspond to a unique light scattering efficiency.
Recently, White (1986) has noted that regression
estimates of fight scattering efficiencies could be confounded by interactions between aerosol constituents.
The calculations to be presented in this paper are
intended to provide a realistic evaluation of the
sensitivity of light scattering efficienciu to interactions
between aerosol chemical coMtituents. The light scattered by model aerosols is calculated exactly. The light
scattering eff|ciencies of the individual chemical species
are evaluated by calculating Bsp before and after a
change in the mass of one species. The difference
between these calculations, ABsp, is divided by the
mass change, AMt, to give the light acattering effic-
1025
1026
CHIUb~HNES. SLOANE
x (vJWi)Mi(D)/M i dD]
(3)
EXTINCTION BUDGETS
f [3q
i H20
D)/2pa M(D)dD
= f[3Q(m,~,D)/2DPo'I(y/Yo)Mo(D)dD
(2)
for the illustrative case of fight scattered by homogencous, spherical particles of similar composition.
Here, (m,,t, D) is the scattering cross-section, ,l is the
wavelength of the fight, m is the index of refraction, and
D is the particle diameter. Q(m, lt, D) equals
4(m,~,D)/~D 2, the ratio of the scattering crosssection to the geometrical cross-section; that ratio is
the optical scattering efficiency (Kerker, 1969).
(3Q/2D) is the mattering cross-section per unit volume
of particles with diameter D. n(D)dD is the particle
number density, which is related to the mass size
distribution by n(D)dD = (6/p(D)xD3)M(D)dD.
M(D)dD is the mass of aerosol particles (per unit
volume of air) with diameters in the size interval dD. p
is the ambient aerosol density.
The total aerosol mass (including water) is generally
not measured. Therefore, Bsp is more usefully expressed in terms of the size distributions of the 'dry'
species (i.e. species other than water): n(D)dD
-, (6/po (D) nD~)Mo (D)dD. Here the subscript 'o' signifies 'dry' particulate mass and Vo (= xD~/6) refers to
the volume an ambient particle of diameter D would
have without a water component.
Equation (2) is easily generalized to describe aerosols in which all particlfs are not homogeneous or of
similar composition (Sloane, 1984). For example, if
aerosol particles of like size were of like chemical
composition, but particles differing in size possibly had
different chemical compositions, then Equation (2)
would be changed by making the density, refractive
index and growth factor (V/Vo) into functions of D,
the particle diameter.
An analogous expression can be obtained for the
mass size distribution differentiated by chemical
SpeCies"
(4)
,_-
f[3 (m.
(,;
M,(D)/Mj dD
+ ~j ~3/2D(v3wl)UQ(m,~-D)
-Q(m',A,D)] Mi(D)/Mj dD
where m' is the refractive index evaluated without
species j. In the likely event that the refractive index of
species j is not appreciably different from that of the
other major species, then the fight scattering efficiency
for species j is simply:
dD
fE30.(.o)/2DpMj(D)/Mj
] dD.
(5)
1027
+ i#jy f[3Q(m,W)/2Op3[(v/Vo)M(O)
HzO
- (V/V'o)M~(O)]/Mj dO
= aj ('traditional')
i)
4. H 2 0
- (V/V'o)Q(m',A,D)M~(D)]/Mj dD
f[3Q
(6)
+ ~ f[3Q(m,~,o)/20p~}
ij
H~O
aj = ~,o f(z/2op~)[(V/Vo)Q(m,).,D
) Mi(D )
- (V/V'o)Q(m',2,D)M~(D)]/Mj dO.
Interactions between the aerosol constituents contribute to the inaccuracy of the 'traditional' prediction.
The calculations presented in this paper will illustrate
the effect of two such interactions. A simple example of
the first is the decrease in particle volume after the
removal of one species for the case where the number
of aerosol particles and the ma_~sesof the other species
are not affected. In that case, the mass size distributions
of the remaining constituents shift to smaller sizes as
illustrated in Figure 1C. The overlap of the final mass
size distributions with 3Q/2D is smaller. The size
distribution could also be distorted if the species
eliminated had not been evenly apportioned within the
overall mass size distribution. This simple effect can
most easily be visualized for internally mixed (i.e.
multicomponent) aerosol particles, but it would also
occur in externally mixed aerosols (i.e. mixtures of
chemically pure particles) if the species mass change
were less than total and the size (as opposed to the
number) of chemically pure particles was affected.
A second type of interaction occurs when the
concentration of one constituent influences the concentration of another in the aerosol. An example is
presented by the affinity of water for water-soluble
salts. A change in the mass of a salt constituent could
affect the aerosol water content. Therefore, if the
formation processes for the other species were not
affected, the mass size distribution of those remaining
constituents would shift in response to the loss of the
salt species and associated water from each particle.
Some distortion of the size distributions would result if
the salt were not evenly apportioned within the overall
mass size distribution. The average index of refraction
of the aerosol could also shift noticeably.
This last type of interaction between aerosol constituents will be referred to as i/chemical interaction in
contrast to the more physical interaction described
CHILSTINE ~. SLOANE
1028
,.,~..,..,.--......
~.
3 Q(m,A,D) I 2D
Distl il)ullon
~,J'~;?:~111!) ii:;?71~?~i?:
,,,
.2
.4
.6
.8
1.2
Diameter (},m)
3Q(m,k,D) / 2D
DIstrlbutl0n
il . . . . . . . . .
I
.2
.I
.4
.6
Diameter ~m}
.8
,1.2
3 Q(m,k,D) /2D
/_
.1
.2
.4
.6
.8
L2
Diameter (~mb
rlg.
~,D)/2DPi]
3. DESCRIPTION
OF
SEMI-EMPIRICAL
PHYSICAL
MODEL
<i) <t>
- M,Ep o.
(M,)
(7)
Here, rh - relative humidity/100, E = water soluble
fraction of dry aerosol mass, Mw = molecular weight
of water, and P0 - dry aerosol density. The ambient
aerosol diameter, D, is taken as the Stokes diameter,
which is related to the aerodynamic diameter through
Dstok. = D,erody,~ ~/(S!/$2 p) where the Ss are the
usual slip correction factors (Friedlander, 1977) and p
is the ambient aerosol density. The Kelvin Effect is not
included in Equation (7) becauseit is not important for
particles in the size range which dominates visible light
scattering. The remaining parameters form the composite function, < i ) ( ~ ) / ( M , ) .
They are derived
from the thermodynamic relationship: rh = water activity [1 +in(s)/n(w)]-' where: n(s) = moles of dissolved solute, n(w) = moles of water and i = van't
Hoff factor, n(s) equals E~ Mo/W, with ~ = dissolved
fraction of soluble aerosol mass Mo is the dry aerosol
mass and W, is the average molecular weight of
dissolved aerosol constituents, n(w) equals (V
- Vo)p(w)/qMw; p(w) is the density of water, q is the
ratio of the partial molar volume of water in the
aerosol particle to the molar volume of water, q = 1 if
volume is conserved when water is absorbed by an
aerosol l~rticle. M , - , W,]q, so M, is sensitive to
deviations from the conservation of volume upon
1029
1030
CHRmTn~ S. StOANS
A~rosol type
30
40
H2SO,
NH,HSO4
(NH4)aSO4 '
NH,NO~
Binary salts*:
(0.5 (NH,),SO, and
0.5 NH,NO3)t
(0.5 (NH4hSO4 and
0.5 H=SO,)
Mixed aerosols :
Sea salt (E = 0.89)1"
1.60
1.40
0.61
, 1.10
0.29
1.20
0.48
0.71
0.23
1.00
0.45
0.53
0.37
~:
:~
0.61
0.04
80
90
0.80
0.46
0.39
0.31
0.66
0.40
0.32
0.32
0.52
0.35
0.27
0.33
0.32
0.33
0.35
0.48
0.45
0.46
0.40
0.35
0.06
0.09
0.13
0.83
0.71
0.57
0.24
0.24
0.30
0.54
0.67
0.58
0.46
0.18
0.18.
0.20
0.34
0.31
0.29
0.24
0.25
0.48
0.62
0.65
0.69
0.56
0.37
0.20
0.40
0.60
0.70
0.75
0.65
0.60
D'
(D/Do)XDo
where
L
J/L
.I
mass of a chemical species will be altered proportionally throughout that species' reference mass size distribution, so that in all size ranges AMi(D)/Mi(D)
= AMilM i.
3.3. lnhomooeneity
In this model, the ambient aerosol is taken to have
an interior core of the water-insoluble material and an
exterior water-soluble layer. The exterior layer is
presumed to take up/release water with changes in the
relative humidity. The refractive index and density of
each layer are represented by volume averages of the
refractive indices and densities, respectively, of the
chemical constituents of the layer (SIoane, 1983). They
are re-evaluated to be consistent with aerosol composition for calculations involving changes in the
masses of individual species. Values of the constituent
refractive indices and densities used in these calculations are given in Table 4.
The chemical composition of all particles in each size
interval is proportioned according to the mass size
distribution for each chemical constituent as given in
Section 3.2. Thus, the composition and refractive index
of the core layer vary with particle diameter. The
composition and refractive index of the exterior layer
1031
'
~ .
g.
e~
~" --"
..=
m
m
" ~
o
.=
~s
e4
.G
.o
"~'--
e~
t~s
o
.q
o
. ~
~k
,- ~,o..
~ ,,.~
1032
CnmsrmE S. St.oxw:
NH,,HSO,
Organic carbon
Elemental carbon
Residue
Index of refraction
Density
1.47"
1.55t
1.90-0.60i~
,1.53-0.005i H
1.78"
1.401"
2.00
2.3~
The light scattering efficiencies of individual chemical species have been evaluated by taking the difference
between Bsp calculated for a reference aerosol composition (Section 3.2.a) and Bsp calculated for an
aerosol which differs from the reference aerosol by the
removal (or addition) of mass for one or more chemical
constituents (Section 3.2.b). That difference in Bsp is
divided by the mass change.
Results of the calculations based on the 'typical'
urban aerosol model are presented in Table 5. Section
A ofTable 5 shows the calculated scattering efficiencies
for the sulfate, carbonaceous and residual aerosol
constituents given as functions of relative humidity
and percentage change in mass. For example, the light
scattering efficiency (change in light scattering per unit
mass change) for a 50 ~ increase in mass of sulfate
compound at 60 ~ relative humidity is 8.7 m a g- ~ for
this aerosol. The sulfate compound represented by the
dens/ty and index of refraction in Table4 is
NH4HSO,,, the sulfate species likely to have dominated the Detroit aerosol (Sloane and Wolff, 1985).
For comparison, the calculations were repeated for
the hypothetical case where the three constituents
(referred to as $1, $2 and $3 in Table 5B) differed only
in affinity for water and mass fraction. All three
constituents were assigned the same size distribution
(D e -- 0.49, In# s = 0.73), density (1.75 gec- 1) and refractive index (1.50). The solubilities of the three
constituents were taken to be 100, 50 and 0 ~ ,
respectively. And, as in the 'typical' urban aerosol the
average aerosol mass loading was apportioned as S1,
12.7 Sa, 11.8 and $3,6.1/~g m - 3 for a total fine particle
mass (FPM) of 30.6 pgm -3. The results of these
calculations are given in Table 5B. Values shown in
parentheses correspond to calculations in which the
5. DISCUSSION
1033
Mass
chan~
Aerosol con~ituent
S
TC
R
FPM
+ 50 ~
-50%
3.7
3.7
3.9
3.9
4.8
4.7
6.4
6.3
8.7
8.6
11.5
11.9
14.6
15.8
30
50
70
80
3.5
3.3
3.5
3.4
3.9
3.7
4.5
4.5
5.6
5.4
6.8
6.6
8.6
8.5
1.7
1.0
1.6
I.I
1.5
1.1
1.8
0.9
2.6
1.2
2.1
1.9
2.7
2.3
3.2
3~
3.3
3.1
3.8
3.6
4.9
4.6
6.2
5.9
7.9
7.9
9.9
1~4
~ aiMi/Bsp
i
1.1
1.1
1.1
1.1
1.1
1.1
1.1
B. Light scattering efficiencies for constituents differing in water affinity (m2 g-i).
AM~ = + 50% M i
Relative
humidity
Aerosol constituent**
St
S2
$3
FPM
30
4.4
(4.4)
4.4
4.2
(3.5)
4.4
4.0
(4.0)
4.5
4.3
(4.1)
4.5
50
6.8
(6.9)
6.8
5.4
(4.7)
6.9
4.2
(4.2)
6.9
5.8
(5.6)
6.9
70
12.4
(12.3)
12.3
8.5
(7.4)
12.2
4.3
{4.2)
12.1
9.1
(8.7)
12.1
Solubilities (~/o)
Sl 100
St 100
S t 100
$2 50
S~ 50
Sz 100
$3 0
$3 0
$3 100
+ 2 0 ~o
-20~
30
50
70
3.9
6.2
11.6
4,0
6.4
11.5
3.9
6.3
11.9
3.9
6.5
11.5
-99~
3.8
6.1
I 1.8
the g r m t ~ ' t affinity for water (in this case, the most
soluble) is the strongest function o f relative humidity.
The N H , H S O 4 efficiency is calculated to change from
3.9 m 2 g - t at 3 0 ~ rh to 8.6 m a g - l at 60~0 rh within
the composite urban aerosol, over a two-fold increase.
The effgiency o f the carbonaggous m a t , ~ t l ( 4 8 ~
water soluble) changed from 3.4 m a g - x at 30 ~ rh to
5.4 mZg - l at 6 0 ~ rh, a 5 9 ~ increase. The residual
material (10 ~o water soluble and with 2 0 ~ of its mass
in the optically efficient size ranse ) showed an,increase
in light scattering efficiency from 1.1 m a g - 1 at 30 ~ to
1.2 m s g - I at 60 ~ rh. Table 5B further establishes that
1034
CHtlSTINZ S. SLO^NE
-50%
-100%
a(-50,%) M/Bsp
20
30
40
50
60
70
80
3.7
3.7
5.6
6.0
6.9
8.8
11.3
3.3
3.3
5.1
5.4
6.4
8.2
10.6
1.1
1.1
1.1
1.1
1.1
1.1
1.1
16
15
14
13
12
11
10
9
8
765-
AM
*Thermodynamic
(Table 1) assumed.
growth function
MIlS
for NH,HSO,
4-
16-
/'
is14-
/I
121i -
/ 6
//
'-6-
3 I
20
,
40
RELATIVE
,
50
50
60
70
80
s
60
70
HUMIDITY
mluble man
/4/'
,
30
40
//
30
RELATIVE
/~
.:bg 13 -
l0
20
I
80
su.
IL
IH
HUMIDITY
Reduced
soh,d)le mrs
RL
It,l~ve Humiolty
RH
High RIIsUwHumidity
1035
1036
CtlaWnsE S. SLO^NE
6. S U M M A R Y
(4) The response of the light scattered by atmospheric aerosols to a change in the mass of an aerosol
constituent is not linear. Nonlinearities are generally
less than 10% of the light scattering efficiencies.
(5) Simple relationships (such as Bsp = ~ a i M i ( 1
i
AE 20:5-0
1037