Chemistry Reference
In-Depth Information
constituents measured in the ambient samples in unit emissions of all different
sources [
32
]. The required source profiles can be either obtained from
measurements at single emission sources or taken from literature [
33
,
34
]. In
contrast to CMB, multivariate statistical models only require qualitative or semi-
quantitative a posteriori information about the source emission profiles. Both,
source profiles and contributions of the sources are estimated based on the chemical
speciation of ambient particulate matter as measured at the receptor site. A widely
used receptor model is positive matrix factorization (PMF), [
35
,
36
]. Applications
of PMF for source apportionment of ambient particulate patter are, for example,
presented by Lee et al. [
37
] and Pandolfi et al. [
38
]. It is important to note that CMB
and multivariate receptor models rely on the assumptions that characteristic and
constant source profiles exist and that the mass of the considered chemical
constituents of PM is conserved during transport from the source to the receptor
site. Besides determination of source contributions to total PM mass concentration,
CMB and multivariate receptor models also allow for quantitative source
attribution of the considered chemical constituents such as organic carbon (OC)
and elemental carbon (EC).
PMF has successfully been applied to data from aerosol mass spectrometers
(AMS, Aerodyne Research Inc., Billerica) for identification of the main sources of
particulate organic matter, OM [
9
,
12
,
39
]. AMS instruments allow measurement of
the mass spectra of the non-refractory fraction of approximately PM1 with high
temporal resolution and determination of the concentration of particulate OM,
which can be converted to OC by multiplication by a conversion factor [
40
].
An alternative to the above described approaches is the radiocarbon method
that allows a distinction of contemporary carbon (from biogenic emissions and
combustion of biomass) and carbon from combustion of fossil fuels in particulate
carbonaceous matter [
15
,
41
,
42
]. In contrast to fossil fuels where the 14C isotope is
completely depleted, CM emitted from WB shows a contemporary radiocarbon
level. Radiocarbon measurements are often combined with measurements of com-
plementary source specific tracers (macro-tracer) for additional information of
source impacts [
14
,
43
,
44
].
The so-called aethalometer model (AE model, see [
12
,
45
-
47
]) is another method
for the distinction between light absorbing carbon (or black carbon, BC) from wood
burning and combustion of fossil fuels. The AE model relies on the stronger light
absorption of aerosol particles emitted from WB in the ultraviolet (UV) compared to
aerosol particles from fossil fuel (FF) combustion. Multi-wavelength aethalometers
such as model AE-31 from Magee Scientific Corporation (Berkeley) continuously
measure the aerosol
). With known
Angstrom exponents for aerosol particles from FF combustion and WB, assumption
of the Lambert-Beer law and the absorption coefficients
b
abs
(
light absorption at several wavelengths (
λ
λ
)measuredattwo
different wavelengths, the contribution of WB and FF to
b
abs
(
λ
) can be determined.
The BC mass concentration can then be obtained from
b
abs
(
λ
) divided by the mass-
specific aerosol light absorption cross section
σ
abs
is
calculated from EC analysis (e.g. thermal optical transmission method). In summary,
the AE model allows the determination of EC from wood burning.
σ
abs
.BCisconsistentwithECif
