Chemistry Reference
In-Depth Information
2.3 Mass Closure/Tracer Based Approaches
A further, from computational point of view, rather simple method is based on
general information about emission source profiles which is used to assort the
measured PM compounds into different source categories [
6
]. This method can be
applied, for instance, easily and rather accurately in case of marine aerosols which
mainly comprise sodium and chloride. Based on the measurements of one or both
compounds the fraction of PM10 generated by sea spray emissions can be computed
from the well-known sea water composition. Sodium can be considered as the more
suitable tracer for inland sites [
7
] as the marine aerosol is frequently subject to an
exchange of chloride for nitrate (reaction with gaseous nitric acid) during the
transport. Similarly, contributions by mineral dust from soil erosion may be assessed
measuring silicon and/or aluminium concentrations (e.g. [
8
]). Elemental carbon is
the most widely used indicator compound for traffic exhaust emissions, while
elements like Cu, Ba and Sb are suitable tracers for non-exhaust traffic emissions
from brake and clutch wear [
9
]. Organic carbonaceous matter in many cases reflects
both natural (secondary aerosols from natural VOC emissions) and anthropogenic
(e.g. wood burning) sources. A major difficulty with tracer-based approaches is to
find the correct relationships between the tracer compound and the total particle
mass assigned to the source or process. The multiplication factors to be used may
have significant spatial and temporal variability.
In addition, most chemical analyses used for the mass closure and tracer
approaches are made from bulk samples collected on filters. Consequently, particles
with different origin but similar chemical composition can hardly be distinguished.
To cope with such overlaps methods based on electron microscopy coupled with
X-ray spectroscopy have been developed [
10
,
11
].
2.4 Statistical Receptor Models
The problem mentioned before has led to the development of statistical receptor
models (Fig.
3
) which nowadays are the most widely used tools for PM source
apportionment. They can be applied even to a single site and need a time series of
PM mass concentrations and corresponding chemical composition data. Depending
on the method, these analyses are based only on the receptor data or additionally use
information on the chemical composition from the relevant emission sources (emis-
sion profiles).
In Germany, as well as generally in Europe, multivariate methods (PCA and
PMF) are the predominantly applied tools (see [
1
]). Their main advantage is that no
information about emission sources is needed, and sources or processes so far not
registered in an emission inventory may be detected. On the other hand, the
analytical effort to be invested is considerably higher than for the previous methods
since these models need “enough” data to disentangle the sources and source
