On Composition, Morphology, and Size Distribution of Airborne Mineral Dust - Mineral Dust: A Key Player in the Earth System - page 23

transported dusts collected, for example, at Fuerteventura (Canary islands) are
characterized by much higher values (Bergametti et al. 1989 ).
Finally it is important to note that the calculation of the fraction of “pure” desert
dust in Asian aerosol samples shows that even during heavy dust-storm conditions,
the amount of mineral dust at maximum reaches about 80 % and that 20 % of the
aerosol concentration are still of non-crustal origin (Zhang et al. 2003b ;Shenetal.
2007 ). Comparable studies of northern African dust are not available at the moment.
Today, only a comparatively small number of studies is devoted to the (inorganic)
ionic composition of northern African (e.g., Viana et al. 2002 ; Formenti et al. 2003 ;
Alastuey et al. 2005 ; Müller et al. 2010 ; Paris et al. 2010 ) and eastern Asian (e.g.,
Arimoto et al. 2004 ; Cheng et al. 2005 ;Shenetal. 2009a ;Wangetal. 2012 ) mineral
dust. This data is mainly helpful to decipher the alteration and processing of mineral
dust and its mixing with different pollution aerosols. The obtained data is extremely
heterogeneous concerning absolute values and ratios between different ions and will
here not be considered further.

2.3 Individual-Particle Analysis

For several decades now, scanning electron microscopy (mainly coupled with
energy-dispersive X-ray microanalysis) sometimes accompanied by transmission
electron microscopy (TEM) was used to characterize individual particles of mineral
dust samples in more detail. However, only recently techniques become available
enabling the automated measurement of several 10,000 particles in a single study.
Particles are collected on different substrates by active (e.g., impactor) or passive
(e.g., dry and/or wet deposition) sampling. Subsequently, particles are analyzed for
their chemical elements (usually atomic number 6), their size (area, diameter), and

their shape (aspect ratio, shape factor). Based on net counts or after some correction
procedures (e.g., standardless ZAF correction), the chemical analyses of the parti-
cles are interpreted in terms of particle classes (e.g., Si rich, Ca(Mg) rich, S rich,
etc.) grossly comparable to mineralogical groups (e.g., quartz, calcite

dolomite,
sulfates, etc.) or are further evaluated by statistical treatments such as principal
component analysis (PCA) (e.g., Anderson et al. 1996 ). Automated individual-
particle analysis offers the possibility to determine the size-segregated composition
and shape of aerosols in more detail and is helpful for a better understanding of the
external and internal mixing state of mineral dust. Furthermore, from the obtained
data set, it is possible to derive an average complex refractive index (Kandler et al.
2007 ) which can be compared with direct optical measurements.
For northern African dust, a series of studies exist using automated individual-
particle analysis for a robust characterization of mineral dust samples (Reid et al.
2003 ;Morenoetal. 2006 ; Kandler et al. 2007 , 2009 , 2011a ; Chou et al. 2008 ;
Deboudt et al. 2010 ; Matsuki et al. 2010a , b ; Lieke et al. 2011 ; Scheuvens et al.
2011 ; Deboudt et al. 2012 ;Pósfaietal. 2013 ). Their results are summarized in
the following paragraphs, and some images of “typical” mineral dust particles

C