Processing and Ageing in the Atmosphere - Mineral Dust: A Key Player in the Earth System - page 78

Fe takes the form of amorphous ferrihydrite nanoparticles, which are more soluble
than the Fe originally associated with dust particle surfaces (Shi et al. 2009 ). Acid-
processing reactions have formed the basis of several attempts to model changes in
Fe solubility in atmospheric dust, with a variety of iron-containing minerals being
used as models for dust (Meskhidze et al. 2003 ; Johnson et al. 2010 ;Ito 2013 ).
Photochemical redox changes (e.g. reduction of Fe (III) to the more soluble Fe (II)),
often enhanced by organic components, such as oxalate, have also been suggested
to enhance the solubility of Fe in the atmosphere (Erel et al. 1993 ; Pehkonen et al.
1993 ; Siefert et al. 1994 ).
Dust reactivity changes as a result of atmospheric transport have also been linked
to the physical consequences of winnowing. Baker and Jickells ( 2006 ) suggested
that decreasing particle size spectrum during transport led to increasing particle
surface area to volume ratios and hence to higher fractional solubility of iron
and other dust components. Journet et al. ( 2008 ) linked changes in mineralogy to
changes in overall solubility of Fe in dust and suggested that >96 % of soluble
Fe was derived from clay minerals, rather than Fe oxides. Cwiertny et al. ( 2008 )
also suggested in a study on authentic dust samples that clay minerals played an
important role in Fe solubility. Furthermore, it was also noted that Fe present as
Fe(II) in clays were particularly soluble.
The extent to which atmospheric processing enhances trace element solubility
is not very well defined. Measurements of solubility in dust actually at source
are very scarce and measurements on source soil materials (also few in number)
are probably not representative of suspended dust at source because abrasion and
fractionation during dust uplift significantly alter the dust with respect to the parent
soil (Bullard et al. 2004 ; Mackie et al. 2006 ). Nevertheless, analysis of atmospheric
dust collected over a broad range of dust concentrations appears to show a general
hyperbolic increase in Fe solubility as dust concentration (or total Fe concentration,
as a proxy for dust concentration) decreases (e.g. Chen and Siefert 2004 ; Baker
and Jickells 2006 ; Sedwick et al. 2007 ; Theodosi et al. 2010 ). Sholkovitz et al.
( 2012 ) compiled (

1,000) available measurements of iron solubility and total Fe
concentration from more than 20 studies of (marine) aerosols and showed that
this hyperbolic relationship is observed consistently on a global scale. In Fig. 4.3 ,
we replot the Sholkovitz data set (excluding samples extracted with strong acid
solutions, which have been shown to release a significantly higher fraction of Fe
from dust than the majority of methods used in the data set (Witt et al. 2010 )). It is
apparent from Fig. 4.3 that similar behaviour is exhibited under four, rather different,
conditions for the extraction of soluble Fe (ultrapure water, ammonium acetate,
formate and seawater leaching solutions) and that there is considerable scatter about
the general solubility - concentration trend within, and between, these methods.
Several authors have suggested that non-dust Fe might contribute to elevated Fe
solubility values, particularly at low dust concentrations (e.g. Guieu et al. 2005 ;
Sedwick et al. 2007 ; Sholkovitz et al. 2012 ). In Fig. 4.4 , we compare Fe solubility
for subsets of the data in Fig. 4.3 for total Fe concentration ranges of 30-100 ng TFe
m 3 and 300-1,000 ng TFe m 3 . These data appear to indicate that Fe solubility
increases as total Fe decreases, even when non-dust sources of Fe are unlikely to be