By Doris Cadavid1 and Andreu Cabot1,2
The rusting, burning, or decomposition of solids in contact with air or liquids has been known since antiquity. In the 19th century, we understood these phenomena to be related to chemical reactions involving a rearrangement
of atoms. Since then, we have dreamed of
watching the underlying atomic reorganizations in real time. On page 303
of this issue, Sun et al. (1) used x-ray
scattering and molecular dynamics
simulations to follow with subnano-meter spatial resolution the three-dimensional (3D) compositional and
morphological evolution of colloidal
iron particles as they oxidize.
Solids can be modified by atomic
reorganization, replacement, subtraction, and addition. Atoms within
a solid can rearrange when supplied with enough energy to form or
change the crystal phase. They can
be replaced by other atoms through
galvanic (2) or ion-exchange reactions
(3) or subtracted through reduction or
etching processes. More commonly,
the composition of a solid is modified
through atomic addition, chemically
reacting with external atoms in gas,
liquid, or solid media or intercalating
them. A particularly familiar example
is the oxidation of metals in air. Oxygen molecules physically adsorb on
the metal surface and dissociate to
form chemisorbed oxygen atoms. These atoms and the surface metal atoms rearrange
to form the first oxide layer. Further oxidation requires the diffusion of metal or oxygen
or both through the already-formed oxide.
Such reactive diffusion processes not
only involve a change of composition and
atomic organization but potentially also a
morphology change. The oxide layer thick-
ness, its adherence, porosity, and final ge-
ometry are determined by the material
volume expansion with the atomic addition,
the elements’ reactivities, their diffusivities
through the oxide layer, and the initial ge-
ometry. Overall, an extended phenomenol-
ogy can be obtained (4).
For example, in a diffusion-limited reaction regime, when the oxygen diffuses faster
than the metal through the oxide layer,
the reaction takes place at the metal-oxide
interface, potentially forming highly compact layers. However, these same conditions
translate into highly porous oxide layers
with “sea urchin” morphology in surfaces
with a small radius of curvature, moderate
volume expansions, and fast growth rates
(4). In the same diffusion-limited reaction
regime, when the metal diffuses faster than
the oxygen, a net outward transport of material takes place, and voids are formed at
the metal-oxide interface through the so-called Kirkendall effect (5–7). These latter
conditions can be used to produce hollow
structures with dimensions up to several
hundreds of nanometers (8).
The real-time monitoring at the atomic
level of such reactive diffusion processes is
extremely challenging. Oxidation processes
are usually studied through the characterization of the material after the chemical
reaction. In situ monitoring can be realized
with x-ray diffraction and optical spectros-copies, but these techniques provide averaged structural and chemical information.
Near-atomic resolution can be reached with
in situ transmission electron microscopy,
but only in 2D projections.
Sun et al. monitored the 3D compositional and morphological evolution of iron
during oxidation with near atomic resolution (see the figure). This process had been
analyzed before with transmission electron
microscopy (9), but they described it with an
unprecedented level of detail by combining
synchrotron x-ray scattering and molecular
dynamics simulations. A key enabler was
the use of colloidal iron nanoparticles with
extremely narrow size, shape, and
composition distributions as the solid
model system. The level of detail in
their description of reactive-diffusion processes will allow not only a
qualitative understanding of the phenomenon but also quantification of
parameters such as reaction kinetics
and element diffusivities.
The method of Sun et al.
approaches the point where we can see,
in real-time, material restructuring
at the atomic scale. The resulting improved control of reactive-diffusion
processes could allow more precise
engineering of intermetallic compounds and doped materials, such
as carbon steel or phosphorus-doped
silicon, for instance. Additionally,
a deeper understanding of the processes could enable a more rational
design of protective materials with
enhanced corrosion resistance, as
well as catalysts with enhanced
activity, selectivity, and stability,
among other numerous applications
involving atomic diffusion. Particularly interesting would be to study the intercalation of lithium within electrodes of
lithium-ion batteries, as this process generally involves very large volume expansions
that limit the lifetimes of these devices. j
1. Y.Sun etal.,Science356, 303(2017).
2. X.Xia,Y. Wang, A.Ruditskiy,Y.Xia, Adv. Mater.25,6313
3. L.De Trizio, L.Manna, Chem. Rev. 116,10852(2016).
4. M.Ibáñez etal., Chem.Mater.23,3095(2011).
5. A.D.Smigelskas, E.O.Kirkendall, Trans.AIME 171,130(1947).
6. Y.Yin etal., Adv.Funct.Mater.16,1389(2006).
7. Y.Yin etal.,Science 304,711(2004).
8. A. Cabot et al., ACS Nano2, 1452 (2008).
9. A.Cabot etal., J.Am.Chem.Soc.129,10358(2007).
Oxidation at the atomic scale
X-ray scattering and molecular dynamics follow the rusting
of iron nanoparticles in three dimensions
1Catalonia Institute for Energy Research (IREC), 08930 Sant
Adrià de Besòs, Barcelona, Spain. 2ICREA, Pg. Lluís Companys
23, 08010 Barcelona, Spain. Email: email@example.com
reacts with the
(gray) to form
…and provide insight into
corrosion on a macroscopic
scale and how to prevent it.
21 APRIL 2017 • VOL 356 ISSUE 6335 245
Atomic oxidation visualized
Sun et al. used x-ray scattering and molecular dynamics simulations
to follow oxidation of iron colloidal particles.