ME TALLURGY
Direct observation of individual
hydrogen atoms at trapping sites in
a ferritic steel
Y.-S. Chen,1 D. Haley,1 S. S. A. Gerstl,2 A. J. London,1 F. Sweeney,3 R. A. Wepf,2,4
W. M. Rainforth,3 P. A. J. Bagot,1 M. P. Moody1
The design of atomic-scale microstructural traps to limit the diffusion of hydrogen is one
key strategy in the development of hydrogen-embrittlement–resistant materials. In the
case of bearing steels, an effective trapping mechanism may be the incorporation of finely
dispersed V-Mo-Nb carbides in a ferrite matrix. First, we charged a ferritic steel with
deuterium by means of electrolytic loading to achieve a high hydrogen concentration. We
then immobilized it in the microstructure with a cryogenic transfer protocol before atom
probe tomography (APT) analysis. Using APT, we show trapping of hydrogen within the
core of these carbides with quantitative composition profiles. Furthermore, with this
method the experiment can be feasibly replicated in any APT-equipped laboratory by using
a simple cold chain.
Hydrogen embrittlement, in which the pres- ence of hydrogen within a material’s mi- crostructure causes a severe loss in ductility, can lead to catastrophic and unpredictable failure of structural components in service
(1, 2). This is critical in many applications because hydrogen is a near-ubiquitous element and
can enter the material either at the time of manufacture or later during use. Hydrogen embrittlement affects many applications, such as fasteners
in aircraft and bolts used in the construction of
bridges. As such, it is a serious concern in many
marine and civil engineering applications (3, 4).
Hydrogen embrittlement is particularly important
in the area of advanced high-strength steels for
the automotive sector, in which the steel strength
now exceeds 1 GPa, resulting in a material much
more susceptible to hydrogen embrittlement. Such
hydrogen-embrittlement concerns can lead to the
selection of higher-cost or reduced-performance
alloys—particularly in terms of tensile strength.
Mitigating hydrogen embrittlement usually in-
volves either annealing to remove hydrogen that
has been introduced during manufacture or ap-
plying barrier coatings to minimize further ingress
of hydrogen from external sources (5). However,
a third option is to imbue materials with an
intrinsic resilience. Our poor understanding of
how hydrogen interacts with microstructure at
the atomic scale limits this option, despite gen-
eral agreement that the diffusion of hydrogen
within a microstructure is a key factor for hydro-
gen embrittlement (6). More generally, despite
the clearly serious nature of hydrogen embrittle-
ment, and a large body of macroscopic experimen-
tal and simulation research, there exists little in
the way of direct observation of hydrogen within
these materials. This is due to the experimental
difficulty of directly observing hydrogen within
technologically relevant microstructures (7). This
makes it challenging to engineer microstructures
that will mitigate the effect.
Practically, after low to moderate heat-treatment
schedules to remove hydrogen from within parts
after manufacture (8), protective barriers can re-
duce further ingress of hydrogen from environ-
mental sources during service. However, protective
barriers can fail in the presence of abrasive or
degrading environments, providing a pathway
for hydrogen and embrittlement (9). The incor-
poration of finely dispersed nanoscaled carbides
(10–12), such as ~10-nm vanadium carbides (VCs)
(13, 14), offers a suite of options for intrinsically
resistant steels. The carbides act as “traps” to re-
tain hydrogen that might otherwise diffuse through
the microstructure and promote embrittlement.
The traps may operate to limit apparent diffu-
sion rates (15) in a reversible manner, and/or they
may sequester hydrogen within their own micro-
structures (14). Again, little direct experimental
information exists on the interaction of hydro-
gen with such precipitates. Although some theo-
retical predictions do exist, they are limited to
cases that are straightforward to simulate, such as
via density functional theory (16). The uncertainty
in the effectiveness of hydrogen trapping by these
microstructures in turn leads to great uncertainty
in microstructural design.
Hydrogen trapping in bulk samples is studied
with several methods, such as small-angle neutron
scattering (17), thermal desorption spectroscopy
(15, 18), and energy recoil detection (19). These
methods provide only bulk-averaged information
on microstructural interaction, and the direct interpretation from these signals is difficult. Secondary
ion mass spectroscopy can directly image hydrogen (20–23) but cannot measure the precise location of hydrogen atoms, which limits applicability.
At smaller scales, transmission electron microscopy (TEM) can detect hydrogen in very specific
1196 17 MARCH 2017 • VOL 355 ISSUE 6330 sciencemag.org SCIENCE
1Department of Materials, Oxford University, 16 Parks Road,
Oxford OX1 3PH, UK. 2Scientific Center for Optical and
Electron Microscopy, ETH Zürich, Auguste-Piccard-Hof 1,
8093 Zürich, Switzerland. 3Department of Materials Science
and Engineering, Sheffield University, Western Bank,
Sheffield S10 2TN, UK. 4Centre for Microscopy and
Microanalysis, Faculty of Science, University of Queensland,
Brisbane, QLD 4072, Australia.
*Corresponding author. Email: daniel.haley@materials.ox.ac.uk
Fig. 1. As-received ferritic VC steel. 3D map of V in VC Fe-0.096C-1.6Mn-0.026Si-0.51Mo-0.25V-
0.05Al-0.056Nb (weight %) ferritic steel, as obtained via APT. Slice is 10 nm in thickness, to highlight
the V-containing carbides. (Inset) Bright-field TEM image for reference, showing “tracks” of carbides.
Composition “proximity histogram,” computed from a 25% V+Mo isosurface, shows segregation of
V, Mo, and C into the carbide phase. Segregated Nb is not shown in profile for visual clarity.
