Corrosion Science 42 (2000) 1327±1336

Soft X-ray spectromicroscopy for in situ study

of corrosion
K.E. Kurtis a,*, W. Meyer-Ilse b, P.J.M. Monteiro c
a
School of Civil and Environmental Engineering, Georgia Institute of Technology, 790 Atlantic Dr.,
Atlanta, GA 30332, USA
b
Center for X-ray Optics, Lawrence Berkeley National Laboratory, 1 Cyclotron Rd. 2-400, Berkeley, CA
94720, USA
c
Department of Civil and Environmental Engineering, University of California at Berkeley, 725 Davis
Hall, Berkeley, CA 94720-1710, USA

Received 9 April 1999; accepted 20 October 1999

Abstract

Corrosion of ®lings from an ASTM Grade 60 steel reinforcing bar was examined in situ
using the soft X-ray transmission microscope XM-1, operated by the Center for X-ray
Optics and located at the Advanced Light Source at E.O. Lawrence Berkeley National
Laboratory. Iron distribution maps generated from X-ray images of a steel ®ling before and
after the addition of water show the change in elemental iron distribution in the sample.
The decreased presence of iron in the distribution map after the introduction of water
indicates the oxidation of iron Ð the anodic reaction in steel corrosion. These results
demonstrate the utility of transmission soft X-ray microscopy for in situ corrosion
studies. 7 2000 Elsevier Science Ltd. All rights reserved.

Keywords: Steel; Iron; Oxidation; Microscopy

1. Introduction

Both corrosion reactions and corrosion products are extremely sensitive to

* Corresponding author. Tel.: +1-404-385-0825; fax: +1-404-894-0211.

E-mail address: [email protected] (K.E. Kurtis).

0010-938X/00/$ - see front matter 7 2000 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 1 0 - 9 3 8 X ( 9 9 ) 0 0 1 4 7 - X
1328 K.E. Kurtis et al. / Corrosion Science 42 (2000) 1327±1336

environmental conditions. Changes in ionic concentrations and relative

humidity in the environment surrounding a corroding metal and changes in
moisture content of the corrosion product will a€ect the electrochemical
process. If these changes occur during examination of the reaction, the change
in environmental conditions may in¯uence the interpretation of the
observations. Because most existing high-resolution characterization techniques
require drying of the sample or examination in vacuum, in situ studies of
corrosion are challenging to perform. A high-resolution technique, which
allows the examination of corrosion and the formation reaction products over
time without drying, at normal pressures will promote understanding of the
reactions and provide further insight into means for mitigating or preventing
the damage by corrosion.
The impetus for this research was the need for high-resolution microscopy
technique for observation of the corrosion process over time. Only
transmission X-ray microscopy allows through-sample observation of reactions
in a wet environment at normal pressures with high spatial resolution. With
X-rays, examination is not limited to surface features; internal features can be
studied in samples by transmission X-ray microscopy. In addition, the
spectroscopy capabilities available in the soft X-ray range with transmission
soft X-ray microscopy are ideally suited for the examination of steel corrosion
in situ, as regions of elemental iron concentration can be identi®ed.
The research presented here represents the initial stages of research toward
in situ examination of steel corrosion in a highly alkaline environment, such
as the concrete pore solution. In highly alkaline (pH > 11.5) solutions, a thin
passive ®lm forms on steel, providing corrosion resistance even in environments
where corrosion is thermodynamically favorable. In concrete where the pore
solution pH lies between 12.5 and 13.5, corrosion of embedded reinforcing steel
can be considered to be negligible due to passivation in the high pH environment
[1]. However, carbonation decreases the pH, leading to depassivation, and
chloride ingress damages the passive layer, through mechanisms which are not
understood [2]. In reinforced concrete structures, depassivation of the steel
resulting from carbonation and/or chloride ion ingress can lead to the onset of
corrosion. With estimates for repair ranging from $1 to $3 trillion, corrosion of
reinforced concrete is a leading cause of infrastructure damage [3]. Improved
understanding of the mechanisms of passivation, depassivation, and corrosion
onset in concrete are integral for the development of e€ective design, construction,
and repair strategies. However, these reactions must be characterized in a hydrous
alkaline environment Ð a constraint that poses a great challenge.
To demonstrate the applicability of a new technique for examining
electrochemical reactions in situ, reaction of ®lings from a Grade 60 steel
reinforcing bar in water was examined using the soft X-ray transmission
microscope XM-1. Details about the X-ray microscopy technique for imaging and
spectromicroscopy of iron will be provided. X-ray images of the corrosion of steel
®lings in water will be presented and further applications of this technique will be
discussed.
 K.E. Kurtis et al. / Corrosion Science 42 (2000) 1327±1336 1329

2. Instrument

The Center for X-ray Optics (CXRO) built and operates a high-resolution
transmission soft X-ray microscope, XM-1. The microscope is an endstation of
beamline 6.1 located at the Advanced Light Source (ALS), a third generation
synchrotron facility, at the E.O. Lawrence Berkeley National Laboratory (LBNL)
in Berkeley, California. XM-1 uses intense light in the ultraviolet and soft X-ray
regimes generated by the ALS for microscopy and spectromicroscopy. Soft X-ray
wavelengths range between 1 and 50 nm, which is 50 to 100 times longer than
hard X-rays and 50 to 100 times shorter than visible light. XM-1 was originally
developed for imaging of wet biological specimens [4], but has recently found
applications in cement chemistry [5,6] and environmental science [7].
A condenser zone plate lens in the microscope focuses bending magnet
radiation supplied by the ALS onto a sample. Chromatic aberrations of the
condenser are used to select photons of the energy required, because zone plate
lenses are most e€ective with monochromatic radiation. These photons pass
through a 15 mm diameter pinhole onto the sample, which are located very close
together [8,9]. After passing through the sample, an objective zone plate lens
forms an enlarged image on an X-ray sensitive CCD detector, which transfers a
digital image of the sample to a computer. A schematic representation of the
optical layout of the instrument in depicted in Fig. 1. The optical path, including
the zone plates, pinhole and CCD camera is in vacuum, but the sample remains at
atmospheric pressure. The instrument operates optimally (achieving a resolution
of 43 nm) when imaging samples up to 10 mm thick in water [10], but has been
used to image samples 50 mm thick in water.
Interactions between X-ray photons and electrons in the sample produce
contrast which allows sample imaging by X-ray microscopy. The complex index of
refraction (Eq. (1)), depends in a characteristic way on the photon energy of the
X-rays and the electronic con®guration of the sample [11]. Beer's law describes the

Fig. 1. Schematic of the soft X-ray transmission microscope XM-1 optics.

1330 K.E. Kurtis et al. / Corrosion Science 42 (2000) 1327±1336

attenuation I=I0 † of an electromagnetic wave of wavelength, l, passing through a

sample of thickness, z, with absorption index, b, and refractive index decrement,
d,
I 4p
ˆ eb l z 1†
I0
The wave will experience a phase shift, j, of
2p
jˆ d
z 2†
l
While phase contrast is used to some extent, the photoelectric absorption is the
primary mechanism for producing contrast in images produced by high-resolution
X-ray microscopy [11]. For each element, absorption varies with the X-ray photon
energy. Absorption, or attenuation length, versus photon energy for water and
iron is shown in Fig. 2. In a wet sample, contrast depends primarily upon
disparity between absorption of the solid and liquid phases in the sample. Thus,
for imaging, X-ray photons should be selected to maximize contrast. For water
and iron, Fig. 2 shows contrast in images will be greatest with photon energies
between 350 and 543 eV (which corresponds to the K-shell binding energy for
oxygen), and with photon energies above 705 eV. At these energies, iron absorbs
nearly an order of magnitude more than water, indicating that soft X-ray
microscopy performed within these limits is well-suited for steel corrosion studies
in wet environments.
Near the ionization threshold for an element, the interaction between the X-ray
photons and the sample electrons is more complex. X-ray photons may be selected

Fig. 2. X-ray 1/e attenuation length of iron and water as a function of photon energy.
 K.E. Kurtis et al. / Corrosion Science 42 (2000) 1327±1336 1331

within a narrow region from both sides of an ionization threshold (or ``edge'') for
a particular element present within the sample. For instance, in Fig. 2, an
ionization threshold for iron is very close to 706 eV. To utilize the spectroscopic
capabilities of the soft X-ray microscope, an image is made with photons from the
more absorbing region, and a second image is produced using photons from the
less absorbing region. By comparing the two images, regions where the element of
interest is present can be identi®ed. When this technique is combined with high
spatial resolution, the technique is referred to as spectromicroscopy [11].

3. Experimental method

3.1. Materials

Steel ®lings were observed using soft X-rays of two di€erent photon energies on
either side of the iron edge to facilitate the observation of the iron oxidation
during corrosion in water. The ®lings were obtained by mechanically abrading an
ASTM Grade 60 steel reinforcement bar, obtained from McGrath Steel,
Emeryville, CA. The alloying composition of the steel by weight percent is
described in Table 1. In the table, the remainder of the composition is iron. The
®lings, which ranged in size from several microns to hundreds of microns, were
stored in acetone prior to examination by transmission soft X-ray microscopy.

3.2. Sample preparation

Steel ®lings were ®rst observed dry. The dry sample was prepared by dusting
the steel ®lings across a 120 nm thick silicon nitride (Si3N4) window glued to the
bottom half of the sample carrier (Fig. 3). The acetone was allowed to volatize
before imaging the steel.
After imaging the dry steel sample, a 2±3 ml drop of deionized water was placed
on the Si3N4 window containing the steel ®lings. A second 120 nm thick Si3N4
window which was positioned on the top half of the carrier was placed on top of
the bottom window and bottom half of the carrier, which held the steel ®lings and
water (Fig. 3). Three screws were tightened to seal the carrier. A rubber O-ring
inserted between the top and bottom halves of the carrier sealed the carrier edges,
decreasing the rate of evaporation. Polystyrene beads of 6 mm nominal diameter
were mixed with the water used in the sample preparation because of the thinness

Table 1
Alloying composition of the steel

C Mn P S Si Cr Ni Cu Mo V CEa

0.28 1.04 0.02 7 0.027 0.17 0.18 0.17 0.42 0.02 0.23 0.49

a
Carbon Equivalent, often used as an indication of weldability.
1332 K.E. Kurtis et al. / Corrosion Science 42 (2000) 1327±1336

and brittleness of the windows, and the surface tension generated in the wet
sample. The purpose of the beads was to separate the two Si3No4 windows,
preventing breakage of the windows.

3.3. Spectromicroscopy method

Kinematic mounts on the sample chamber ensure that the carrier is precisely
seated in the X-ray microscope. When regions of interest within the dry steel
sample were identi®ed, x, y, z coordinates for these locations were recorded. After
removing the dry steel sample and adding water to the sample chamber, the
previously identi®ed regions of interest can be revisited with accuracy. In this way,
changes in the sample which occur after the addition of water can be observed
and compared with images of the sample prior to water addition.
Initially, pairs of images were obtained of the steel ®lings using soft X-rays of
energy (or wavelength) on either side of the Fe L-III absorption edge. Images
were made at 697 eV (1.780 nm) which is just below the Fe L-III edge, and at 706
eV (1.757 nm) which is just on the Fe L-III edge. On the Fe L-III edge,
absorption by the iron is at a maximum, and just below the edge, absorption by
the iron is severely reduced. Images taken at 706 and 697 eV can be used to locate
elemental iron within a sample [12], because the ionization threshold for the L-III
electron on iron is located at 706 eV. By comparing the image taken on the Fe L-
III edge with the image taken just below the edge, a distribution ``map'' by mass/
area of the iron may be generated. Essentially, the images may be divided to show
where the elemental iron is located in the sample.
After adding the water to the sample, another pair of images were obtained
using the same wavelengths on either side of the iron edge. The sample was
imaged in a wet state at normal pressures. Again, a comparison of the image

Fig. 3. Schematic of the sample carrier used for wet preparation at XM-1.
 K.E. Kurtis et al. / Corrosion Science 42 (2000) 1327±1336 1333

obtained on the Fe L-III edge with the image obtained just below that edge will
reveal the distribution of iron in the wet sample.

4. Results and discussion

Fig. 4 shows four X-ray images of a steel ®ling in both an initial or ``dry'' state
and after the addition of water, termed ``wet'' state in the ®gure. As described in
``Spectromicroscopy Method'', use of kinematic mounts on the sample chamber
and an indexible coordinate system on the sample stage ensure that locations of
interest can be revisited. While the ``dry'' and ``wet'' images were made at the
same x, y locations in the sample, comparison of the images also reveals distinct
structural similarities, ensuring the sample location previously imaged ``dry'' is
indeed the same as the location imaged ``wet''.
In the dry sample, contrast is quite good both above (706 eV) and below (697
eV) the iron edge. The dry image, below the iron threshold, was obtained using a
beam current of 336.6 mA and an exposure time of 2.963 s. Above the iron
threshold, the image was obtained with a beam current of 333.5 mA with
exposure time of 2.987 s. In the wet sample, good contrast is apparent in the
image made with photon energy above the iron edge. The wet image at 706 eV
was obtained using a beam current of 255.5 mA and an exposure time of 116.761
s. However, contrast below the iron edge is reduced because X-ray absorption by
oxygen (present in the water) and iron is similar at that energy (Fig. 2). The wet
image at 697 eV was obtained using a beam current of 261.0 mA and an exposure
time of 114.197 s. Exposure times for wet samples are typically longer than those
for dry samples to allow sucient X-ray transmission for high-resolution imaging.
From these four images, two iron distribution maps have been generated Ð one
for the dry state and one for the wet state. The distribution maps are shown in
the lower portion of Fig. 4, and, in both maps, iron concentrations appear as
white to light gray regions.
The distribution map of the dry steel ®ling shows regions of elemental iron
concentration throughout the ®ling, as expected. Comparison of the map with the
dry image obtained below the iron edge at 697 eV, shows a small region of the
®ling which appears not to contain iron. It may be speculated that this region,
indicated by the label ``no iron'' in Fig. 4, may be rich in the alloying elements
present in the steel (Table 1).
The distribution map for the wet sample shows signi®cantly decreased presence
of iron as compared to the image of the ®ling before the addition of water. These
results indicate the occurrence of the anodic half-cell reaction, Fe 4 Fe2+ +2eÿ
whereby iron is oxidized. The sample preparation technique, described in Section
3.2, ensures a high water-to-solid ratio to maximize X-ray transmission. One e€ect
may be that the iron ions generated by the anodic reaction remain in solution or
form soluble products. It is also possible that the products were formed in regions
of the sample not under observation. While the introduction of water to the
sample may have dislodged portions of the ®ling leading to changes in the iron
 1334
K.E. Kurtis et al. / Corrosion Science 42 (2000) 1327±1336

Fig. 4. Iron distribution maps (bottom) generated from images of a steel ®ling dry and after the addition of water. The maps were digitally generated by
comparing a pair of images, with one made at the Fe-L-III absorption edge and the other made below the edge.
 K.E. Kurtis et al. / Corrosion Science 42 (2000) 1327±1336 1335

distribution map, no solid material was found in the regions surrounding the
particle to support this belief.
In both images of the wet sample, one of the 6 mm diameter polystyrene beads,
used to separate the two brittle Si3N4 windows, can be seen as a light region in
the upper right of the image. As is evidenced by the images, the polystyrene bead
is highly penetrable by X-rays, and its presence should not in¯uence image
interpretation.
Examination of iron oxidation in a wet sample by transmission soft X-ray
microscopy has demonstrated the potential of this microscopy technique for in
situ examination of electrochemical reactions. While the technique described here
can be applied to various systems, examination of corrosion in highly alkaline
environments is of particular interest to the authors. Potential extensions of the
research, relevant to corrosion of reinforcing steel in concrete include:
. examination of the in¯uence of pore solution, in particular the e€ect of
decreasing pH (such as during carbonation), on steel corrosion
. study of depassivation of steel in the presence of chloride ions, and
. characterization of new reinforcing steels designed for improved corrosion
resistance.
In addition, ongoing re®nements to XM-1 are expected to enhance the
spectromicroscopy capabilities. In the future, it is anticipated that di€erent
chemical states of iron may be distinguished. This development would be a
signi®cant advancement for in situ steel corrosion studies, as compositions of iron-
containing reactants and products could be more exactly identi®ed.

5. Conclusion

Iron distribution maps generated from X-ray images of a steel ®ling before and
after the addition of water show the change in elemental iron distribution in the
sample. The decreased presence of iron in the distribution map after the
introduction of water indicates the oxidation of iron Ð the anodic reaction in
steel corrosion. These results, obtained from experiments performed at the Center
for X-ray Optic's (CXRO) XM-1, demonstrate the utility of transmission soft
X-ray microscopy for in situ corrosion studies.

Acknowledgements

The authors are grateful for the insightful comments of Dr Greg Denbeaux of
the Center for X-ray Optics at Lawrence Berkeley National Laboratory. Research
at XM-1 is supported by the United States Department of Energy, Oce of Basic
Energy Sciences under contract DE-AC 03-76SF00098.
1336 K.E. Kurtis et al. / Corrosion Science 42 (2000) 1327±1336

References

[1] C.M. Hansson, Cem. Concr. Res 14 (1984) 574±584.

[2] K.E. Kurtis, P.K. Mehta, in: V. Malhotra (Ed.), Proc. CANMET/ACI International Conference
on Durabilty of Concrete, ACI, Farmington Hills, MI, 1997, pp. 535±549.
[3] M. Fickleborn, Materials and Structures 23 (1990) 317.
[4] C. Magowan, J.T. Brown, J. Liang, J. Heck, R.L. Coppel, N. Mohandas, W. Meyer-Ilse, Proc.
Natl. Acad. Sci 94 (1997) 6222±6227.
[5] K.E. Kurtis, P.J.M. Monteiro, J.T. Brown, W. Meyer-Ilse, Cem. Concr. Res 28 (1998) 411±421.
[6] K.E. Kurtis, P.J.M. Monteiro, J.T. Brown and W. Meyer-Ilse, J. Microscopy, 196 (1999) 288±
298.
[7] S.C.B. Myneni, T.K. Tokunaga, G.E. Brown Jr, Science 278 (1997) 1106±1109.
[8] W. Meyer-Ilse, H. Medecki, L. Jochum, E. Anderson, D. Attwood, C. Magowan, R. Balhorn, M.
Moronne, D. Rudolph, G. Schmahl, Synchrotron Radiation News 8 (1995) 29±33.
[9] W. Meyer-Ilse, H. Medecki, J.T. Brown, J. Heck, E. Anderson, C. Magowan, A. Stead, T. Ford,
R. Balhorn, C. Petersen, D.T. Attwood, in: J. Thieme, G. Schmahl, E. Umbach, D. Rudolph
(Eds.), X-ray Microscopy and Spectromicroscopy, Springer±Verlag, Heidelberg, Germany, 1998.
[10] J.M. Heck, W. Meyer-Ilse, J.T. Brown, E. Anderson, H. Medecki, D.T. Attwood, in: J. Thieme,
G. Schmahl, E. Umbach, D. Rudolph (Eds.), X-ray Microscopy and Spectromicroscopy,
Springer±Verlag, Heidelberg, Germany, 1998, pp. II130±II135.
[11] W. Meyer-Ilse, X-ray Microscopy and Microanalysis, Encyclopedia of Applied Physics, VCH
Publishers, 1999, pp. 324±341.
[12] J.C. Fuggle, N. Martensson, J. Electron Spectrosc. Relat. Phenom 21 (1980) 275.