TEM investigation of intergranular stress corrosion cracking for
316 stainless steel in PWR environment
Y.Z. Huang*, J.M. Titchmarsh
Department of Materials, Oxford University, Parks Road, Oxford, OX1 3PH, United Kingdom
Received 29 July 2005; received in revised form 25 September 2005; accepted 5 October 2005
Available online 28 November 2005
Type 316 stainless steel foils containing stress corrosion cracks grown in high temperature aqueous environments have been examined
by transmission electron microscopy. It was found that the crack tips are oxidized and have a three-layered morphology where all the
layers taper towards the crack tip. The inner layer is a microcrystalline spinel sandwiched between the outer layers of a nano-crystalline
oxide. The outer layers are enriched in Cr, and the inner with Fe, relative to the matrix. Cu was observed to segregate at the interface
between oxide and matrix at one crack in type 316 steel. The inner oxide growth is dominated by different mechanisms before and after
the grain boundary cracks.
? 2005 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
Keywords: Stainless steel; Stress corrosion cracking; Crack tip; Grain boundary; Transmission electron microscopy
The investigation of intergranular stress corrosion
cracking (IGSCC) of austenitic stainless steels is of interest
since it is one of their most prominent failure modes, affect-
ing their application in the electrical utility industry and
thus having significant economic implications.
The mechanism of IGSCC is rather complex and still
remains unclear, even though a number of investigations
have been conducted to date [1–3]. It has been reported
that the chemical and microstructural differences of the
grain boundary compared to the matrix material are
responsible for intergranular cracking when stainless steels
are subjected to a certain environment, such as high tem-
perature water [4,5], which occurs in pressurized water
reactors (PWR). For example, the chromium depletion
zone formed in the vicinity of a grain boundary, as a result
of sensitization during thermal exposure or aging treat-
ment, is prone to cause intergranular attack on stainless
steels. The segregation of impurities such as phosphorus,
sulfur, or silicon at the grain boundaries has emerged as
another important factor capable of facilitating intergran-
ular attack, even in the absence of chromium depletion .
Apart from chemical redistribution at the grain bound-
aries, oxides/corrosion products formed along grain
boundaries also play an important role in the occurrence
of IGSCC. The presence of significant amounts of oxide/
corrosion products in cracks or crevices has been reported
in studies of metal components exposed to high tempera-
ture water , where the crack flanks were found to be
strongly corroded compared, for instance, with the surface
. Corrosion products, acting as crack initiation sites, can
promote crack propagation and thus cause pronounced
grain boundary attack in front of the crack tip . Andre-
sen et al.  demonstrated that intergranular cracking
occurs by mechanical failure of oxide particles that create
electrochemical crevices and stress concentrators, from
which intergranular cracks can initiate and propagate.
In this paper, we investigate at a high spatial resolution
the corrosion products along cracks, in terms of their struc-
ture, morphology and chemistry, since these characteristics
of corrosion products can be used to identify local corro-
sion and electrochemical processes promoting IGSCC in
1359-6454/$30.00 ? 2005 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
*Corresponding author. Tel.: +44 1865 273660; fax: +44 1865 283333.
E-mail address: email@example.com (Y.Z. Huang).
Acta Materialia 54 (2006) 635–641
of oxidation along grain boundaries. In principle, solute
diffusion occurs in the solid state through interactions be-
tween lattice defects and specific solute species. The diffu-
sivities of solute species are varied due to the fact that
their values of migration enthalpy are not the same and
are dependent on many factors, i.e., temperature, composi-
tions, aging treatment. According to the inverse Kirkendall
mechanism, the relative rates of solute diffusion are caused
by vacancy exchange. According to the diffusion rates of
the metal ions in the crystal lattice, Cr atoms move more
slowly than Ni atoms but are unable to diffuse as fast as
Fe atoms. As a consequence, Cr atoms are retained and
thus enriched in the NLs with Fe rich in the PL. It is re-
ported that microstructural evolution with time gives rise
to the variation in the solute diffusion rates [19,20], that
is, the diffusion rate is dependent on the morphological
structure. In terms of the nano-crystalline structure with
smaller lattice defects and vacancies, it may act as a barrier
to preferentially suppress the further diffusion of Cr atoms
onwards to the PL but not have much effect on the diffu-
sion of Ni atoms to penetrate the NLs; however, there is
not enough data available to calculate the solute diffusion
rate within the nano-structure so far.
4.2.2. Oxide growth along the crack flank (cracked grain
The asymmetry of the majority of the diffusion profiles
suggests that the growth of both PL and NL is simulta-
neous and correlated each other. From this point of view,
it is reasonable to assume that the growth process of oxides
is a competitive event with the PL growing faster than the
NL on the cracked grain boundary. As a matter of fact,
with the onset of oxide growth from virgin grain bound-
aries, the essential factor controlling the oxide growth, is
the relative diffusion of solutes under the solid-state growth
mechanism. However, in the case of oxide growth after
grain boundaries are cracked and open to the external envi-
ronment, the PL growth becomes clearly dominant (simply
compare the width for both layers) and relies on both the
metal dissolution and oxide precipitation mechanism and
the solid-state growth mechanism. On the one hand, the
dissolved metal species from active sites at the base of the
pores enable precipitation of oxides onto the PL. These
oxides, either coming from the oxidation of local metal spe-
cies or migrating from other places, deposit on top of the
PL, resulting in the formation of a spongy PL (as shown
in Fig. 2). On the other hand, the PL may also grow into
the NL through the solid-state growth mechanism. The
porous grain boundary within the PL acts as a preferential
path that allows the easy diffusion of solutes (Fe, Ni, O and
Cr). A number of small crystals (Fig. 2), which nucleate
from the NL, grow up and form a transition zone between
two layers. Meanwhile, the NLs seem to grow into the
matrix grains but be suppressed by a thin copper layer,
which is segregated in the interface between the NLs and
matrix grains. This interface appears as a line with a lot
of bumps, suggesting that the growth rate of the NL is
not equivalent at different sites. The variations of local
environments such as phase composition, defect and ele-
mental diffusivities may be responsible for that. Further
investigation is in progress.
Type 316 stainless steel foils containing stress corrosion
cracks grown in high temperature aqueous environments
have been examined by TEM. It was found that crack tips
are oxidized and have a three-layered morphology. All the
layers taper towards the crack tip in which the inner layer is
a microcrystalline spinel sandwiched between outer layers
of a nano-crystalline oxide. The outer layers are enriched
in Cr, and the inner with Fe, relative to the matrix. Cu
was observed to segregate at the interface between oxide
and matrix at one crack in type 316 steel. Oxide growth
relies on different mechanisms before and after grain
Y.Z. Huang is grateful to Rolls Royce Marine Power for
financial support and the provision of samples. The support
of INSS (JMT), AEA Technology and the Royal Academy
of Engineering (JMT) is gratefully acknowledged.
 Was GS, Bruemmer SM. J Nucl Mater 1994;216:326.
 Bruemmer SM, Simonen EP, Scott PM, Andresen PL, Was GS,
Nelson JL. J Nucl Mater 1999;274:299.
 Gertsman VY, Bruemmer SM. Acta Mater 2001;49:1589.
 Bruemmer SM. Mater Sci Forum 1999;294:75.
 Stiller K, Nilsson JO, Norring K. Metall Mater Trans A 1996;27:327.
 Armijo JS. Corrosion 1968;8:24.
 Andresen PL. Corros Sci 1993;49:714.
 Stroosnijder MF, Guttmann V, de Wit JHW. Metall Mater Trans A
 Cookson JM, Was GS, Andresen PL. Corros Sci 1998;54(4).
 Huang YZ, Lozano-Perez S, Langford RM, Titchmarsh JM. J
 Lister DH, Davidson RD, Mcalpine E. Corros Sci 1987;27:113.
 Ghosh DC, Biswas R. Int J Mol Sci 2002;3:87.
 Rebak RB, Szklarska-Smialowska Z. Corros Sci 1996;38:971.
 Guttmann V, Stein K, Bakker WT. Mater High Temp 1997;14:127.
 Stellwag B. Corros Sci 1998;40:337.
 Warzee M, Sonnen C, Berge, PH. Corrosion of carbon steels and
stainless steels in pressurised water of high temperatures, European
Atomic Energy Commission, Report EURAEC 1896; 1967.
 Tapping RL, Davidson RD, McAlpine E, Lister DH. Corros Sci
 Szklarska-Smialowska Z, Chou KC, Xia Z. Corros Sci 1991;32:609.
 Robertson J. In: Proceedings of the conference on water chemistry of
nuclear reactor systems, vol. 5. London: British Nuclear Engineering
Society; 1989. p. 81.
 Robertson J. Corros Sci 1989;29:1275.
 WinklerR, HtittnerF,Michel
Y.Z. Huang, J.M. Titchmarsh / Acta Materialia 54 (2006) 635–641