Available online at www.sciencedirect.

com

Engineering Fracture Mechanics 75 (2008) 2683–2694

www.elsevier.com/locate/engfracmech

Delamination fracture of prestressing steel:

An engineering approach
J. Toribio *

Department of Materials Engineering, University of Salamanca, E.P.S., Campus Viriato, Avda. Requejo 33, 49022 Zamora, Spain

Available online 12 March 2007

Abstract

This paper provides a composites engineering approach to explain the stress corrosion behaviour of high-strength pre-
stressing steel wires. To this end, two eutectoid steels in the form of hot rolled bar and cold drawn wire were subjected to
slow strain rate tests in aqueous environments in corrosive conditions corresponding to localized anodic dissolution and
hydrogen assisted cracking. While a tensile crack in the hot rolled bar always propagates in mode I, in the cold drawn wire
an initially mode I crack deviates significantly from its normal mode I growth plane and approaches the wire axis or cold
drawing direction, thus producing a mixed mode propagation. In hydrogen assisted cracking the deviation happens just
after the fatigue pre-crack, whereas in localized anodic dissolution the material is able to undergo mode I cracking before
the deflection takes place. Therefore, a different behaviour is observed in both steels and even in the same steel under dis-
tinct environmental conditions. An explanation of such behaviour can be found in the pearlitic microstructure of the steels.
This microstructural arrangement is randomly-oriented in the case of the hot rolled bar and markedly oriented under the
wire axis direction in the case of the cold drawn wire. Thus both materials behave as composites at the microstructural level
and their plated structure (oriented or not) would explain the different time-dependent behaviour in a corrosive
environment.
 2007 Elsevier Ltd. All rights reserved.

Keywords: High-strength steel; Stress corrosion cracking; Hydrogen assisted cracking; Cold drawing; Anisotropy; Crack deflection; Crack
branching

1. Introduction

The manufacture of pearlitic steels to produce prestressing steel wires for use in prestressed concrete
structures is carried out by cold drawing a previously hot rolled bar in several passes to increase the yield
strength of the material on the basis of a strain hardening mechanism. This technique produces important

Abbreviations: CD, cold drawn; HAC, hydrogen assisted cracking; HEDE, hydrogen enhanced delamination; HELP, hydrogen en-
hanced localized plasticity; HR, hot rolled; LAD, localized anodic dissolution; TTS, tearing topography surface.
*
Tel.: +34 980 54 50 00; fax: +34 980 54 50 02.
E-mail address: toribio@usal.es

0013-7944/$ - see front matter  2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.engfracmech.2007.03.015
2684 J. Toribio / Engineering Fracture Mechanics 75 (2008) 2683–2694

microstructural changes in the material, which influence its posterior fracture and stress corrosion cracking
performance [1].
Previous works [2,3] describe a kind of fracture by longitudinal splitting called flat-bottomed or layer-type of
corrosion damage [2] or exfoliation corrosion [3]. The problem of hydrogen embrittlement of these materials has
also received attention. Work [4] deals with hydrogen induced delayed failure of smooth wires, and a paper [5]
studies the role of notches in the hydrogen embrittlement of steel prestressing tendons.
A previous paper [6] showed the anisotropic stress corrosion behaviour of cold drawn wires exhibiting
mixed mode crack growth as a consequence of their microstructural orientation induced by the manufacturing
process (cold drawing). Another paper [7] demonstrated that the crack growth rate is strongly dependent on
the particular material (hot rolled or cold drawn) and the specific crack growth direction under consideration.
This paper goes further in the analysis and provides a rationalization of such a stress corrosion behaviour
of the hot rolled bar and the cold drawn wire considering micromechanical aspects of the damage/fracture
process on the basis of the more or less oriented lamellar microstructure of the steels which thus behave as
composite materials from the fracture mechanics viewpoint.

2. Experimental procedure

A high-strength eutectoid steel was used in this work. The chemical composition is given in Table 1. This
steel was tested in two conditions: firstly, as a hot rolled patented cylindrical bar of 12 mm diameter and, sec-
ondly, as a commercial 7 mm diameter cold drawn prestressing wire obtained from the bar. After rolling, the
12 mm diameter bar was patented by cooling from the austenitic condition in a molten lead bath to produce
fine pearlite. The wire was obtained by cold drawing this bar in six passes, to achieve an overall reduction of
66%, and finally a stress-relieving process was applied involving exposure to about 400 C for a few seconds.
The mechanical properties of both the bar and the wire are presented in Table 2. The fracture toughness KIC
was determined using cylindrical pre-cracked specimens obtained from the bar and the wire—for which the
plane strain condition is achieved at the inner points of the crack [8]—together with an expression for the maxi-
mum stress intensity factor at the deepest point of the crack (assumed semi-elliptical) calculated in [9] using the
finite element method combined with a virtual crack extension technique.
The microstructure of both steels consists of fine pearlite with an interlamellar spacing of 0.1 lm measured
according to the procedure described in [10]. Fig. 1 shows the microstructure of both steels in transverse and
longitudinal cross sections. While the hot rolled bar has a randomly-oriented microstructure in both transverse
and longitudinal sections (Fig. 1a and c), the cold drawn wire presents a randomly-oriented appearance in the
transverse cross section (Fig. 1b), but a marked orientation in the longitudinal cross section (Fig. 1d), which
implies an effect of manufacturing on the resulting microstructure. Thus the cold drawn wire presents features
consisting mainly of alternate lamellae of ferrite and cementite aligned parallel or quasi-parallel to the wire
axis or cold drawing direction.

Table 1
Chemical composition (wt%) of the steel
C Mn Si P S Cr Ni Mo
0.74 0.70 0.20 0.016 0.023 0.01 0.01 0.001

Table 2
Mechanical properties of the bar and the wire
Steel Young’s modulus Yield strength U.T.S. Uniform Reduction of Fracture toughness
(GPa) (MPa) (MPa) elongation (%) area (%) (MPa m1/2)
1. Hot rolled 195 725 1300 8.0 30 53
(HR)
2. Cold drawn 190 1500 1830 5.8 37 84
(CD)
 J. Toribio / Engineering Fracture Mechanics 75 (2008) 2683–2694 2685

Fig. 1. Microstructure of both the hot rolled and the cold drawn steels in transverse and longitudinal metallographic sections: (a) hot
rolled-transverse, (b) cold drawn-transverse, (c) hot rolled-longitudinal and (d) cold drawn-longitudinal.

To evaluate the behaviour of the steels in aggressive media, slow strain rate tests were performed on trans-
versely pre-cracked rods immersed in aqueous environment under electrochemical control and with a cross-
head speed of 8.3 · 108 ms1, as described elsewhere [6]. Pre-cracking of the samples was carried out by
axial fatigue in air environment, using different fatigue loads during the last step (just previous to the fracture
test). The maximum stress intensity factor (K)-levels in fatigue were Kmax = 0.28KIC, 0.45KIC, 0.60KIC and
0.80KIC, where KIC is the fracture toughness of the material in air (cf. Table 2).

3. Experimental results

3.1. Engineering approach: fracture load

The macroscopic effects of the environment on fracture were quantified through the ratio of the failure load
in the solution (critical value FC) to the failure load in air (reference value F0), as depicted in Fig. 2. All results
showed the anodic and cathodic regimes of environment-sensitive cracking [11]: for higher potentials
(E = 400 mV SCE) the anodic regime, associated with localized anodic dissolution (LAD); for lower poten-
tials (E = 1200 mV SCE) the cathodic regime, associated with hydrogen assisted cracking (HAC), cf. [6].
An important Kmax-effect is observed due to compressive residual stresses in the vicinity of the crack tip dur-
ing fatigue pre-cracking of the samples. This phenomenon has been discussed in [11], the main conclusion being
that high values of maximum K during fatigue pre-cracking (Kmax) produce strong compressive residual stresses
in the vicinity of the crack tip, thus delaying the hydrogen entry (in HAC) or the metal dissolution (in LAD).
Both the hot rolled bar and the cold drawn wire are more susceptible to HAC than to LAD (in engineering
macroscopic terms). The cold drawing process is beneficial against LAD phenomena, since it clearly increases
the fracture load in the anodic regime. However, cold drawing is damaging against HAC processes, since it
lowers the fracture load in a hydrogen environment (cathodic regime) for the whole range of Kmax-values.

3.2. Physical approach: microscopic fracture modes

Fig. 3 gives the microscopic fracture modes. In HAC conditions (cathodic regime) the hot rolled bar fails in
mode I associated with the so-called tearing topography surface (TTS) fracture mode [12,13] followed by
2686 J. Toribio / Engineering Fracture Mechanics 75 (2008) 2683–2694

0.9 HR/HAC
HR/LAD
CD/HAC
CD/LAD
0.8

Fc/Fo
0.7

0.6

0.5
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Kmax/KIC
Fig. 2. Macroscopic results of the slow strain rate tests, quantified through the ratio of the failure load in the solution (critical value FC) to
the failure load in air (reference value F0), for both materials (HR: hot rolled, CD: cold drawn) and the two environmental conditions
(HAC: hydrogen assisted cracking, LAD: localized anodic dissolution).

Fig. 3. Microscopic modes of fracture for an intermediate Kmax-level of 0.45KIC: (a) hot rolled bar in HAC conditions showing TTS; (b)
cold drawn wire in HAC conditions showing shear topography and isolated cleavage facets; (c) hot rolled bar in LAD conditions showing
cleavage and (d) cold drawn wire in LAD conditions showing mode I dissolution and crack deflection.

cleavage-like propagation, whereas the cold drawn wire exhibits a shear topography with some evidence of
isolated cleavage facets, and the crack approaches the axis direction producing a mixed mode stress state
(longitudinal splitting or delamination). There are two embryos of fracture located symmetrically in relation
to the initial crack plane (at an angle of about 80), but only one of them becomes the final fracture path, and
thus the initial crack branching progresses along only one of the branches, probably for statistical reasons,
which makes it the fracture path of lower fracture resistance or that with the higher concentration of
hydrogen.
Under LAD conditions (anodic regime) the hot rolled bar fails in mode I by cleavage-like topography
whereas the cold drawn wire exhibits a short mode I crack growth path (50 lm in depth), and a posterior crack
deflection with a deviation angle of about 80 from the initial crack path in mode I. Thus the mixed mode
 J. Toribio / Engineering Fracture Mechanics 75 (2008) 2683–2694 2687

crack growth associated with longitudinal splitting or delamination also appears in anodic (pure stress corro-
sion cracking) conditions.

3.3. Fracture mechanics approach: critical stress intensity factor

The time-dependent stress corrosion properties are a consequence of the behaviour of both steels in corro-
sive media in conditions of HAC and LAD, and the important differences arise when a crack deviation from
its initial plane appears and a mixed mode propagation takes place in the form of crack branching (in the case
of HAC) or crack deflection (in the case of LAD), as shown in Fig. 4. As a matter of fact, the stress corrosion
resistance of the cold drawn steel is strongly dependent on the capacity of this material to undergo mode I
cracking across the more resistant microstructural paths, since when the crack deviates from its original direc-
tion it follows a path of minimum resistance to stress corrosion cracking associated with the very oriented
microstructure of the cold drawn steel.
The environmental effects on fracture of the considered materials can be analyzed in Table 3 which sum-
marizes the results of the slow strain rate tests for the slightest fatigue pre-cracking regime (Kmax = 0.28KIC),
i.e., for that which produces the minimum previous mechanical pre-damage before the stress corrosion tests
and thus allows the maximum degree of environmental (stress corrosion) damage and a more proper estima-
tion of stress corrosion behaviour of the steels.
In the case of HAC (cf. Fig. 4 and Table 3) the hot rolled bar exhibits better behaviour than the cold drawn
wire since, although the fracture load is the same when compared with its respective value in air, the ratio of
stress intensity values in aggressive medium (hydrogen) to the reference value in air is higher in the hot rolled
material as a consequence of the subcritical crack growth by TTS taking place along the original cracking
direction in mode I.
For LAD (cf. Fig. 4 and Table 3) the situation is the opposite, and in this case the effect of cold drawing on
stress corrosion (LAD) performance of the steels is clearly beneficial and the subcritical crack growth in mode
I (xLAD) extends the time to failure and raises the fracture load before the crack deflection (and the mixed
mode propagation) takes place.

a b
XTTS C

c d
C X LAD

Fig. 4. Schema showing the environment-sensitive cracking modes: (a) hot rolled bar in HAC conditions, (b) cold drawn wire in HAC
conditions, (c) hot rolled bar in LAD conditions, (d) cold drawn wire in LAD conditions.

Table 3
Summary of results of the slow strain rate tests (Kmax = 0.28KIC)
Environment Material Critical crack Critical load Critical SIF Crack path
HAC HR aC = a0 + xTTS FC/F0  0.56 KC/K0 > 0.56 Mode I
CD aC  a0 FC/F0  0.56 KC/K0  0.56 Mixed modea [Branching]
LAD HR aC  a0 FC/F0  0.70 KC/K0  0.70 Mode I
CD aC = a0 + xLAD FC/F0  0.96 KC/K0 > 0.96 Mixed modea [Mode I+Deflection]
The subindex c indicates critical (failure) instant in the tests in aggressive solution and the subindex 0 refers to the pre-crack length before
the tests (a0) or to the reference fracture value in air (F0, K0).
HAC: hydrogen assisted cracking; HR: hot rolled; LAD: localized anodic dissolution; CD: cold drawn; TTS: tearing topography surface;
a: crack length; SIF: stress intensity factor K; F: applied load.
a
Angle of 70–90.
2688 J. Toribio / Engineering Fracture Mechanics 75 (2008) 2683–2694

3.4. Discussion

In HAC, both the hot rolled and the cold drawn materials suffer a marked reduction of fracture load due to
the hydrogen degradation process, and a subsequent reduction of critical stress intensity factor KHAC (’’appar-
ent toughness’’) which is KHAC/K0  0.56 in the cold drawn wire and KHAC/K0 > 0.56 in the hot rolled bar as a
consequence of the subcritical mode I cracking by TTS (cf. Table 3).
In the case of LAD, the hot rolled bar exhibits a marked reduction of ’’apparent toughness’’ in relation to
the actual toughness obtained in air (KLAD/K0  0.70) in spite of the cleavage fracture micromechanism oper-
ating in both cases. Although the cleavage topography is predominant in both situations, for the fracture in air
(K0) it is mechanical cleavage, whereas for the fracture in corrosive medium (KLAD) it could be the so-called
film-induced cleavage [14,15], i.e., a brittle film which injects a crack into the substrate.
With regard to the critical stress intensity factor for LAD in the cold drawn wire, it could be even higher
than the corresponding value in air (KLAD/K0 > 0.96). It can be explained by the blunting of the crack tip
caused by the anodic dissolution itself [16–18] which produces an apparently higher critical K-value. This
blunting effect is enhanced by the oriented microstructure of the drawn material (specially at the ferrite/
cementite interface) and can block the film-induced cleavage mechanism which otherwise would be operative
(e.g. in the hot rolled material).
The fundamental idea summarizing the results of the cold drawn steel is that such a material is highly sus-
ceptible to HAC—even more than the hot rolled initial material—but at the same time extremely resistant to
LAD (without appreciable loss of fracture load). This elevated LAD resistance is associated with a clear incre-
ment of time to failure and is achieved because of the important fact that—in spite of the clear orientation of
the steel microstructure—the crack is able to propagate in mode I breaking the strongest links, blunting the
crack tip and blocking the film-induced cleavage mechanism. This is the key point related to the cold drawn
wire: that the crack is able to undergo mode I cracking before final fracture, thus increasing both the critical
stress intensity factor and the time to failure. In the following section a micromechanical model is proposed to
explain this behaviour.

4. Micromechanics of time-dependent behaviour of the cold drawn material

4.1. Material anisotropy

A previous paper [6] showed that the pearlitic microstructure of the cold drawn steel becomes oriented along
the wire axis direction as a direct consequence of cold drawing. In this section a simple composite approach is
used to model the oriented microstructure of the cold drawn wire. Two approaches may be followed: a 1D-
approach in the form of a fiber-reinforced composite to reproduce the alignment of the microstructure and
a 2D-approach in the form of a laminate to account for the plated microstructure at the finest microscopic level.
Considering that both microstructural levels are very markedly oriented (angle of 70–90 to the radial direction
or initial crack growth plane in mode I), the modelling assumes that they are totally oriented and the angle is
90, i.e., the fibers in the 1D modelling (or the plates in the 2D modelling) are completely oriented in the wire
axis or cold drawing direction. Since the problem is axisymmetric the fiber model may be adopted and thus the
prestressing steel (cold drawn) wire can be considered as a fiber-reinforced composite.
Such a microstructural arrangement has consequences of both a mechanical and a chemical nature, namely:

(i) Strength anisotropy (mechanical anisotropy), i.e., fracture toughness KIC (or similar strength parameter)
as a function of the orientation angle: KIC = KIC(h), where h is the common fracture mechanics angle
between the specific direction under consideration and the original plane of the macroscopic mode I
crack perpendicular to the wire axis or main direction of the fiber alignment in the composite approach.
In particular, the fracture toughness values in directions perpendicular and parallel to the fibers are
respectively
K IC? ¼ K IC ðh ¼ 0 Þ for fibers fracture ðtransverse fractureÞ ð1Þ

K ICk ¼ K IC ðh ¼ 90 Þ for longitudinal splitting ð2Þ
 J. Toribio / Engineering Fracture Mechanics 75 (2008) 2683–2694 2689

the former is the fracture toughness for breaking the fibers in the 1D model (fracture of the plates in the
2D model or the lamellae in the real material), whereas the latter is the fracture toughness for delami-
nation or, generally speaking, debonding of any microstructural elements at any scale, e.g., the pearlitic
colonies which are also very slender and markedly oriented.
Accounting for microstructural reasons, it may be assumed that KIC? is clearly higher than KICk, i.e.,
K IC?  K ICk ð3Þ
(ii) Chemical anisotropy, i.e., hydrogen diffusion coefficient D as a function of the orientation angle:
D = D(h), h having the same meaning as above. This coefficient is a key item in HAC processes in pearl-
itic steels where stress-assisted diffusion has been shown to be the main hydrogen transport mechanism
[19]. In particular, the diffusion coefficients in direction perpendicular and parallel to the fibers are
respectively
D? ¼ Dðh ¼ 0 Þ diffusion in the transverseðcrackÞdirection ð4Þ
Dk ¼ Dðh ¼ 90 Þ diffusion in the longitudinal direction ð5Þ
the former corresponds to diffusion in the macroscopic crack direction, perpendicular to the fibers (or
lamellae or plates, according to the modelling), whereas the latter is associated with diffusion parallel
to the fibers.

Considering again microstructural reasons, Dk is clearly higher than D?, i.e.,

Dk  D? ð6Þ
which indicates that hydrogen tends to diffuse in the direction h = 90, thus creating a pre-damage at the
microscopic level. This could explain why in HAC tests there are two potential fracture initiation planes, sym-
metrically oriented in relation to the direction h = 0, one of them producing the final fracture (Fig. 4b).

4.2. Micromechanisms of HAC

In the matter of hydrogen transport, diffusion is the main operative mechanism, as explained above. When
stress-assisted diffusion is considered, hydrogen transport is driven by both the gradients of concentration and
hydrostatic stress [20,21], which implies that the distribution of hydrostatic stress r in the vicinity of the crack
tip governs the hydrogen transport. This distribution is a function of the polar angle h (in relation to the crack
line in the common fracture mechanics sense). For a plane strain mode I crack this distribution is [22]
2 KI h
r ¼ ð1 þ mÞ pffiffiffiffiffiffiffi cos ð7Þ
3 2pr 2
where m is the Poisson coefficient, KI the stress intensity factor and r the radial distance from the crack tip.
For the case of a plane stress crack it is
2 KI h
r¼ pffiffiffiffiffiffiffi cos ð8Þ
3 2pr 2
In both cases it is proportional to cos(h/2), with maximum for h = 0. Thus hydrogen tends to diffuse, in an
isotropic material, towards the points located at h = 0 and the crack grows in mode I. When the material is
anisotropic, such as the cold drawn wire, although the maximum hydrostatic stress is located at h = 0 (polar
angle form the crack tip), there are two important reasons for the crack deviation from its initial propagation
path in mode I:

(i) strength anisotropy of the cold drawn wire (lower toughness in the wire axis direction),
(ii) chemical anisotropy (higher hydrogen diffusion coefficient in the wire axis direction),

so that hydrogen diffuses faster in the longitudinal direction (Dk relatively high) than in the transverse direc-
tion (Dk relatively low), as depicted in Fig. 5a.
2690 J. Toribio / Engineering Fracture Mechanics 75 (2008) 2683–2694

D //
a b
HEDE
D

D//

Fig. 5. Micromechanism of HAC: (a) hydrogen diffusion in longitudinal and transverse directions, (b) fracture by hydrogen enhanced
delamination or debonding (HEDE).

With regard to the critical mechanism of failure in HAC, two micromechanisms could be operative, accord-
ing to a terminology proposed by Gerberich et al. [18]:

(i) Hydrogen enhanced localized plasticity (HELP).

(ii) Hydrogen enhanced delamination (HEDE).

In the latter case the original term coined by Gerberich was hydrogen enhanced decohesion, with the same
acronym HEDE, but in this work the term delamination is more adequate when the plated microstructure of
the steel is taken into account, but it could also be hydrogen enhanced debonding (or splitting) if the fiber-
reinforced model is considered.
With regard to HELP, its importance in the HAC of prestressing steel is probably low, since the lamellar
structure of the steel (markedly oriented, analogous to a fiber-reinforced material or a laminate at the micro-
scopical level) probably delays or even blocks the dislocation movement which otherwise could be operative in
an isotropic material. In a lamellar structure such as that of the cold drawn wire, the cementite plates act as
obstacles to the movement of dislocations, so they can block (or diminish) the dislocation multiplication. In
addition, HELP is a mechanism of failure associated with hydrogen transport by dislocation dragging which,
as reported by Toribio [19] does not seem to be the main hydrogen transport mechanism in these kinds of steel.
In the matter of HEDE, its importance in the HAC of prestressing steel is probably high for the same rea-
son as explained above, i.e., the lamellar structure of the steel (markedly oriented) which produces anisotropy
regarding fracture and hydrogen diffusion, so that hydrogen diffuses mainly in the direction of the plates and
can weaken the bonds or interfaces between the ferrite and the cementite lamellae (which are the weakest links
even before the hydrogen presence) thus contributing to the hydrogen induced fracture by delamination or
debonding between two similar microstructural units, i.e, at the ferrite/cementite interface or at the pearlitic
colony boundaries, cf. Fig. 5b. Thus the micromechanism of HAC is TTS in the hot rolled bar (isotropic) and
HEDE in the cold drawn wire (strongly anisotropic).

4.3. Micromechanisms of LAD

A question arises here as to why the crack does not change its propagation path in spite of the clearly ori-
ented microstructure of the steel: a kind of fibrous, plated or composite microstructure which in HAC pro-
duces a sudden change of 70–90 in crack propagation direction at the beginning of the subcritical
environment-sensitive crack growth rate, i.e., just after the fatigue pre-crack. The explanation could lie in
the local (or crack tip) strain rate required to promote the anodic dissolution process [23] which is achieved
only at the crack tip in the h = 0 direction (i.e., in the crack plane) due to the stress–strain distribution in its
vicinity.
Then, why does the crack finally change its propagation direction after a certain subcritical crack growth by
LAD in mode I? Because of the presence in the heavily drawn material of pearlitic pseudocolonies extremely
slender, with anomalous (too large) local interlamellar spacing and even with pre-damage (microcracks which
act as local fracture precursors) which makes them preferential fracture paths with minimum local resistance
[24]. The pearlitic pseudocolonies are aligned to the cold drawing direction in the metallographic analysis of
the cold drawn steel and they are potential fracture sites for two reasons:
 J. Toribio / Engineering Fracture Mechanics 75 (2008) 2683–2694 2691

(i) The very high local interlamellar spacing which makes them weaker or potentially fracturable by shear
cracking of pearlitic plates according to the mechanism proposed by Miller and Smith [25].
(ii) The presence of some microcracks and defects consisting of plates prefractured in the pseudocolony dur-
ing the manufacturing process (cold drawing) as a consequence of the very high stresses applied on the
wire.

The mechanism of LAD in cold drawn steel could be explained as follows: dissolution is produced in mode
I along a distance xLAD (see Fig. 6). The crack continues in mode I along the initial plane and only deviates
when it reaches a defect (pre-damage) in the material: the afore- mentioned pearlitic pseudocolonies which are
potential fracture sites. When this happens, final fracture takes place for purely mechanical reasons, so the
distance xLAD has statistical significance, since it depends on the position of the pseudocolonies and thus
on their frequency of appearance which is higher when drawing becomes heavier.
The longer the mode I path (xLAD) in LAD processes, the slower the crack growth rate and the more
delayed the final fracture. This delay is the reason why the cold drawn wire is very resistant to LAD, while
at the same time is strongly susceptible to HAC, because in this case the crack is unable to propagate in mode
I and the deviation suddenly appears, producing mixed mode propagation and a quite fast crack growth rate.

4.4. Evaluation of directional fracture toughness in aggressive environments

To evaluate the directional fracture toughness (i.e. mechanical strength) of the steel wires in aggressive envi-
ronments, an expression is needed of the stress intensity factor (SIF) for the considered geometry (a cylinder
with a transverse crack perpendicular to its axis and with semi-elliptical shape). In this paper, the following
general expression [26] is used:
pffiffiffiffiffiffi
K I ¼ Y ða=DÞr pa ð9Þ

where r is the remote axial stress, a the crack depth and Y(a/D) a dimensionless function given by (cf. [26])
h i1=2 h i1=4
2 2
Y ða=DÞ ¼ 0:473  3:286ða=DÞ þ 14:797ða=DÞ ða=DÞ  ða=DÞ ð10Þ

which was obtained using the finite element method together with a compliance technique to obtain a global
value of the stress intensity factor from the energy release rate.
A fracture criterion on the basis of the energy release rate G will be used to account for the anisotropic
behaviour, i.e.,
G ¼ GC ð11Þ

In materials with strength anisotropy—as the cold drawn wires considered in this paper, cf. Fig. 4b and d—the
specific energy for fracture depends on the propagation angle h in relation to the crack plane in the standard
fracture mechanics sense, Gc(h), and it can be related to the directional fracture toughness KIC(h) as follows:
GC ðhÞ ¼ K 2IC ðhÞ=E0 ð12Þ

where E 0 = E in plane stress and E 0 = E/(1  m2) in plane strain. In addition, and taking into account the prop-
agation step (cf. Fig. 4b and d) oriented 90 in relation to the radial direction (original crack propagation

LAD

Fig. 6. Micromechanism of LAD by anodic dissolution of the fibers (mode I crack growth) and posterior mechanical fracture of a slender
pearlitic pseudocolony.
2692 J. Toribio / Engineering Fracture Mechanics 75 (2008) 2683–2694

direction in mode I), the relation between the energy release rates in the radial (h = 0) and axial (h = 90)
directions is [27]
Gð90 Þ ¼ 0:2615Gð0 Þ ð13Þ
which indicates clearly that the energy release rate for crack deflection is quite lower than the correspondent
value for crack propagation in mode I, and thus crack deviation from the mode I direction only happens if the
fracture resistance in the axial direction—measured in terms of specific energy for fracture—is low enough.
To obtain the approximate value of the directional toughness of cold drawn wires in aggressive environ-
ment promoting HAC and LAD (the final aim of the present paper), the following hypotheses will be used:

(i) The subcritical crack growth in mode I (before the 90 deflection, cf. Fig. 4) is neglected, i.e., the crack
length for calculating the critical value of the SIF at any load level is the fatigue crack length: aC  a0 in
all cases.
(ii) In HAC crack deflection takes place at the very beginning of the test and thus it is assumed that the crit-
ical load in hydrogen (FC/F0  0.56 in Table 3) is associated with a deflection angle h = 90 (propagation
in axial direction by delamination). Although the load probably increases further due to hydrogen
assisted (subcritical) cracking, this probably slight increase is neglected and thus the critical load is
slightly underestimated.
(iii) In LAD the crack is able to propagate in mode I and thus it is assumed that the critical load in hydrogen
(FC/F0  0.96 in Table 3) is associated with a null deflection angle h = 0 (propagation in radial direction
by anodic dissolution). When finally crack deflection takes place because the slowly propagating crack
reaches a pre-defect in the drawn material (oriented in the drawing direction) the load suddenly drops
due to this fact.

Now, the fracture criterion based on the energy release rate is applied to the fracture instants of Fig. 4b
(HAC) and d (LAD) to obtain the directional fracture toughness of the steels in aggressive environments
in directions h = 90 (Fig. 4b, HAC) and h = 0 (Fig. 4d, LAD). With regard to HAC, and considering
hypothesis (ii), the fracture criterion (11) yields
GðaHAC ; F HAC ; 90 Þ ¼ GHAC ð90 Þ ð14Þ
where the subindex HAC is now used to indicate the critical situation under HAC conditions. Considering the
relation (13) between the energy release rates in different directions
0:2615GðaHAC ; F HAC ; 0 Þ ¼ GHAC ð90 Þ ð15Þ
or, in terms of SIF, accounting for (12) and considering that in this mode I case G ¼ K 2I =E0
pffiffiffiffiffiffiffiffiffiffiffiffiffiffi
0:2615 K I ðaHAC ; F HAC Þ ¼ K HAC ð90 Þ ð16Þ
which gives the directional toughness in axial (h = 90) direction for HAC conditions. From Table 3 a value
KHAC (90)  0.29K0, i.e., the fracture toughness (or fracture resistance) for hydrogen enhanced delamination
(or debonding) causing longitudinal cracking in the cold drawn wires is about 30% of the fracture toughness in
air of the same material.
In LAD (Fig. 4d), and considering hypothesis (iii), the fracture criterion (11) gives
GðaLAD ; F LAD ; 0 Þ ¼ GLAD ð0 Þ ð17Þ
and considering now G ¼ K 2I =E0 , the criterion may be expressed in terms of SIF as
K I ðaLAD ; F LAD Þ ¼ K LAD ð0 Þ ð18Þ
which gives the directional toughness in radial (h = 0) direction for LAD conditions. From Table 3 a value
KLAD (0) > 0.96K0, i.e., the fracture toughness (or fracture resistance) for anodic dissolution in mode I caus-
ing transverse cracking in the cold drawn wires is, considering the underestimation of both the critical crack
and the critical load, approximately the same as the fracture toughness in air of the same material.
Therefore, the composites engineering approach to the phenomenon of environmentally assisted cracking
of prestressing steels show that the cold drawn wires are extremely susceptible to HAC (by hydrogen enhanced
 J. Toribio / Engineering Fracture Mechanics 75 (2008) 2683–2694 2693

delamination fracture in axial direction) while at the same time extremely resistant to LAD (by mode I anodic
dissolution in transverse direction) because of the fact that in the later case the material is able to undergo
transverse cracking.

5. Conclusions

1. The cold drawn steel is strongly susceptible to HAC—even more than the hot rolled initial material—but at
the same time extremely resistant to LAD. This is because in the case of LAD the crack is able to undergo
mode I growth, breaking the strongest links, blunting the crack tip and blocking the film-induced cleavage
mechanism which would otherwise appear before the final fracture.
2. The pearlitic microstructure of the cold drawn steel becomes oriented in the wire axis direction as a direct
consequence of cold drawing. This happens at the two basic microstructural levels: the pearlitic colony and
the pearlite lamellae. Therefore a simple approach may be used to model the oriented microstructure in the
form of a fiber-reinforced composite which exhibits anisotropy of both a mechanical (strength) and a chem-
ical nature.
3. The micromechanism of HAC in the cold drawn steel (strongly anisotropic) is HEDE, i.e., hydrogen
enhanced delamination or, generally speaking, hydrogen enhanced debonding between two similar micro-
structural units, i.e, at the ferrite/cementite interface or at the boundaries between pearlitic colonies.
4. The mechanism of LAD in cold drawn steel could be explained by dissolution in mode I along a certain
distance. The crack deviates when it reaches a pearlitic pseudocolony very slender and with anomalous local
interlamellar spacing which is a potential fracture site and fails for purely mechanical reasons.
5. These micromechanisms have direct consequences in the stress corrosion behaviour of the steels. In HAC
the subcritical crack growth is strongly dependent on time and on the degree of cold drawing, since the
hydrogen diffusion coefficient is a directional property in relation to the cold drawing axis. In LAD, the
dissolution takes place over the strongest links and thus produces a time-dependent effect in the form of
delay of final fracture.
6. The directional fracture toughness of the anisotropic cold drawn material was evaluated in aggressive envi-
ronments promoting HAC and LAD. The composites engineering approach shows an axial toughness for
hydrogen assisted delamination KHAC(90)  0.29K0 and a radial toughness for transverse cracking by ano-
dic dissolution KLAD(0) > 0.96K0 where K0 is the reference value (fracture toughness) of the material in air.

Acknowledgements

The author wishes to thank the financial support of his research at the University of Salamanca provided by
the following institutions: Spanish Ministry for Scientific and Technological Research (MCYT-FEDER;
Grant MAT2002-01831), Spanish Ministry for Education and Science (MEC; Grant BIA2005-08965), FED-
ER-INTERREG III (Grant RTCT-B-Z/SP2.P18), Regional Government Junta de Castilla y León (JCYL;
Grants SA078/04 and SA067A05), Spanish Foundation ‘‘Memoria de D. Samuel Solórzano Barruso’’ (Grants
for Scientific and Technological Research) and University of Salamanca (Grants for Research Projects).

References

[1] Parkins RN, Elices M, Sánchez-Gálvez V, Caballero L. Environment sensitive cracking of pre-stressing steels. Corros Sci
1982;22:379–405.
[2] Cherry BW, Price SM. Pitting, crevice and stress corrosion cracking studies of cold drawn eutectoid steels. Corros Sci
1980;20:1163–84.
[3] Sarafianos N. Environmentally assisted stress-corrosion cracking of high-strength carbon steel patented wire. J Mater Sci Lett
1989;8:1486–8.
[4] Langstaff DC, Meyrick G, Hirth JP. Hydrogen induced delayed failure of high strength alloy steel wires. Corrosion 1981;37:429–37.
[5] Price SM, Warden PG, Chia TS, Scarlett NVY, Mahoney JH. The role of notches in the hydrogen-assisted cracking of steel
prestressing tendons. Mater Forum 1992;16:245–51.
[6] Toribio J, Lancha AM. Effect of cold drawing on environmentally assisted cracking of cold drawn steel. J Mater Sci 1996;31:6015–24.
[7] Toribio J, Lancha AM. Anisotropic stress corrosion cracking behaviour of prestressing steel. Mater Corros 1998;49:34–8.
2694 J. Toribio / Engineering Fracture Mechanics 75 (2008) 2683–2694

[8] Elices M. Fracture of steels for reinforcing and prestressing concrete. In: Sih GC, DiTommaso A, editors. Fracture mechanics of
concrete: structural application and numerical calculation. Dordrecht: Martinus Nijhoff; 1985. p. 226–71.
[9] Astiz MA. An incompatible singular elastic element for two- and three- dimensional crack problems. Int J Fract 1986;31:105–24.
[10] ASM. Metals handbook. Metallography and microstructures, vol. 9. Metals Park, OH: ASM; 1985.
[11] Toribio J, Lancha AM, Elices M. Factors influencing stress corrosion cracking of high strength pearlitic steels. Corros Sci
1993;35:521–30.
[12] Thompson AW, Chesnutt JC. Identification of a fracture mode: the tearing topography surface. Metall Trans 1979;10A:1193–6.
[13] Costa JE, Thompson AW. Hydrogen cracking in nominally pearlitic 1045 steel. Metall Trans 1982;13A:1315–8.
[14] Saito M, Smith GS, Newman RC. Testing the film-induced cleavage model of stress-corrosion cracking. Corros Sci 1993;35:411–7.
[15] Newman RC, Saito M. Anodic stress-corrosion cracking: slip-dissolution and film-induced cleavage. In: Magnin T, Gras JM, editors.
Corrosion–Deformation Interactions (CDI’92). Les Editions de Physique, Les Ulis, 1993; p. 3–26.
[16] Chu WY, Hsiao CM, Li SQ. A new engineering fracture toughness parameter KISCC(q). Scripta Metall 1979;13:1057–62.
[17] Ray KK, Rao GR. A new test principle for determining threshold stress intensity factor KIEAC in environment assisted cracking. Int J
Fract 1993;61:R69–75.
[18] Gerberich WW, Marsh P, Hoehn S, Venkataraman S, Huang H. Hydrogen/plasticity interactions in stress corrosion cracking. In:
Magnin T, Gras JM, editors. Corrosion–Deformation Interactions (CDI’92). Les Editions de Physique, Les Ulis, 1993; p. 325–53.
[19] Toribio J. Hydrogen–plasticity interactions in pearlitic steel: a fractographic and numerical study. Mater Sci Engng
1996;A219:180–91.
[20] Van Leeuwen HP. The kinetics of hydrogen embrittlement: a quantitative diffusion model. Engng Fracture Mech 1974;6:141–61.
[21] Toribio J, Kharin V. K-dominance condition in hydrogen assisted cracking: the role of the far field. Fatigue Fract Engng Mater
Struct 1997;20:729–45.
[22] Broek D. Elementary engineering fracture mechanics. The Hague: Martinus Nijhoff Publishers; 1982.
[23] Toribio J. The role of crack tip strain rate in hydrogen assisted cracking. Corros Sci 1997;39:1687–97.
[24] Toribio J, Ovejero E, Toledano M. Microstructural bases of anisotropic fracture behaviour of heavily drawn steel. Int J Fract
1997;87:L83–8.
[25] Miller LE, Smith GC. Tensile fractures in carbon steels. J Iron Steel Inst 1970;208:998–1005.
[26] Valiente A. Criterios de fractura para alambres. PhD thesis, Polytechnic University of Madrid, 1980.
[27] Amestoy M, Bui HD, DangVan K. Analytic asymptotic solution of the kinked crack problem. In: Francois D, editor. Advances in
fracture research-ICF5. Oxford: Pergamon Press; 1981. p. 107–13.