Engineering Failure Analysis 58 (2015) 485–498

Contents lists available at ScienceDirect

Engineering Failure Analysis

journal homepage: www.elsevier.com/locate/efa

Hydrogen damage of steels: A case study and hydrogen

embrittlement model
M.B. Djukic ⁎, V. Sijacki Zeravcic, G.M. Bakic, A. Sedmak, B. Rajicic
University of Belgrade, Faculty of Mechanical Engineering, Kraljice Marije 16, Belgrade 11120, Serbia

a r t i c l e i n f o

Article history: Many efforts have been made to understand the effects of hydrogen on steels, resulting in an
Received 14 December 2014 abundance of theoretical models and papers. However, a fully developed and practically appli-
Received in revised form 30 April 2015 cable predictive physical model still does not exist industrially for predicting and preventing
Accepted 17 May 2015
hydrogen damage. In practice, it is observed that different types of damages to industrial boiler
Available online 3 June 2015
components have been associated with the presence and localization of hydrogen in metals. In
this paper, a damaged boiler tube made of grade 20 – St.20 (or 20G, equivalent to AISI 1020)
Keywords: was investigated. The experimental research was conducted in two distinctive phases: failure
Low carbon steel
analysis of the boiler evaporator tube sample and subsequent postmortem analysis of the via-
Hydrogen-assisted cracking
ble hydrogen embrittlement mechanisms (HE) in St.20 steel. Numerous tested samples were
Hydrogen embrittlement
Fractography cut out from the boiler tubes of fossil fuel power plant, damaged due to high temperature hy-
Impact strength drogen attack (HTHA) during service, as a result of the development of hydrogen-induced cor-
rosion process. Samples were prepared for the chemical composition analysis, tube wall
thickness measurement, tensile testing, hardness measurement, impact strength testing (on in-
strumented Charpy machine), analysis of the chemical composition of corrosion products – de-
posit and the microstructural characterization by optical and scanning electron microscopy –
SEM/EDX. The HTHA damage mechanism is a primary cause of boiler tube fracture. Based on
the multi-scale special model, applied in subsequent postmortem investigations, the results in-
dicate a simultaneous action of the hydrogen-enhanced decohesion (HEDE) and hydrogen-
enhanced localized plasticity (HELP) mechanisms of HE, depending on the local concentration
of hydrogen in investigated steel. The model is based on the correlation of mechanical proper-
ties to the SEM fractography analysis of fracture surfaces.
© 2015 Elsevier Ltd. All rights reserved.

1. Introduction

Hydrogen damage of boiler evaporator tubes is a very serious problem in many industrial boiler plants worldwide, including
even the most developed countries. This is due to a very specific and complex relationship between the hydrodynamic processes
in an evaporator and the boiler water treatment regimes, as well as the design and operating characteristics of a particular boiler
unit [1–3]. In practice, it is observed that different types of damages to industrial components of fossil fuel power plant have been
associated with the presence and localization of hydrogen in metals [2–5]. Due to the complexity of the processes involved and
difficulties to experimentally simulate the operating conditions of evaporator tubes of the steam drum boiler in power plants,
there is a very limited number of available information items about necessary preconditions for the initiation and development
of hydrogen damage of the boiler tubes made of low carbon steel.

⁎ Corresponding author. Tel.: +381 11 3370 375; fax: +381 11 3370 364.
URL: mdjukic@mas.bg.ac.rs (M.B. Djukic).

http://dx.doi.org/10.1016/j.engfailanal.2015.05.017
1350-6307/© 2015 Elsevier Ltd. All rights reserved.
486 M.B. Djukic et al. / Engineering Failure Analysis 58 (2015) 485–498

During exploitation of one natural circulation boiler in a 210 MW coal-fired power plant, significant failures of evaporator
(water wall) tubes occurred after 73,000 h of operation. Water wall tubes (K 60 × 6 mm) were made of plain carbon steel,
grade 20 – St.20 (or 20G, equivalent to AISI 1020) and were exposed to the operating parameters, pressure: P = 15.5 MPa
and temperature: t = 350 °C. Damage appeared on the boiler evaporator tubes located in the zone of membrane evaporator
panels, exposed during operation to the peak thermal load (boiler burner zone).
Numerous tested specimens were cut out from the boiler tube sample damaged during service. The experimental research was
conducted in two phases: a case study and failure analysis of the boiler evaporator tube sample and a subsequent – postmortem
analysis of the viable hydrogen embrittlement mechanisms in the investigated steel, which are unevenly enriched with hydrogen
as a result of the development of intensive and uneven local hydrogenation of the tube metal during boiler exploitation. A com-
prehensive case study and failure analysis of the boiler evaporator tube damaged during service due to the development of the
hydrogen-induced corrosion process [2,6] and the high temperature hydrogen attack (HTHA) damage mechanism were carried
out. A special experimental concept was applied during subsequent postmortem analysis of the viable hydrogen embrittlement
mechanisms in investigated St.20 steel, unevenly saturated with hydrogen due to the effect of HTHA, as a main mechanism re-
sponsible for the destruction of evaporator tubes during boiler exploitation. The proposed model is based on the correlation of
mechanical properties to the scanning electron microscopy fractography analysis of fractured surfaces in the presence of simulta-
neously active hydrogen embrittlement micro-mechanisms. A background for the failure analysis of the boiler evaporator tube
and the postmortem analysis of the viable hydrogen embrittlement mechanisms in investigated St.20 steel presented in this
paper is an overview-form of discussion, following in a Section 1.1, on the hydrogen damage of industrial components with spe-
cial emphasis on the hydrogen embrittlement mechanisms and their possible simultaneous action in steels.

1.1. Hydrogen damage and simultaneous action of hydrogen embrittlement mechanisms

Hydrogen-assisted mechanical degradation of the structural materials is one of the most complex phenomena of the metallic
materials degradation, due to a number of new unknowns in the degradation process kinetics, compared to the degradation in the
absence of hydrogen in metal. Hydrogen atoms dissolved in iron, steel and alloys show unique characteristics such as variable and
different diffusivity and solubility, in particular alloy and microstructure. Hydrogen may degrade the mechanical behavior of me-
tallic materials, including steels, and lead them to failure [7,8]. As a result, different types of hydrogen damage of industrial com-
ponents often occurred due to hydrogen presence in metals. Hydrogen damage is a commonly used generic name given to a large
number of metallic materials degradation processes and different mechanisms due to material interaction with hydrogen. Hydro-
gen damage includes a wide range of hydrogen-assisted mechanical degradation processes, the environmental types of failure,
material microstructural changes and hydrogen–material interactions caused by the presence of hydrogen in metals. In the
world literature, there is a strong disagreement and polarization of opinions about the possible classification and inherent nature
of the particular hydrogen damage mechanism [5,8–11].
Typically, researches in the world about the hydrogen damage mechanisms were carried out using the previous uniform en-
richment of metallic materials with hydrogen throughout the whole volume of the samples. On the other hand, during the oper-
ation metallic materials of industrial components are usually very unevenly saturated with hydrogen [12,13]. A critical experiment
that would allow realistic simulation of the kinetics of the development of a certain type of hydrogen damage, which is in full
compliance with the actual kinetics in the components of industrial plants exposed to hydrogen during service, is very difficult
to conduct [6].
Hydrogen embrittlement (HE), as one of the mechanisms of hydrogen damage, is the degradation of the mechanical properties
of metallic materials, loss of ductility and tensile strength, which usually result in a decrease of fracture resistance and sub-critical
cracking due to the presence of dissolved hydrogen. The HE of iron, steel and their alloys is an extremely interesting phenomena
since these materials are widely used in many industrial applications, while a fully developed and practically applicable predictive
physical model still does not exist [14–17]. A number of studies carried out in the world are an unambiguous confirmation of that
statement. Bearing in mind the multiple effects of hydrogen in metals, the specific mechanism of hydrogen embrittlement is man-
ifested, depending on numerous factors [17]. Still, there is no universal mechanism to simultaneously describe and take into ac-
count all forms of the HE [1,14–18].
Most often, a multidisciplinary approach based on the use of different advanced experimental-laboratory methods, theoretical
and quantum mechanical models, modern micro-and nano-research, micro-fracture mechanics, atomistic investigations and solid-
state physics, provides conditions for a much better understanding of extremely complex, multiple and often simultaneously ac-
tive mechanisms of the HE [15–18]. A recent study on the hydrogen embrittlement mechanism in iron indicates that a mechanism
map for the prediction of hydrogen embrittlement is a function of numerous parameters, including the loading rate, hydrogen
chemical potential, initial crack size, effective hydrogen diffusion activation enthalpy, temperature, and cleavage stress intensity
[14]. In numerous non hydride-forming metals, contemporary research has recognized the hydrogen-enhanced decohesion
(HEDE) and the hydrogen-enhanced localized plasticity (HELP) as two mechanisms responsible for the hydrogen embrittlement
[15,16,19,20]. Since over a wide range of hydrogen concentration the enhancement in the mobility of dislocations was detected
from very low to high, it can be concluded that the HELP mechanism might be active, but in varying degrees, depending on
the concentration of hydrogen in metal [21]. On the other hand, the prerequisite for the activation of the HEDE mechanism,
based on the hypothesis that interstitial hydrogen lowers the cohesive strength by dilatation of the atomic lattice, which results
in a very sudden and sharp ductile–brittle transition (ductility loss), is reaching sufficiently high – so called “critical” hydrogen
concentration in a metal [21,22].
 M.B. Djukic et al. / Engineering Failure Analysis 58 (2015) 485–498 487

The coexistence of different HE mechanisms and their simultaneous effects in metallic materials, including steels, is still not
well documented, while recognition of the dominant mechanism, one or more, is an extremely challenging and crucial problem
[23–25]. These two seemingly opposing hydrogen embrittlement mechanisms (HEDE and HELP) are extremely difficult to detect
at the same time. Activation of the particular mechanism of HE and the extent of its influence on the fracture process is primarily
caused by the successive processes at the micro and nano levels in the fracture process zone at the crack tip [19]. Taketomi et al.
[26] conducted their study applying the dislocation dynamics calculation, which clearly indicates that the possible change of hy-
drogen embrittlement mechanisms in alpha iron, from HELP to HEDE, depends on the boundary environmental and mechanical
conditions: hydrogen concentration and applied stress intensity rate. Recently, simultaneous action of the HEDE and HELP mech-
anisms were detected, depending on the local concentration of hydrogen in investigated low carbon steel [6]. However, the over-
view of the hydrogen embrittlement phenomena in steels under various conditions is still unclear.

2. Experimental concept and procedure

The experimental research was conducted in two distinctive phases: (1) a case study and failure analysis of the boiler evap-
orator tube sample, and (2) subsequent postmortem analysis of the viable hydrogen embrittlement mechanisms in investigated
St.20 steel unevenly enriched with hydrogen as a result of the development of intensive local hydrogenation of the tube metal
during boiler exploitation. The results of both experimental research phases are presented separately in this paper and also
some general and more specific conclusions about the boiler tube failure mechanism and subsequent research on the hydrogen
embrittlement mechanisms in St.20 steel.
The experiments in this study were performed on the specimens cut out from the real boiler evaporator tube sample (diam-
eter: D = 60 mm and nominal wall thickness: t = 6 mm) made of plain carbon steel, grade 20 – St.20 (or 20G, equivalent to AISI
1020) and damaged during boiler exploitation. The chemical composition of 20 steel specimens is given in Table 1.
Several different experimental methods and techniques were used to determine the boiler tube failure mechanism and after-
ward also the viable hydrogen embrittlement mechanisms in the investigated steel. Numerous tested specimens were cut out
from the boiler tube sample, Fig. 1a [2].
In order to determine and analyze the main causes of boiler tube fracture, traditional experimental methods were applied dur-
ing the first phase of the experimental research – (1) the boiler tube failure analysis.
The following experimental techniques were successively used during the failure analysis: visual examination, tube wall thick-
ness measurement, material chemical composition analysis, hardness measurement, tensile testing, metallographic investigation,
and analysis of the chemical composition of corrosion products and deposit on the tube inner side in the vicinity of the thick
wall “window” type fracture, Fig. 1a.
Measuring of the tube wall thickness, along the fracture edge, Fig. 1c, was done using a mechanical external measuring gauge,
Kroeplin D1R10.
Analysis of the chemical composition of St.20 steel was performed by mass spectroscopy using an ARL 31000 quantometer,
while for the hardness measurement (Vickers hardness HV5) at a different distance from the edge of the “window” type hydro-
gen, Fig. 1c, a stable Vickers hardness device, type HPO 250 VEB-WPM Leipzig, was used.
Tensile testing of specimens was carried out on a mechanical tensile machine AVK, type10t at ambient (20 °C) temperature,
according to EN 10002-1. Testing was performed on the total of three specimens (No. 1–3). Two specimens were prepared
from the fractured tube the subject of this study (No. 1–2). The first specimen (No. 1) is cut out from the evaporator tube
side, directly exposed to the boiler furnace (“fireside”), in the vicinity of the “window” type fracture, while the second specimen
(No. 2) is cut out from the other (“opposite”) side of the damaged boiler tube. For comparison purposes, one additional specimen
(No. 3) is also cut out from the new boiler tube which was out of service.
The metallographic investigation by optical microscopy was carried out on two specimens (ME1 and ME2) cut out in the vi-
cinity and at some distance (6 mm) from the “window” type fracture, using a microscope Olympus PME3-313UN, Fig. 1a.
The chemical composition of corrosion products and deposit on the tube inner side in the vicinity of the “window” type frac-
ture was analyzed by a Perkin–Elmer atomic absorption spectrometer (AAS), model AAnalyst 400. An acetylene–air flame was
used; the gas flow rates and the burner height were adjusted to obtain the maximum absorbance signal for each element to
be determined. The Zimmermann–Reinhardt (Z–R) permanganometric titration method was also used to determine iron in cor-
rosion products on the tube inner side. The iron is titrated with potassium permanganate (KMnO4) in order to determine the con-
tent of ferrous ions (Fe2+ and Fe3+).
During the second experimental phase – (2) subsequent-postmortem analysis of the viable hydrogen embrittlement mecha-
nisms in investigated St.20 steel, a special experimental concept was applied.

Table 1
Chemical composition of St.20 steel.

Element (wt%)

C Si Mn S P

0.24 0.28 0.48 0.025 0.013

488 M.B. Djukic et al. / Engineering Failure Analysis 58 (2015) 485–498

Fig. 1. Boiler tube, hydrogen damage: (a) the thick wall, “window” type fracture, metallographic specimens – ME1 and ME2; (b) position and designation of the
“Roman tile” Charpy specimens and hardness measurements positions in the vicinity of the fracture; (c) hardness measurements positions along a tube, x-distance
from the fracture edge.

The applied special experimental concept is based on the correlation of the material macro mechanical characteristics: impact
strength values (KCVTOT.) and its components of crack propagation (KCVP) and crack initiation (KCVI) obtained by impact tough-
ness testing, as well as material tensile properties and hardness to scanning electron microscopy – SEM fractography analysis of
the fracture surfaces of the Charpy specimens. Only SEM fractography of the fracture surfaces is not sufficient for understanding
the deformation and fracture mechanisms in steels enriched with hydrogen. These Charpy specimens were used to systematically
test the viability potential of the active hydrogen embrittlement mechanisms, as well as their possible simultaneous action.
Numerous Charpy specimens were cut out from the evaporator boiler tube sample in the vicinity of the “window” type hydro-
gen damage fracture at a different distance from the fracture edge. These specimens were cut out from the locally thinned zone of
the evaporator tube due to corrosion gouging, where the local corrosion and hydrogen damage occurred. The position and desig-
nation of the Charpy specimens (S1–S6) in the vicinity of the fracture are shown in Fig. 1b. The orientation of each specimen is
transverse to the axis of the tube, whereby the notches on the set of three specimens, on both sides of the fracture, are oriented
toward the fracture edge. The non-standard, “Roman tile”-like geometry [27,28], Charpy V-1 notched specimens (1 mm/45°) were
used (dimension: 3 × 3 × 44 mm and 3 × 6 × 44 mm), Fig. 1b. Before testing, specimens were flattened as necessary at extremely
slow strain rate. The use of this particular specimen geometry for impact strength testing is explained by the impossibility to pre-
pare the standard Charpy specimens, due to the low thickness and boiler tube important curvature. Hardness measurement po-
sitions on the tube outer surface for all six Charpy specimens are also marked in Fig. 1b.
The impact testing was performed in accordance with the EN 10045-1 i.e. ASTM E23-95 on the instrumented Charpy machine
Schenck Trebel 150 J at 20 °C. The total impact energy, as well as crack initiation and crack propagation energy were estimated.
Specimens for the SEM fractography examination of the fracture surfaces of the Charpy specimens were prepared in a standard
way. The fracture surface of the impact testing specimens was examined in order to identify the fracture mode and characteristic
changes in fracture features caused by changes of the dominant hydrogen embrittlement mechanism [18].
Fractography examination was carried out on a scanning electron microscope – SEM unit, type JEOL JSM-6460LV at different
magnifications, including the high one. In order to identify the chemical content, the energy-dispersive X-ray spectroscopy (EDS)
analysis of the ferrite matrix and characteristic phases were also performed. Elemental mapping and line profiles were acquired
from the SEM/EDS and the corrosion products on the tube inner side in the vicinity of the “window” type fracture were also in-
vestigated, Fig. 1a.
It should be noted that the term “critical hydrogen concentration” is often used in studies on the problems of hydrogen deg-
radation of metallic materials, although in different studies this term has different physical meanings [12]. When reaching a crit-
ical hydrogen concentration, different authors suggest that different phenomena occur in steel: changes in the damage
mechanism from decohesion to the formation of micro crevices at the grain boundaries [29]; local cracking initiation – establish-
ment of the local fracture criteria [30] and critical loss of the material local strength at the notch in the presence of hydrogen [27].
In this study, the term “critical hydrogen concentration” defines the concentration level which causes critical drop in the impact
strength of the material and sharp ductile–brittle fracture transition [6,12,21] in the presence of hydrogen. Also, such an
 M.B. Djukic et al. / Engineering Failure Analysis 58 (2015) 485–498 489

interpretation allows the establishment of the hydrogen embrittlement model based on simultaneous action of both HELP and
HEDE mechanisms in the investigated St.20 steel.

3. Results

3.1. Visual examination

The thick wall, “window” type fracture on the boiler tube fireside was visually examined in detail and macrographs taken in as
received conditions are shown in Figs. 1a and 2. Schematic representation of the “window” type fracture is shown in Fig. 2a. On
the outer surface of the tube specimen, the general corrosion and local deposit are observed. Thick, coarse and hard corrosion de-
posit and local area with a strong corrosion gouging were visible on the tube fireside inner surface, especially in the vicinity
downstream of the welded joint with defect. An uneven local wall thinning in the form of a strong corrosion gouging of the ir-
regular shape is also visible at some distance from the fracture edge.
It is also observed that the weld joint in the damaged zone has a poor weld overlay in the form of excessive root penetration
(up to 1.5 mm), Fig. 2b. Then, the fracture surfaces were observed to be very smooth indicating the brittle type failure with no
gross deformation.

3.2. Measurements of the tube wall thickness

Measurements positions along the tube fracture edge are shown in Fig. 3a, while the results of the wall thickness measure-
ments are shown in Fig. 3b.
Nominal evaporator tube thickness is 6 mm. The change in thickness along the fracture surface is very uneven. The thick wall,
“window” type fracture has an elliptical shape and extends along the tube’s longitudinal axis. The thickness in front of the weld
joint in relation to the direction of the fluid flow, on the side closer to the furnace wall membrane (positions 3–5, Fig. 3a), is in the
range 4.5–4.9 mm and close to the nominal thickness. The same trend can be also observed in the case of measuring positions
behind the weld joint, near the furnace wall membrane (positions 6–18). The tube wall thickness is in the range 4.3–4.8 mm,
Fig. 3b. A significant decrease in the tube thickness was observed at the positions 19–28, with the measured thickness in the
range 2.7–3.65 mm. These positions correspond to the tube inner surface area with a strong corrosion gouging, Fig. 2b.

3.3. Characterization of corrosion products

The corrosion products on the tube surface inner side, in the vicinity of the “window” type fracture, have been identified. The
corrosion products formed at various locations in the area with an extreme corrosion gouging (see Fig. 2b) were characterized by
energy-dispersive X-ray spectroscopy (EDS) and their morphology was also observed by scanning electron microscopy (SEM). The
atomic absorption spectroscopy (AAS) procedure was also used for the chemical elements’ quantitative determination, as well as
the Zimmermann–Reinhardt (Z–R) method to determine iron in corrosion products in the area with a thick deposit (see Fig. 2b).
Fig. 4a shows the SEM micrograph of the secondary corrosion pit (cavity) in the area with a corrosion gouging, near the fracture
surface. The results of chemical elements distribution (elemental mapping) in the corrosion products deposited within the

Fig. 2. Fractured boiler tube: (a) schematic representation of the “window” type fracture; (b) tube inner surface in the vicinity of the fracture.
490 M.B. Djukic et al. / Engineering Failure Analysis 58 (2015) 485–498

Fig. 3. Measurements of the tube wall thickness: (a) schematic representation of the measurement positions; (b) thickness distribution along the tube fracture
surface.

Fig. 4. SEM/EDS of the secondary corrosion pit (cavity) in the area with a corrosion gouging: (a) micrograph; (b) measurement positions – line profile along the
pit; (c) distribution of chemical elements.
 M.B. Djukic et al. / Engineering Failure Analysis 58 (2015) 485–498 491

secondary corrosion pit and line profile along the pit acquired by the SEM/EDS with measurement positions are shown in
Fig. 4b and c.
The results of chemical elements’ quantitative determination in the deposit by the AAS and the content of ferrous ions (Fe2+
and Fe3+) by the Z–R method are shown in Table 2.

3.4. Metallographic investigation

Optical micrograph of specimens ME1 cut out at some distance (6 mm) from the “window” type fracture and ME2 cut out in
the vicinity of the fracture edge (see Fig. 1a) are shown in Fig. 5.
The St.20 steel has ferrite–pearlite microstructure, 6 mm away from the fracture edge, specimen ME1, as seen in Fig. 5a. Also, a
distinctly banded structure with nonhomogeneous distribution of very fine elongated MnS inclusion is visible. The evaporator
tube was made of plain carbon steel, grade 20 – St.20, whose initial microstructure consisted primarily of ferrite and pearlite,
with ∼25% of pearlite. The microstructure of the ME1 specimen after 73,000 h of operation corresponds to the initial microstruc-
ture of St.20 steel.
The microstructure of the investigated sample ME2 in the vicinity of the fracture edge is pure ferrite and decarburized with a
complete decomposition of pearlite, as seen in Fig. 5b. Also, numerous micro-fissures along ferrite grain boundaries and intergran-
ular (IG) microcracks are visible.

3.5. Mechanical properties

3.5.1. Tensile test and embrittlement index

The tensile test properties of specimens with standard recommended values for plain carbon steel, grade 20 – St.20 (or 20G,
equivalent to AISI 1020) are given in Table 3.
During studies of the hydrogen embrittlement, the elongation and reduction of the area are often used as the evaluation
criteria, due to the well-established knowledge of hydrogen effect on the material ductility [31]. The degree of hydrogen embrit-
tlement Iδ5 was evaluated on the basis of the decrease in elongation, which was determined from the formulas:

δ5new  δ5H
Iδ5 ¼  100% ð1Þ
δ5new

where Iδ5 is the embrittlement index, δ5new is the elongation of the specimen cut out from the new, unused boiler tube, with the
same dimension and made of the same material (St.20) as a fractured tube and δ5H is the elongation of the specimen cut out from
the fractured boiler tube. The higher Iδ5, the more reduction in material ductility occurs due to the local hydrogenation and em-
brittlement of metal.

3.5.2. Hardness measurements

The degree of hydrogen embrittlement of the material in the vicinity of the fracture edge was also evaluated on the basis of
decrease in the hardness with increasing distance (x (mm)) from the edge of the “window” type hydrogen damage fracture (see
Fig. 1c), as shown in Fig. 6.
Hardness variation as a function of the distance from the fracture edge clearly indicates a substantial increase in hardness in
the vicinity of the fracture characterized by very uneven distribution. Hardness value in the vicinity of the opening is very high
(206HV5) and at a distance of ∼3 mm from the fracture it remained significantly increased (185HV5) compared to the maximum
acceptable standard hardness value (max. 162HV5) for St.20 steel.
Also, it was observed that in a relatively narrow zone at a distance 3.7–5.5 mm from the fracture edge, the hardness remains
elevated and constant (∼175HV5). Here, it can be seen that the slope of the hardness-distance curve significantly changes before
and after the “plateau”, characterized by relatively uniform hardness values. Near the fracture (up to 3.7 mm), the slope of the
hardness curve is significantly higher compared to that at a distance from the opening higher than 5.5 mm. The mean hardness
values of the six Charpy specimens (see Fig. 1b) are given in Table 4.

3.6. The impact testing properties and SEM fractography analysis

Variation of the KCVTOT., KCVP and KCVI for all six Charpy specimens as a function of the mean specimen hardness – hydrogen
concentration (see Table 2) is shown in Fig. 7. The position of all specimens is shown in Fig. 1b.

Table 2
AAS and Z–R analysis of deposit, shown in Fig. 2b.

Element (wt%)

Fe Ni Cu Mn Cr Si Zn Ca Fe2+ Fe3+

72.2 0.28 22.8 1.22 1.1 0.06 0.23 1.2 3.16 71.3
492 M.B. Djukic et al. / Engineering Failure Analysis 58 (2015) 485–498

Fig. 5. Optical micrograph of specimens: (a) specimen ME1: ferrite–pearlite, elongated MnS inclusion; (b) specimen ME2: complete decarburization, ferrite, micro-
fissures and IG micro cracks.

The SEM micrograph of the fracture surface in the Charpy specimen S5, at an approximate distance of 3 mm from the fracture
edge (left side), is shown in Fig. 8a. It can be generally described as a “quasi-cleavage”, based on surface appearance. The fracture
modes reflected by the fracture surfaces are neither intergranular fracture nor cleavage fracture; however they are showing the
signs of plasticity.
Based on careful analysis of the fracture surfaces with very fine features (magnification 3000×), two distinctive fracture fea-
tures can be distinguished: areas of transgranular (“cleavage-like”) fracture, marked as TG in Fig. 8b, and very fine microvoid co-
alescence (b2 μm) with relatively poorly defined dimples, marked as MVC in Fig. 8b. The transgranular fracture (TG) zone is
observed in the vicinity of elongated manganese sulfide – MnS inclusions, strong hydrogen trapping sites [10], present in the mi-
crostructure of St.20 steel, as shown in Fig. 5a. Also, a distinctly banded structure with nonhomogeneous distribution of very fine
elongated MnS inclusion was visible.
Nevertheless, in the case of this specimen (S5) a significant drop in the properties of plasticity (KCVTOT., KCVP and KCVI) was
not detected (see Fig. 7 – specimen S5), while the mean hardness is relatively low: 157HV5 (see Table 4).
The SEM micrograph of fracture surfaces of the Charpy specimen S3 at a distance of 9 mm from the fracture edge – right side
and S4 in the vicinity of the fracture edge – left side are shown in Fig. 9a and b, respectively. The mean hardness values of these
two specimens with slightly elevated mean hardness are very close: 164HV5 (S3) and 166HV5 (S4), as shown in Table 4.
The general fracture appearance for both specimens (S3 and S4) is characterized by a distinctive mixed fracture mode with the
simultaneous presence of brittle transgranular – TG fracture features of ferrite (confirmed by EDS analysis and not shown here),
and locally ductile fine MVC fracture feature of pearlitic microconstituent.
In the case of specimen S3, a negligible drop in the impact strength (KCVTOT.) and its component of crack propagation energy
(KCVP) and steady value of crack initiation component (KCVI) were observed, as shown in Fig. 7 (specimen S3). Also, for specimen
S3 additional significant drop in the properties of plasticity (KCVTOT., KCVP and KCVI), in comparison with specimen S5, was not
detected.
In the case of specimen S4, a sharp ductile–brittle fracture transition occurs, as shown in Fig. 7 (specimen S4). A sudden drop
of plasticity is accompanied by a sharp drop in KCVTOT (reduced by a factor of two) and especially KCVP (reduced by a factor of
five), with a negligible increase in the KCVI (see Fig. 7 – specimen S4) compared to previously analyzed specimen S3.
The SEM micrograph of the Charpy specimen S2 fracture surface, with an elevated mean hardness: 183HV5 (see Table 4) and
at a distance of 3 mm from the fracture edge – right side is shown in Fig. 10.
The general fracture appearance is characterized by the occurrence of the brittle intergranular (IG) area on the fracture surface
as well as voids (b10 μm), fissures and micro cracks along the ferrite grains. Also, a small remaining area with local plasticity, frac-
tured by the MVC (1–2 μm) can be observed.
Besides, there is a significant additional drop in both KCVI (reduced by a factor of two) and KCVTOT. (reduced by a factor of
1.5), compared with specimen S4, as shown in Fig. 7 (see specimens S4 and S2).

Table 3
Mechanical properties of specimens.

Specimen Yield strength σs (MPa) Tensile strength σb (MPa) Elongation δ5 (%) Embrittlement index Iδ5 (%)

No. 1 – damaged tube, “fireside” 403 481 13.2 57.8

No. 2 – damaged tube, “opposite” 323 460 26.3 15.9

No. 3 – new tube 312 471 31.3 –

 M.B. Djukic et al. / Engineering Failure Analysis 58 (2015) 485–498 493

Fig. 6. Variation of the Vickers hardness (HV5) as a function of the distance (x) from the edge of the fracture (see Fig. 1c).

4. Discussion

4.1. Failure analysis of boiler tube

One of the key factors – root causes, responsible for the occurrence of boiler tube hydrogen damage – high temperature hy-
drogen attack (HTHA) is the presence of poorly executed weld overlay which penetrates the tube wall to the inside surface [2,32],
Fig. 2. This type of weld geometry-contour defect, excessive root penetration, is undesirable because the excess material on the
tube inner side can disrupt two-phase fluid flow, leading to the destabilization of fluid boiling/flowing regime, followed by the
appearance of micro-departure from nucleate boiling (micro-DNB) of two-phase fluid, downstream of the weld defect [2]. This
phenomenon, micro-DNB, also provoked local increase in the concentration of acid components in a working fluid in the vicinity
of tube metal, the intensification of local deposit precipitation rate on metal surface, tube metal thermal cycling and, finally, the
destruction of protective magnetite layer [2,33]. As a result, localized underdeposit corrosion in the form of hydrogen-induced cor-
rosion process was initiated, especially downstream of the weld with defect [32,33]. Visual examination showed that thick, coarse
and hard corrosion deposit and local area with a strong corrosion gouging were visible on the tube inner surface, especially in the
vicinity – downstream of the welded joint, Fig. 2b.
The analysis of chemical composition of the corrosion products and deposit on the tube inner side in the vicinity of the “win-
dow” type fracture provided valid data on the corrosion mechanisms, Fig. 4 and Table 2. The corrosion process was actively de-
veloping locally on the tube surface inner side in the vicinity of the “window” type fracture. The EDS results for the positions near
the corrosion pit (spectrum No. 1, 4 and 5 in Fig. 4c) indicated the iron-oxide forming elements (Fe: 76.18–82.5% and O:
11.42–11.59%) as the main components. In addition to the iron and oxygen, a high content of chlorine, up to 7.4%, and calcium,
up to 1.2 mm (in deposit, see Table 2), which provoked acid corrosion, as well as copper, up to 23.14%, responsible for the elec-
trochemical galvanic corrosion, were also detected in the deposit and corrosion pit. However, it should be noted that the corrosion
process at the bottom of a corrosion pit is more intense: sudden increase and extremely high chlorine content (see spectrum No.
8 in Fig. 4c) and high copper content (spectrum No. 7 and 11), compared to the areas near the pit (spectrum No. 1–5). Besides the
obvious dominant acid corrosion, due to chlorine with an additional effect of copper, there is also intense transition of iron oxide
from desirable-protective magnetite (Fe3O4) to unprotective hematite (Fe2O3) type (see low Fe content and high content of O,
spectrum No. 8 and 11 in Fig. 4c). Based on the results of the Z–R method, it is observed that the content of Fe2+ ions is signif-
icantly lower compared to Fe3+ ions, Table 2. This indicates the formation of a defective-multilayer magnetite and hematite as a
result of under deposit corrosion. The decrease in the stability of magnetite and increasing in susceptibility to oxidation is also a
consequence of the presence of sodium (not shown here), detected in the deposit [34].
Excessive deposits of corrosion products themselves are not sufficient for the development of hydrogen damage of the boiler
evaporator tube in the form of high temperature hydrogen attack. The second condition suggested by Dooley and Bursik [32] is
the presence of acidic contaminants – chlorine and their concentration underneath the deposits, just as in this case study. The
hydrogen formed during acidic corrosion in aqueous medium beneath the defective-multilayer magnetite oxide layer covered

Table 4
Mean hardness of the Charpy specimens.

Charpy specimen S1 S2 S3 S4 S5 S6

Mean hardness (HV5) 198 183 164 166 157 152

494 M.B. Djukic et al. / Engineering Failure Analysis 58 (2015) 485–498

Fig. 7. Variation of the impact strength (KCVTOT.) and its crack propagation component (KCVP) and crack initiation component (KCVI) of the Charpy specimens
(S1–S6) as a function of the specimen hardness (hydrogen concentration).

with deposits freely diffused into the tube surface layer and caused high temperature hydrogen attack (HTHA) damage of the boil-
er tube in the form of a characteristic blow out window fracture according to a well-defined mechanism [2–5,21,22,32,33].
Metallographic examinations have confirmed that the local HTHA is the responsible failure mechanism and the main cause of
the boiler tube thick wall, “window” type fracture. In Fig. 5b the advanced stage of HTHA is visible and it is characterized by com-
plete decarburization, a group of micro-fissures joining together to form micro cracks. The observed decarburization, as well as the
complete degradation of pearlite indicates that the hydrogen penetration occurred through the tube wall. Many micro-fissures
and intergranular cracks are the result of the pressure effect of methane molecules generated during the HTHA [32,33] and pre-
cipitated along the ferrite grain boundaries.
A detailed metallographic examination through the tube wall thickness in the vicinity of the damage (inner and outer tube
surface) and at a different distance from the fracture edge was also carried out and not shown here. Close to the tube inner sur-
face in contact with a source of hydrogen - the underdeposit corrosion mechanism, there is some, to a varying degree, degrada-
tion of the microstructure due to the HTHA (complete or partial decarburization, pure ferite, micro-fissures), but only in the
vicinity and very close to the “window” type fracture edge, up to 3 mm (see Figs. 5b and 10). On the tube outer surface, even
in the vicinity of hydrogen damage (see Fig. 1c), where hardness was measured (see Fig. 6), the microstructure corresponds to
the initial i.e. ferritic–pearlitic, as shown in Fig. 5a. Therefore, the influence of the microstructural degradation due to HTHA during
boiler exploitation on the measured hardness change is negligible. As a result of the uneven and local enrichment of metal with
hydrogen, due to the activity of the HTHA process during boiler tube service, hydrogen embrittlement of the tube metal has ap-
peared afterward in a varying degree.

Fig. 8. SEM micrograph of the Charpy specimen S5: (a) magnification 450×; (b) detail, magnification 3000×.
 M.B. Djukic et al. / Engineering Failure Analysis 58 (2015) 485–498 495

Fig. 9. SEM micrograph of the Charpy specimens: (a) specimen S3 (HELP > HEDE); (b) specimen S4 (HEDE > HELP).

4.2. Hydrogen embrittlement mechanisms and mechanical properties

Tensile testing results for the material in the vicinity of the “window” type hydrogen damage fracture on the tube “fireside”
(see Fig. 2a), specimen No. 1, indicated a noticeable material embrittlement, Table 3. Yield strength (403 MPa) was significantly
higher than the minimum standard value for St.20 (216 MPa according to the GOST 1050-88 standard) or values for specimens
No. 2 and 3 (323 MPa; 312 MPa), and it is close to the minimum standard recommended value for tensile strength (min.
420 MPa). Tensile strength is 481 MPa, while deformation characteristic, expressed by elongation (δ5 = 13.8%), is much lower
than the standard recommended one (δ5 = 24%) for St.20 steel, which is an indication of a significant increase of material brit-
tleness. In addition, a very high embrittlement index (Iδ5 = 57.8%) for specimen No. 1 clearly indicates local significant drop of
the boiler evaporator tube material ductility characteristics, in the vicinity of the “window” type fracture, due to a hydrogen en-
richment during boiler exploitation.
The effect of hydrogen on the hardness of metallic materials have been reported by numerous researchers. They found that
cathodic hydrogen charging produced major surface hardening. In general, hydrogen charging increases hardness of metallic ma-
terials (plain carbon steels, austenitic stainless steels, stainless steels and Ti alloy) [3,35–37], but for some alloys a limited effect on
hardness, or even opposite effect: H-induced softening, was observed [38].
The hardness change trend, as a function of the distance from the edge of the fracture, showed “the plateau appearance” and
also a significant change in the slope of the hardness curve, Fig. 6. This indicates the material brittleness and change in the dom-
inant hydrogen embrittlement mechanism, which was confirmed by impact toughness testing and fractographic studies of the
fractured surface in the Charpy specimens [6]. The abrupt change in mechanical properties is a function of the hydrogen content
in the metal, whereby it appears in exceeding the “critical hydrogen concentration” [6,12,21,27].
The transition from ductile to brittle failure in steels due to hydrogen remains a phenomenon that is still not sufficiently ex-
plored and clarified either. In the world literature, there is disagreement and polarization of opinions about the inherent nature,
trigger mechanism and hydrogen embrittlement mechanisms responsible for the transgranular, microvoid coalescence, cleavage

Fig. 10. SEM micrograph of the Charpy specimen S2 (HEDE + HTHA).

496 M.B. Djukic et al. / Engineering Failure Analysis 58 (2015) 485–498

and quasi-cleavage fracture due to hydrogen [6]. The transition from one mode to another is still a matter of debate
[6,18,25,39,40].
The general fracture appearance for specimens S3–S5, Figs. 8. and 9, is arguably similar to plasticity-related hydrogen-induced
cracking (PRHIC) [41]. Over recent decades the “quasi-cleavage” has been used to describe any fracture surface appearance that
cannot otherwise be explained as either intergranular (IG), microvoid coalescence (MVC) or true cleavage. Ultimately, the
hydrogen-assisted fracture mechanisms leading to these imprecisely defined fracture surface appearances are still unknown
[42]. It is also important to note that the emergence of dominant macro brittle fracture (TG and IG) of the material enriched
with hydrogen, determined by the choice of experimental parameters, usually precludes the possibilities of proper detection of
the phenomenon of hydrogen-assisted micro-local plasticity i.e. simultaneous effects of both HEDE and HELP mechanisms.
Simultaneous action of the HELP and HEDE mechanisms (HELP + HEDE) in St.20 steel is characterized by a distinctive mixed
fracture mode with the simultaneous presence of brittle transgranular – TG fracture features of ferrite, predominantly by the
HEDE mechanism, and locally ductile fine MVC fracture features of pearlitic microconstituent due to the HELP mechanism, as
shown in Fig. 9. At lower hydrogen concentration (lower hardness), the HELP mechanism is dominant (HELP > HEDE), which
is manifested by an increase in ductile MVC fracture features in the “quasi-cleavage”-like fracture surface of specimen S3,
MVC ≫ TG, as shown in Fig. 9a. The initial dominant HELP activity in specimen S3 is followed by a negligible drop in the impact
strength (KCVTOT.) and its component of crack propagation energy (KCVP) and steady value of crack initiation component (KCVI),
as shown in Fig. 7, and decreases with increasing in hydrogen concentration (hardness).
On the other hand, prevailing transgranular fracture features – TG of ferrite of specimen S4, without obvious traces of plastic-
ity, TG ≫ MVC, Fig. 9b, followed by a sharp drop in KCVTOT. (reduced by a factor of two) and especially KCVP (reduced by a factor
of five), with a negligible increase in the KCVI (see Fig. 7 – specimen S4) are a consequence of increased activity of the HEDE
mechanism (HEDE > HELP) i.e. hydrogen-induced decohesion, with increasing in hydrogen concentration. The expressed obser-
vations are consistent with a model of the simultaneous effects of the HELP and HEDE [23,26]. Simultaneous appearance of the
HELP and HEDE mechanisms is further supported in the models proposed by Jokl et al. [43], Gerberich et al. [23] and Djukic
et al. [6].
The occurrence of the brittle intergranular (IG) area on the fracture surface of specimen S2 with high hardness: 183HV5 and
hydrogen concentration is associated not only with the HEDE action and reduction of the HELP mechanism effects, but also as a
consequence of the high temperature hydrogen attack (HTHA) mechanism activity during boiler tube exploitation (HEDE
+ HTHA), Fig. 10. The HTHA effects initially occurs in the areas where the concentration of individual methane void (b10 μm),
due to local HTHA is significantly increased, which is reflected in an additional increase in the material brittleness. The HTHA ef-
fect is visible only in the vicinity of the fracture edge (specimen S2). In addition to the formations of the voids, fissures and micro
cracks along the ferrite grains (complete degradation of pearlite microconstituent during exploitation and HTHA activity), there is
also a significant additional drop in both KCVI (reduced by a factor of two) and KCVTOT. (reduced by a factor of 1.5) compared to
specimen S4, Fig. 7.

5. Conclusions

Based on the comprehensive failure analysis of a failed boiler tube, the following conclusions can be drawn about the cause
and mechanism of failure in the boiler tube made of plain carbon steel, grade 20 – St.20 (or 20G, equivalent to AISI 1020).
(1) The local and very intensive underdeposit acidic corrosion – a source of hydrogen, in the form of a local corrosion gouging,
is mainly provoked by chlorine with simultaneous action of an electrochemical galvanic corrosion due to copper, in the
presence of organic matter in the boiler water. The local metal enrichment with hydrogen is the result of localized
underdeposit corrosion of the tube metal during boiler operation. High content of carbon and chlorine in deposits is the
result of an undesirable enrichment of boiler water with organic matters and acidic ingredients during exploitation. The
development of all these processes, including hydrogen-induced corrosion were also enhanced as a result of thermal cy-
cling of the tube metal during operation: frequent starts and stops of the boiler and common temperature fluctuations dur-
ing boiler operation.
(2) The resulting local metal enrichment with hydrogen on the boiler evaporator tube inner surface, in the vicinity of welded
joint with excessive root penetration, provoked hydrogen damage in the form of the “window” type fracture generated by
the high temperature hydrogen attack (HTHA) damage mechanism activity during boiler exploitation.
(3) As a result of comprehensive investigations it was found that the local initiation and growth of the HTHA damage and fail-
ure mechanism was mainly provoked by the weld geometry-contour defect; excessive root penetration – root cause of fail-
ure and accompanying undesirable processes. The HTHA damage mechanism, as a primary cause of boiler tube fracture,
was also accompanied by postmortem uneven tube metal enrichment with hydrogen in the vicinity of the “window
type fracture”. As a result, the local occurrence of the tube metal hydrogen embrittlement in a varying degree was detected
in the vicinity of the HTHA damage. The hydrogen embrittlement degree was evaluated on the basis of the hydrogen em-
brittlement index Iδ5 and the decrease in the hardness with an increasing distance from the HTHA “window” type fracture
edge.
The testing procedure of the boiler tube sample, unevenly enriched with hydrogen and damaged due to the HTHA mechanism
during actual boiler operation, as well as the selected experimental concept, were specially designed to credibly investigate the
possible mechanisms of hydrogen embrittlement and their possible coexistence in this plain carbon steel, grade 20 – St.20 (or
 M.B. Djukic et al. / Engineering Failure Analysis 58 (2015) 485–498 497

20G, equivalent to AISI 1020). The fracture surfaces of the Charpy specimens, in situ (during boiler tube exploitation) enriched
and embrittled by hydrogen to a different degree, were examined by SEM in order to identify the fracture mode and characteristic
changes in fracture features due to changes of the dominant mechanism in hydrogen embrittlement. This approach provides the
possibilities to define the degree of activity and influence of individual HE micromechanism, as well as correlation, assessment
and analysis of their individual and simultaneous effects on the macro mechanical characteristics (KCVTOT., KCVP, KCVI and hard-
ness) of investigated St.20 steel. The principal observations about active hydrogen embrittlement mechanisms and their effects on
the mechanical properties of grade 20 – St.20 steel (or 20G, equivalent to AISI 1020) are:

(1) Simultaneous action of the hydrogen-enhanced localized plasticity (HELP) and hydrogen-enhanced decohesion (HEDE)
were detected and confirmed to be active, depending on the local concentration of hydrogen in the investigated steel.
(2) Despite the fact that hydrogen concentration was not determined in the vicinity of the “window” type HTHA fracture, the
results in the present study show that the macro hardness value could be very well correlated with the hydrogen concen-
tration in the metal, and that it is growing with an increase of hydrogen concentration, regardless of the type of embrittle-
ment mechanism (HELP or HEDE) being dominant.
(3) A coexistence of the HE mechanisms (HELP + HEDE) is characterized by a distinctive mixed, “quasi-cleavage”-like fracture
mode, with the simultaneous presence to a varying degree of the brittle transgranular (TG) fracture features of ferrite, due
to prevailing action of the HEDE mechanism and locally ductile, fine microvoid coalescence (MVC) fracture feature of pearl-
itic microconstituent in account of HELP mechanism activity.
(4) At the lower hydrogen concentration, the HELP mechanism is dominant, which is manifested by an increase in ductile MVC
fracture features in the “quasi-cleavage”-like fracture surface, MVC ≫ TG. There is no significant decrease in the ductility of
the material, despite a slight increase in hardness.
(5) The prevailing TG fracture features, without obvious traces of plasticity, TG ≫ MVC, followed by a sharp drop in the impact
strength (KCVTOT.), and especially crack propagation component (KCVP), with a negligible change in the crack initiation
component KCVI value is a consequence of increased activity and dominance of the HEDE mechanism, with increasing in
hydrogen concentration.
(6) Relatively sudden HEDE mechanism manifestation on reaching a critical value of hydrogen concentration is followed by a
sharp ductile–brittle fracture transition. This drop in material ductility is not only due to the cohesive strength decline,
which may not be significant, but also due to a noticeable reduction in the appearance of hydrogen-enhanced local plastic-
ity. The transition from the dominant HELP to the dominant HEDE mechanism is not followed by a significant change of
material hardness.
(7) Simultaneous action of both HE mechanisms (HELP + HEDE) is responsible for the decline in steel ductility. Despite the
fact that critical hydrogen concentration was not determined in this study, the results clearly indicate that the degree of
decline in the material impact strength and macro hardness values can be successfully correlated with the hydrogen con-
centration in metal. The impact strength of steel decreases with an increase of hydrogen concentration, regardless of what
type of the embrittlement mechanism is dominant.
(8) The special experimental approach applied and presented in this paper has provided a valid detection of both HELP and
HEDE mechanisms and assessment of their effects on the macroscopic mechanical properties and ductile–brittle fracture
transition of the tested steel. However, the overview of the hydrogen embrittlement in steels under various conditions is
still unclear.
(9) The coexistence of different hydrogen embrittlement mechanisms and their simultaneous action in steels is still not well
documented, while recognition of a dominant mechanism is an extremely challenging and crucial task.
(10) The proposed hydrogen embrittlement model is practical for use as a predictive maintenance in thermal power plants,
since it is based on the use of standard macro mechanical tests. The model is based on the correlation of mechanical prop-
erties to the scanning electron microscopy fractography analysis of fracture surfaces for the case of simultaneously active
hydrogen embrittlement micro-mechanisms (HELP + HEDE). The development of the model is to follow in future
researches.

References

[1] Gangloff RP. Science-based prognosis to manage structural alloy performance in hydrogen. In: Somerday B, Sofronis P, Jones R, editors. Effects of hydrogen on
materials: proceedings of the 2008 international hydrogen conference. Jackson Lake; September 7–10 2008.
[2] Djukic M, Sijacki Zeravcic V, Bakic G, Milanovic D, Andjelic B. Model of influencing factors for hydrogen damages of boiler evaporator tubes. In: Proceedings of the
11th International Conference on Fracture 2005 (ICF11), March 20–25; Turin (Italy), vol. 6. Red Hook (NY): Curran Associates, Inc.; 2010. p. 3998–4003.
[3] J. Woodtli, R. Kieselbach, Damage due to hydrogen embrittlement and stress corrosion cracking, Eng Fail Anal 7 (2000) 427–450.
[4] T.J. Carter, L.A. Cornish, Hydrogen in metals, Eng Fail Anal 8 (2001) 113–121.
[5] R.K. Dayal, N. Parvathavarthini, Hydrogen embrittlement in power plant steels, Sadhana 28 (2003) 431–451.
[6] M.B. Djukic, V. Sijacki Zeravcic, G. Bakic, A. Sedmak, B. Rajicic, Hydrogen embrittlement of low carbon structural steel, Procedia Mater Sci 3 (2014) 1167–1172.
[7] R.N. Iyer, H.W. Pickering, Mechanism and kinetics of electrochemical hydrogen entry and degradation of metallic systems, Annu Rev Mater Sci 20 (1990)
299–338.
[8] R.P. Gangloff, Critical issues in hydrogen assisted cracking of structural alloys, Elsevier Science, Oxford, 2005.
[9] A. Pundt, R. Kirchheim, Hydrogen in metals: microstructural aspects, Annu Rev Mater Res 36 (1) (2006) 555–608.
[10] S. Myers, M. Baskes, H. Birnbaum, J. Corbett, G. DeLeo, S.K. Estreicher, et al., Hydrogen interactions with defects in crystalline solids, Rev Mod Phys 64 (1992)
559–617.
[11] C. Borchers, T. Michler, A. Pundt, Effect of hydrogen on the mechanical properties of stainless steels, Adv Eng Mater 10 (2008) 11–23.
[12] B.A. Kolachev, Hydrogen in metals and alloys, Met Sci Heat Treat 41 (3–4) (1999) 93–100.
498 M.B. Djukic et al. / Engineering Failure Analysis 58 (2015) 485–498

[13] O.D. Smiyan, G.M. Grigorenkoa, A.B. Vainman, Effect of hydrogen on corrosion damage of metal of the high-pressure energetic boiler drum, Int J Hydrogen Energy
27 (7–8) (2002) 801–812.
[14] J. Song, W.A. Curtin, Atomic mechanism and prediction of hydrogen embrittlement in iron, Nat Mater 12 (2013) 145–151.
[15] H.K. Birnbaum, Hydrogen effects on deformation and fracture: science and sociology, MRS Bull 28 (7) (2003) 479–485.
[16] M. Dadfarnia, P. Novak, D.C. Ahn, J.B. Liu, P. Sofronis, D.D. Johnson, et al., Recent advances in the study of structural materials compatibility with hydrogen, Adv
Mater 22 (10) (2010) 1128–1135.
[17] A. Barnoush, H. Vehoff, Recent developments in the study of hydrogen embrittlement: hydrogen effect on dislocation nucleation, Acta Mater 58 (16) (2010)
5274–5285.
[18] S. Lynch, Hydrogen embrittlement phenomena and mechanisms, Corros Rev 30 (3–4) (2012) 105–123.
[19] Y. Katz, N. Tymiak, W.W. Gerberich, Nanomechanical probes as new approaches to hydrogen/deformation interaction studies, Eng Fract Mech 68 (2001)
619–646.
[20] W.W. Gerberich, P.G. Marsh, J.W. Hoehn, Hydrogen induced cracking mechanisms – are there critical experiments? in: N.R. Moody, J.W. Hoehn (Eds.), Hydrogen
effects in materials, TMS, Warrendale (PA), 1996.
[21] D.F. Teter, I.M. Robertson, H.K. Birnbaum, The effects of hydrogen on the deformation and fracture of β-titanium, Acta Mater 49 (2001) 4313–4320.
[22] R.A. Oriani, P.H. Josephic, Equilibrium aspects of hydrogen-induced cracking of steels, Acta Metall 22 (9) (1974) 1065–1074.
[23] Gerberich WW, Stauffer DD, Sofronis P. A coexistent view of hydrogen effects on mechanical behavior of crystals: HELP and HEDE effects of hydrogen on mate-
rials. In: Somerday B, Sofronis P, Jones R, editors. Effects of hydrogen on materials: proceedings of the 2008 international hydrogen conference, September 7–10;
Jackson Lake Lodge, Grand Teton National Park, Wyoming, USA. Ohio: ASM International; 2009. p. 38–45.
[24] P. Novak, R. Yuan, B.P. Somerday, P. Sofronis, R.O. Ritchie, A statistical, physical based, micro-mechanical model of hydrogen-induced intergranular fracture in
steel, J Mech Phys Solids 58 (2) (2010) 206–226.
[25] M. Koyama, C.C. Tasan, E. Akiyama, K. Tsuzaki, D. Raabe, Hydrogen-assisted decohesion and localized plasticity in dual-phase steel, Acta Mater 70 (2014)
174–187.
[26] S. Taketomi, H. Imanishi, R. Matsumoto, N. Miyazaki, Dislocation dynamics analysis of hydrogen embrittlement in alpha iron based on atomistic investigations,
Proceedings of the 13th international conference on fracture 2013 (ICF13), June 16–21; Beijing (China), vol. 7, Curran Associates, Inc., Red Hook (NY) 2013,
pp. 5721–5729.
[27] J. Capelle, J. Gilgert, I. Dmytrakh, G. Pluvinage, Sensitivity of pipelines with steel API X52 to hydrogen embrittlement, Int J Hydrogen Energy 33 (2008) 7630–7641.
[28] H. Adib-Ramezani, J. Jeong, G. Pluvinage, Structural integrity evaluation of X52 gas pipes subjected to external corrosion defects using the SINTAP procedure, Int J
Press Vessels Pip 83 (2006) 420–432.
[29] E. Lunarska, Y. Ososkov, Y. Jagodzinsky, Correlation between critical hydrogen concentration and hydrogen damage of pipeline steel, Int J Hydrogen Energy 22 (2/
3) (1997) 279–284.
[30] J. Lufrano, P. Sofronis, Enhanced hydrogen concentrations ahead of rounded notches and cracks – competition between plastic strain and hydrostatic stress, Acta
Mater 46 (5) (1998) 1519–1526.
[31] H. Xu, X. Xia, L. Hua, Y. Sun, Y. Dai, Evaluation of hydrogen embrittlement susceptibility of temper embrittled 2.25Cr–1Mo steel by SSRT method, Eng Fail Anal 19
(2012) 43–50.
[32] R.B. Dooley, A. Bursik, Hydrogen damage, PPChem 101 – Boiler HRSG Tube Fail 12 (2) (2010) 122–127.
[33] Djukic M, Sijacki Zeravcic V, Bakic G, Rajicic B, Andjelic B. Weld geometry defect influence on boiler tube structural integrity. In: IIW international congress
“welding and joining technologies for a sustainable development and environment”, the 1st south-east European welding congress, proceedings of the 2006
IIW international congress; May 24–26, vol. 3. Timisoara (Romania): The International Institute of Welding (IIW); 2006. p. 169–79.
[34] E. Lunarska, K. Nikiforov, Evaluation of hydrogen degradation of water-wall pipes by the surface-layer properties of magnetite, Mater Sci 38 (4) (2002) 550–559.
[35] Bahr DF, Overman NR, San Marchi CW, Nibur KA. Orientation specific mechanical assessment of hydrogen precharged stainless steels using nanoindentation. In:
Somerday B, Sofronis P, Jones R, editors. Effects of hydrogen on materials: proceedings of the 2008 international hydrogen conference; September 7–10 2008;
Jackson Lake Lodge, Grand Teton National Park, Wyoming, USA. Ohio: ASM International; 2009. p. 179–85.
[36] J. Zhao, H. Ding, W. Zhao, X. Tian, H. Hou, Y. Wang, Influence of hydrogenation on microstructures and microhardness of Ti6Al4V alloy, Trans Nonferrous Met Soc
China 18 (3) (2008) 506–511.
[37] W. Godoia, N.K. Kuromotoa, A.S. Guimarãesb, C.M. Lepienskia, Effect of the hydrogen outgassing time on the hardness of austenitic stainless steels welds, Mater
Sci Eng, A 354 (1–2) (2003) 251–256.
[38] Y.S. Kim, D.W. Kim, S.S. Kim, W.J. Nam, H. Choe, Effects of hydrogen diffusion on the mechanical properties of austenite 316L steel at ambient temperature, Mater
Trans 52 (2011) 507–513.
[39] M.L. Martin, J.A. Fenske, G.S. Liu, P. Sofronis, I.M. Robertson, P. Burgers, On the formation and nature of quasi-cleavage fracture surfaces in hydrogen embrittled
steels, Acta Mater 59 (2011) 1601–1606.
[40] M.L. Martin, I.M. Robertson, P. Sofronis, Interpreting hydrogen-induced fracture surfaces in terms of deformation processes: a new approach, Acta Mater 59
(2011) 3680–3687.
[41] McMahon Jr CJ, Liu X, Kameda J, Morgan MJ. Recent observation of hydrogen-induced cracking of high-strength steels. In: Somerday B, Sofronis P, Jones R, editors.
Effects of hydrogen on materials: proceedings of the 2008 international hydrogen conference; September 7–10 2008; Jackson Lake Lodge, Grand Teton National
Park, Wyoming, USA. Ohio: ASM International; 2009. p. 46–53.
[42] Nibur KA, Somerday BP, San Marchi C, Foulk III JW, Dadfarnia M, Sofronis P, et al. Measurement and interpretation of threshold stress intensity factors for steels in
high-pressure hydrogen gas. Sandia technical report no.: SAND2010-4633. Contract no.: DE-AC04-94AL85000. Sandia National Laboratories; July 2010.
[43] M.L. Jokl, C.J. McMahon, P. Burgers, On the micromechanics of brittle fracture: existing vs injected cracks, Acta Metall 37 (1) (1989) 87–97.