See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/324464835

Effect of hydrogen on the tensile properties of 42CrMo4 steel quenched and

tempered at different temperatures

Article in International Journal of Hydrogen Energy · April 2018

DOI: 10.1016/j.ijhydene.2018.03.158

CITATIONS READS

61 5,490

4 authors, including:

Alfredo Zafra F. J. Belzunce

Imperial College London University of Oviedo
34 PUBLICATIONS 510 CITATIONS 180 PUBLICATIONS 3,083 CITATIONS

SEE PROFILE SEE PROFILE

Cristina Rodríguez
University of Oviedo
166 PUBLICATIONS 2,029 CITATIONS

SEE PROFILE

All content following this page was uploaded by Alfredo Zafra on 31 August 2018.

The user has requested enhancement of the downloaded file.

 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 9 0 6 8 e9 0 8 2

Available online at www.sciencedirect.com

ScienceDirect

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

Effect of hydrogen on the tensile properties of

42CrMo4 steel quenched and tempered at different
temperatures

A. Zafra*, L.B. Peral, J. Belzunce, C. Rodrı́guez

University of Oviedo, Campus Universitario, East Building, 33203, Gijon, Spain

article info abstract

Article history: The influence of hydrogen on the mechanical behaviour of a 42CrMo4 tempered
Received 22 December 2017 martensitic steel was investigated by means of tensile tests on both smooth and
Received in revised form circumferentially-notched round-bar specimens pre-charged with gaseous hydrogen in a
20 March 2018 pressurized reactor.
Accepted 22 March 2018 Hydrogen solubility was seen to decrease with increasing tempering temperature.
Available online 11 April 2018 Moreover, hydrogen embrittlement measured in notched specimens was much greater in
the grades with higher hardness, tempered at the lowest temperatures, where a change in
Keywords: the fracture micromechanism from ductile in the absence of hydrogen to intermediate and
Hydrogen embrittlement brittle in the presence of hydrogen was clearly observed. Results were discussed through
Structural steels FEM simulations of local stresses acting on the process zone.
Notched tensile strength © 2018 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
Failure micromechanisms

embrittlement, such as austenitic stainless steel and

Introduction aluminium alloys. However, these metals have lower strength
and are more expensive than conventional structural steels,
The global trend to replace fossil fuels with cleaner and such as carbon and low-alloy steels [1]. The economical so-
renewable alternatives is boosting the development of lution would be to use medium- and high-strength tempered
hydrogen-based energy sources. In this context, the expected martensitic steels, which allow decreasing pipe and vessel
increase in the demand for hydrogen in the coming years thicknesses and hence the cost of the steel.
means that current storage and delivering facilities and ma- 42CrMo4 is a medium carbon Cr-Mo alloyed steel that is
terials will need to evolve so as to be able to deal with larger frequently used in a quenched and tempered condition when
amounts of hydrogen in an efficient way. Furthermore, to fulfil a good combination of strength and toughness is required. It
this demand, transport and storage units (pipes and vessels) also has high fatigue strength and wear resistance, making it
will be required to withstand higher hydrogen pressures, of up an excellent candidate for manufacturing vessels working
to 70 MPa. under high-pressure hydrogen gas [2]. Nevertheless, it is well
Hydrogen systems have typically been designed to employ known that these steel grades are more sensitive to hydrogen
metallic materials with a high resistance to hydrogen embrittlement (HE) than low-strength steels, this

* Corresponding author.
E-mail address: [email protected] (A. Zafra).
https://doi.org/10.1016/j.ijhydene.2018.03.158
0360-3199/© 2018 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 9 0 6 8 e9 0 8 2 9069

susceptibility increasing with the strength level of the steel

[3,4]. As a result, 42CrMo4 steel is usually tempered at high Experimental procedure
temperatures with reduced strength and hardness so as to
reduce HE-related problems. Materials and heat treatments
Although the deleterious effects of hydrogen on steels
similar to 42CrMo4 have been studied for some time now, the A 42CrMo4 steel was used in the present study. Its chemical
mechanism of hydrogen embrittlement phenomena is still composition in weight % is shown in Table 1.
controversial. Anyway, the common understanding is that Hot rolled plates (250
250
12 mm3) of 42CrMo4 were
mechanical loads, hydrogen uptake, diffusion of mobile austenitized at 845  C for 40 min, quenched in water, and then
hydrogen and, of course, the steel microstructure are involved tempered at different temperatures, 500, 550, 600, 650 and
[4e6]. The mechanism of hydrogen-induced failure in 700  C, for 2 h. The reason for using different heat treatments
quenched and tempered steels usually starts with the stress was to obtain different steel grades in order to analyse the
activation of slip systems in individual grains, being known influence of microstructure and strength level on the sus-
that the enhancement of this dislocation activity is due to the ceptibility of the steel to hydrogen embrittlement. The
hydrogen-enhanced localized plasticity (HELP) mechanism sequence of heat treatments and the nomenclature of the
[7,8]. Hydrogen trapping by dislocations reduces the elastic 42CrMo4 steel grades are shown in Table 2.
energy between them and thereby decreases the Peierls
stress. As a result the mobility of dislocations increases and
Hydrogen charging
plasticity localizes [9]. In the case of high strength steels, their
All the specimens were pre-charged with gaseous hydrogen in
final failure can be intergranular, promoted by carbide pre-
a high-pressure reactor (with a diameter of 73 mm and total
cipitation and solute content along austenitic grain bound-
length of 180 mm), manufactured in accordance with the
aries, or transgranular associated with cracking along
ASTM G146 standard [12]. Before hydrogen charging, air was
martensitic block and packet interfaces. Moreover, despite
removed from the vessel using three alternate inert gas purges
many studies about mobile and trapped hydrogen, it is diffi-
to reduce the oxygen level to a minimum. To ensure that the
cult to establish a consensus on the impact of both hydrogen
specimens were saturated with hydrogen, they were kept in
states on damage mechanisms. In general, segregation of
contact with gaseous hydrogen under a pressure of 19.5 MPa
hydrogen is necessary to reduce the interface energy and
at 450  C for 21 h. After this holding time, the reactor was
induce a cracking process, although it is argued that the mo-
cooled down for 1 h until reaching a temperature of 85  C.
bile hydrogen reduces the cohesion energy greater than
During this phase, the hydrogen pressure was maintained
trapped hydrogen [10].
constant at 19.5 MPa to reduce hydrogen egress but, never-
On the other hand, the effect of hydrogen on the me-
theless, significant hydrogen loss took always place. All the
chanical behaviour of high strength steels may depend on the
pre-charged specimens were subsequently removed from the
dimensions and geometry of the specimen, on the stress
reactor and immersed in liquid nitrogen (196  C), where they
concentration factor, when notched specimens are employed,
were kept until the start of the corresponding mechanical test,
as well as on certain test variables, such as the strain rate [4]. It
once again in order to limit hydrogen losses.
is also important to note that embrittlement evaluation is
often conducted on mechanical tests performed in samples
Measurement of hydrogen content
that were electrochemically charged with hydrogen, but it is
known that electrochemical charging produce high hydrogen
The hydrogen desorption curves of the different steel grades
fugacity near the surface and sometimes causes surface
at room temperature were determined. These curves repre-
cracking or blistering, so these type of tests should be
sent the hydrogen content versus time (ppm vs time). Cylin-
considered with caution when materials are exposed to
drical pins of 10 mm in diameter and 30 mm long (z20 g) were
gaseous hydrogen [11].
used to obtain the steel desorption curves.
In this context, further investigation is also needed for
A LECO DH603 hydrogen analyser, able to measure hydrogen
steels of this kind heat treated at intermediate and high
concentrations between 0.1 and 2500 ppm, was used to measure
strength levels, as these grades are more susceptible to
the hydrogen content. The measuring principle of the equip-
hydrogen embrittlement. In fact, the main objective of this
ment is based on the difference in thermal conductivity be-
study is to analyse the influence of heat treatment (specif-
tween a reference gas flow of pure nitrogen and a secondary
ically the influence of tempering temperature) on the
flow composed of nitrogen and the hydrogen thermally
hydrogen embrittlement susceptibility of 42CrMo4 steel.
extracted from the analysed specimen. This difference is
Hence, the tensile behaviour of both smooth and
measured by means of a Wheatstone bridge and converted into
circumferentially-notched round-bar specimens pre-charged
with gaseous hydrogen was compared. Furthermore,
different displacement rates were used to analyse the influ-
ence of the strain rate on hydrogen embrittlement. FEM
Table 1 e Chemical composition in weight% of 42CrMo4
simulations of the notched tensile tests were performed and steel.
along with the observed fracture micromechanisms, a dis-
Steel C Mn Si P S Cr Mo
cussion of failure processes in the presence of hydrogen was
provided. 42CrMo4 0.42 0.62 0.18 0.008 0.002 0.98 0.22
9070 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 9 0 6 8 e9 0 8 2

notch. The extent of hydrogen embrittlement was assessed by

Table 2 e Heat treatments applied to the 42CrMo4 steel.
means of the embrittlement index (EI), defined in Eq. (1) (EI
Specimen ID Heat treatments varies from 0, no embrittlement at all, to 100%, maximum
42CrMo4_700 845  C/40 min þ water possible hydrogen embrittlement).
quenching þ 700  C/2 h tempering
42CrMo4_650 845  C/40 min þ water X  XH
EI ½% ¼ $100 (1)
quenching þ 650  C/2 h tempering X
42CrMo4_600 845  C/40 min þ water
where X and XH are respectively the measured steel property
quenching þ 600  C/2 h tempering
evaluated without and with hydrogen. The influence of the
42CrMo4_550 845  C/40 min þ water
quenching þ 550  C/2 h tempering displacement rate on the reduction in area, RA, of the
42CrMo4_500 845  C/40 min þ water hydrogen pre-charged notched tensile specimens was also
quenching þ 500  C/2 h tempering assessed.

a signal that can be interpreted by the device's software, which FEM simulation
finally gives the hydrogen concentration in ppm.
The procedure to obtain the desorption curves was the In order to fully characterize the geometry of the notched
following. All the pins were removed from the liquid nitrogen tensile specimens and discuss the obtained experimental re-
at the same time and left in air at room temperature. Then, the sults, the FEM software Abaqus CAE was used to obtain the
hydrogen concentration of the different samples was stresses that developed around the notch region of the spec-
measured at different time intervals. Before starting the imen at the moment of failure. A 2D elastic-plastic finite
measurement, each pin was cleaned in an ultrasonic bath element model with a Hollomon hardening law obtained from
with acetone for 5 min and carefully dried using cold air. The the tensile tests was developed, since the specimens were
analysis to determine the hydrogen concentration consisted axisymmetric (Fig. 1). The geometry of the specimens was
in keeping the sample at 1100  C for 400 s. unevenly meshed, being this mesh finer in the vicinity of the
notch root. 8-node biquadratic axisymmetric quadrilateral
Tensile tests elements with reduced integration (CAX8R) were used with a
minimum size around 40 mm.
Tensile tests on both smooth and circumferentially-notched
round-bar specimens, whose dimensions and geometries are Observation of microstructures and fracture surfaces
shown in Fig. 1, were performed on an Instron 5582 tensile
testing machine. The microstructures obtained after the application of the
The stress concentration factor of the notched tensile different heat treatments were observed by means of a scan-
specimen was calculated according to Neuber [13]. Using a ning electron microscope (SEM JEOL-JSM5600) under an ac-
notch root radius of 0.15 mm and a radius of the notch cross- celeration voltage of 20 kV. The samples were previously
section of 5 mm, a stress concentration factor Kt ¼ 4.25 was ground, then polished with diamond paste, and finally etched
obtained. This value is in agreement with those calculated by with Nital-2%. The fracture surfaces of the tested specimens
other authors for similar specimen geometries [4,14]. were also analysed using the same scanning electron micro-
The behaviour of uncharged and hydrogen pre-charged scope under different magnifications.
specimens was compared in these tests. All uncharged tests
were performed under a displacement rate of 0.4 mm/min, but
different displacement rates, 0.4, 0.04 and 0.004 mm/min, Results
were applied in the case of the pre-charged specimens to
study the influence of this parameter on HE. Steel microstructures
In the case of the notched specimens, the engineering
notched tensile strength (sN) was defined as the maximum The SEM microstructures of the 5 grades of 42CrMo steel are
tensile load divided by the initial cross-sectional area of the shown in Fig. 2 under 10000
magnification. The

Fig. 1 e Geometry and dimensions (mm) of the smooth and notched specimens used in tensile tests.
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 9 0 6 8 e9 0 8 2 9071

Fig. 2 e SEM microstructures of 42CrMo4 steel, quenched and tempered at (a) 500  C, (b) 550  C, (c) 600  C, (d) 650  C, (e) 700  C
and f) Quenched and non-tempered.

microstructure presented in Fig. 2 f corresponds to a quenched increasing tempering temperature, while the opposite occurs
and non-tempered sample, in which a prior austenite grain with tensile elongation, ε, and reduction in area, RA.
size of 15e20 mm was determined. It is well known that hydrogen embrittlement increases
In all cases, the obtained microstructure was tempered with the strength level of the steel [3,4], hence grades
martensite (the profuse carbide precipitation that takes place tempered at lower temperatures are expected to be more
during the tempering stage can be seen). With increasing susceptible to HE.
tempering temperature, microstructure acicularity decreases,
internal residual stresses are released, dislocation rearrange- Hydrogen desorption curves
ment/annihilation and substructure recovery takes place,
given rise to the decrease of martensite lath boundaries [15]. The hydrogen desorption behaviour at room temperature of
Furthermore, elongated carbides break up, globulize, and each 42CrMo4 steel grade is shown in Fig. 3. An initial
finally grow to yield a more uniform distribution. hydrogen content, CH0, between 1.2 and 1.9 ppm was intro-
The aforementioned microstructural differences give rise duced in the different steel grades. As the hydrogen content
to the hardness and tensile properties listed in Table 3. As present in these grades before hydrogen charging was
expected, Brinell hardness, HB, yield strength, sys, and ulti- 0.2 ppm, a significant hydrogen uptake took always place.
mate tensile strength, suts, progressively decrease with Hydrogen uptake increased with decreasing tempering
9072 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 9 0 6 8 e9 0 8 2

Table 3 e Hardness (HB) and tensile properties of the different 42CrMo4 grades under study.
Steel Grades Heat treatment HB sys [MPa] suts [MPa] ε [%] RA [%]
 
42CrMo4_700 845 CþWQþT700 C/2 h 201 622 710 22.4 61.3
 
42CrMo4_650 845 CþWQþT650 C/2 h 246 820 905 16.3 57.1
 
42CrMo4_600 845 CþWQþT600 C/2 h 281 880 985 14.6 55.4
 
42CrMo4_550 845 CþWQþT550 C/2 h 307 1023 1113 13.9 48.9
 
42CrMo4_500 845 CþWQþT500 C/2 h 335 1086 1198 12.7 50.6

Fig. 3 e Hydrogen desorption curves of 42CrMo4 steel grades at room temperature.

temperature, as hydrogen microstructural trapping is greater Tensile tests on smooth specimens

in distorted high energy martensitic microstructures. The
final hydrogen content, CHf, is the hydrogen strongly trapped Tensile tests were performed on smooth hydrogen pre-
and retained in the steel microstructure after a long time at charged specimens of each grade of 42CrMo4 steel. Fig. 4
room temperature. This value is even more dependent on the shows the stress-strain curves obtained with grades
microstructure of the steel than the former value, decreasing tempered at the highest and lowest temperatures (700 and
as the tempering temperatures increases, mainly due to stress 500  C). No significant difference was observed between tests
relaxation and the reduction in dislocation density and in- carried out at different displacement rates. Fig. 5 shows the
terphases (martensite laths, blocks and packets). Moreover, embrittlement index (EI) corresponding to fracture strength
the diffusible hydrogen, i.e. the difference between these two and reduction in area versus steel hardness for the lowest
values, CH0-CHf, is the hydrogen able to move freely within the displacement rate used in these tests (0.04 mm/min). A small
steel at room temperature, overcome traps, and diffuse out of reduction of the yield strength was also observed in the
the steel. All the hydrogen values measured in the different presence of internal hydrogen.
42CrMo4 steel grades are shown in Table 4. It is also worth The embrittlement index for strength remains almost
noting that diffusible hydrogen progressively increases with constant at approximately 2%, which means that hydrogen
increasing tempering temperature. has little influence on the mechanical strength for this ge-
ometry, irrespective of the hardness of the steel. Regarding
the reduction in area, the EI is 0% for tempering temperatures
Table 4 e Initial (CH0), final (CHf), hydrogen retained after a greater than 600  C, increasing slightly to 6e12% for the two
month at room temperature) and diffusible hydrogen lowest temperatures (i.e. the hardest grades).
(CH0-CHf) measured on the different 42CrMo4 steel grades. In conclusion, it may be stated that tensile properties
Steel Grades HB CH0 [ppm] CHf [ppm] CH0-CHf [ppm] measured on smooth specimens pre-charged with gaseous
hydrogen remained practically unaffected, irrespective of the
42CrMo4_700 201 1.2 0.4 0.8
applied displacement rate. Moreover, the fracture micro-
42CrMo4_650 246 1.2 0.7 0.5
42CrMo4_600 281 1.4 0.9 0.5 mechanism observed on the failure surfaces of all these
42CrMo4_550 307 1.7 1.4 0.3 specimens under the scanning electron microscope was al-
42CrMo4_500 335 1.9 1.7 0.2 ways fully ductile (microvoid coalescence).
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 9 0 6 8 e9 0 8 2 9073

Fig. 4 e Stress-strain curves of tensile tests, a) 42CrMo4_700 and b) 42CrMo4_500.

Fig. 5 e Influence of tempering temperature on the embrittlement index (EI) calculated for fracture strength and reduction in
area of hydrogen pre-charged tensile smooth specimens. Displacement rate: 0.04 mm/min.

Tensile tests on notched specimens were rather slight. Furthermore, the displacement rate does
not seem to play an important role in this case. According to
Fig. 6 shows the results obtained for 42CrMo4 quenched and Fig. 3, the amount of pre-charged hydrogen in this steel grade
tempered at 500 and 700  C, which were respectively the was also the lowest (1.2 ppm). Due to this fact and its low yield
grades less and most affected by the deleterious effects of strength, hydrogen accumulation in the notched region of this
hydrogen. The stress-strain curves of the uncharged steel grade was limited and the embrittlement reaction was
(displacement rate of 0.4 mm/min) and hydrogen pre-charged minimised. The steel submitted to the other tempering tem-
notched specimens, tested at different displacement rates peratures gave rise to intermediate behaviours.
(0.4, 0.04 and 0.004 mm/min), are shown in Fig. 6. All these Table 5 summarizes the results obtained in the tensile tests
graphs show a reduced plastic region, as most of the specimen performed on the notched specimens for the different
is only elastically strained, except for a very small region close 42CrMo4 steel grades. The values of the notch tensile strength,
to the stress concentrator, as it will be demonstrated in the sN, and reduction in area, RA, are shown for each grade, un-
next section. charged and hydrogen pre-charged under different displace-
As the test displacement rate decreases, it can be seen that ment rates. Embrittlement indexes calculated for the
the damage caused by hydrogen increases considerably in the aforementioned mechanical properties and prevailing frac-
case of the steel grade tempered at 500  C (sys ¼ 1086 MPa). In ture micromechanisms are also presented in Table 5. More-
fact, the fracture strength of this steel grade decreased almost over, the test duration and approximate average hydrogen
60%, and the reduction in area more than 90%, when spec- content, CH, of the specimens in the course of the tests can
imen was pre-charged and tested at the lowest strain rate. also be seen in the same table. It can be seen that test duration
Quite different behaviour was observed on the barely affected the hydrogen content in these specimens,
42CrMo4_700 grade. In this case, due to the much lower yield except for the lowest displacement rate. Anyway, the values
strength of the steel (sys ¼ 622 MPa), the effects of hydrogen of CH in Table 5 may be taken with caution as they were
9074 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 9 0 6 8 e9 0 8 2

Fig. 6 e Stress-strain curves of the uncharged and hydrogen pre-charged notched tensile specimens (Kt ¼ 4.25) of a)
42CrMo4_700 and b) 42CrMo4_500 steel. Influence of the applied displacement rate.

Table 5 e Results of tensile tests performed on notched specimens (Kt ¼ 4.25) of 42CrMo4 steel uncharged and hydrogen
pre-charged and loaded at different displacement rates.
Steel Grade Test speed Test CH sN RA EI(suts) EI(RA) Fracture
[mm/min] duration [ppm] [MPa] [%] [%] [%] Micromechanisms
42CrMo4_700 (201HB) 0.4 12 min Uncharged 1330 12.5 e e MVC
0.4 9 min 1.2 1273 11.8 4 6 MVC
0.04 80 min 1.2 1258 10.3 5 18 MVC
0.004 14 h 1.2 / 0.7 1294 9.6 3 23 MVC
42CrMo4_650 (246HB) 0.4 17 min Uncharged 1582 11.9 e e MVC
0.4 15 min 1.2 1527 11.8 3 1 MVC
0.04 105 min 1.2 1508 8.9 5 25 MVCþPRHIC
0.004 16 h 1.2 / 0.85 1487 7.9 6 34 MVCþPRHIC
42CrMo4_600 (281HB) 0.4 12 min Uncharged 1724 8.5 e e MVC
0.4 12 min 1.4 1604 7.6 7 11 MVC
0.04 85 min 1.4 1384 6.0 20 29 MVCþPRHIC
0.004 14 h 1.4 / 1.0 1336 3.4 23 60 MVCþPRHICþIG
42CrMo4_550 (307HB) 0.4 8 min Uncharged 1800 7.5 e e MVC
0.4 9 min 1.7 1602 6.3 11 16 MVCþPRHICþIG
0.04 65 min 1.7 1303 4.9 28 35 MVCþPRHICþIG
0.004 11 h 1.7 / 1.6 1248 2.7 31 64 MVCþPRHICþIG
42CrMo4_500 (335HB) 0.4 15 min Uncharged 1833 6.4 e e MVC
0.4 13 min 1.9 1534 4.7 16 27 MVCþPRHICþIG
0.04 80 min 1.9 1050 3.1 43 51 MVCþPRHICþIG
0.004 6h 1.9 / 1.8 771 0.5 58 92 MVCþPRHICþIG

MVC: microvoid coalescence; PRHIC: plasticity-related hydrogen induced cracking; IG: intergranular.

derived from the desorption curves (Fig. 3) obtained with Fracture surfaces analysis of notched specimens
samples with slightly different dimensions.
Fig. 7 shows the notch tensile strength and reduction in The fracture surfaces corresponding to the notched speci-
area obtained with all these grades using uncharged and mens of the different steel grades under study were analysed
hydrogen pre-charged notched tensile specimens (under a using SEM and the main operative fracture micromechanisms
displacement rate of 0.004 mm/min). The presence of were identified.
hydrogen barely modifies the notched tensile strength of the The fracture micromechanism of all the tested grades
grades with the lowest hardness, 201 and 246 HB (respectively without hydrogen was fully ductile, microvoid coalescence
tempered at 700 and 650  C), but causes a significant decrease (MVC), as already shown in Table 5 and as can also be seen in
in the case of the grades with the highest hardness (tempered Fig. 8 (a) and (b).
at the lowest temperatures). This effect greatly increases with The remaining fractographs, shown in Fig. 9, correspond to
increasing hardness (i.e. decreasing tempering temperature). hydrogen pre-charged notched tensile specimens tested
In the case of the reduction in area, the behaviour is different: under displacement rates of 0.004 mm/min. Fig. 9 (a) corre-
the embrittlement effect of hydrogen is evident and practi- sponds to the grade tempered at 700  C. The fracture micro-
cally constant for all the tested grades. mechanism did not change in this grade, but elongated, flat
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 9 0 6 8 e9 0 8 2 9075

Fig. 7 e Tensile strength and reduction in area versus hardness (HB) of notched specimens (Kt ¼ 4.25) without hydrogen and
hydrogen pre-charged. Displacement rate: 0.004 mm/min.

Fig. 8 e Fracture micromechanism of (a) 42CrMo4_700 (uncharged) and (b) 42CrMo4_500(uncharged). 500£.

areas such as those shown in Fig. 9 (a), were noted as a comparable to the microstructure units (martensite blocks)
characteristic feature. shown in Fig. 2 (d); hence, hydrogen accumulation promotes
Furthermore, in the case of 42CrMo4_650, as can be seen in plastic deformation and final failure of these interphases.
Fig. 9 (b) and (c), two different regions were observed in the As to the grades tempered at lower temperatures (600, 550
specimen: a central region with microvoids, some of which and 500), the observed fracture micromechanism was a
were also enlarged, and a narrow peripheral region of around mixture of MVC, PRHIC and intergranular fracture (IG). In the
100 mm that exhibited the appearance of a transgranular bulk of the specimen, MVC was the predominant micro-
fracture surface or quasi-cleavage, usually referred to as mechanism, although PRHIC and IG were the characteristic
plasticity-related hydrogen induced cracking (PRHIC) in micromechanisms in the periphery of these specimens and
martensitic steels [16]. The PRHIC mechanism was first also, in some cases, along some radial “paths” leading from
described by Takeda and McMahon [17] in reference to the the periphery to the centre of the specimen. Examples of these
fracture mechanism observed in a low alloy quenched and characteristic surfaces are shown in Fig. 9 (e, f, g, h, i and j) at
tempered steel in hydrogen gas. It is sometimes called tearing 500, 1000 and 2000
magnification.
topography surface or TTS, which is described as a fracture It is worth noting that the width of the belt characterized
surface characterized by ductile micro-plastic tearing on a by PRHIC and IG fracture increases with increasing hardness
very fine scale, along martensite lath interphases [16,18]. The of the steel (i.e. decreasing tempering temperature), being
appearance of this region can be better seen in Fig. 9 (d). The greater than 250 mm in all cases. Furthermore, the extent of
size of the characteristic features observed in Fig. 9 (d) are intergranular fracture is larger in the grade with the highest
9076 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 9 0 6 8 e9 0 8 2

Fig. 9 e Fracture micromechanisms of 42CrMo4 grades tested with hydrogen at a displacement rate of 0.004 mm/min.
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 9 0 6 8 e9 0 8 2 9077

Fig. 10 e Distribution of sVM/sys along the radial direction when the applied net stress is equal to the notch tensile strength.
42CrMo4 steel quenched and tempered at different temperatures.

Fig. 11 e Distribution of sVM, sh and s22 along the radial direction for six different loading steps. 42CrMo4_650. Applied load
increases from 1 to 6.
9078 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 9 0 6 8 e9 0 8 2

Fig. 12 e Distribution of sh, and s22 along the radial direction when the applied net stress is equal to the notch tensile
strength. 42CrMo4 steel quenched and tempered at different temperatures.

strength, 42CrMo4_500, which also gave rise to the highest Fig. 12 shows now the evolution of sh, and s22 versus
embrittlement indexes, as was seen in Table 5. The estimated distance from the notch tip at an applied stress equal to the
size of the grains directly measured on these fractographs was notch tensile strength corresponding to the five analysed
around 15e20 mm, which corresponds to the prior austenite steel grades. The level of the normal stress and hydrostatic
grain size measured on the polished and etched sections of stress increases, especially in the region closer to the notch
these steels (see Fig. 2 f). tip, as the yield strength of the steel grade is increased and
Now, with the aim of explaining the aforementioned at the same time it also worth noting the displacement of
experimental results, the distribution of the local normal the maximum of both aforementioned stresses to a lower
stress, perpendicular to the notch plane, s22, the Von Mises depth.
stress, sVM (Eq. (2)), and the hydrostatic stress, sh (Eq. (3)) was
simulated along a radial path, defined from the free surface at
the tip of the notch (x ¼ 0) to the centre of the specimen in the Discussion
notched section (x ¼ 2.5 mm).
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi It is well known that the internal hydrogen distribution in
ðs11  s22 Þ2 þ ðs22  s33 Þ2 þ ðs33  s11 Þ2 high strength steels submitted to external loads is governed
sVM ¼ (2)
2 by hydrostatic stress [17]. Due to this effect, hydrogen atoms
diffuse along the steel microstructure and move to regions
s11 þ s22 þ s33 submitted to the highest hydrostatic stress, located near the
sh ¼ (3)
3 notch tip (see Figs. 11 and 12). As a result of the accumulation
being s11, s22 and s33 the principal stresses. of hydrogen in this region, a critical hydrogen concentration
The sVM/sys ratio versus distance from notch tip at an may be reached, leading to premature failure of the specimen
applied stress equal to the notch tensile strength corre- [19].
sponding to the five analysed steel grades is shown in Fig. 10. As a summary of the study, Fig. 13 shows the embrittle-
The extension of the plastic region varies between 100 and ment indexes under a displacement rate of 0.004 mm/min
180 mm, increasing when decreasing the yield strength of the calculated for the notch tensile strength and the reduction in
steel (higher tempering temperature). It is then worth noting area versus the Brinell hardness of each grade (tempering
that in all cases most of the specimen is only elastically temperature). Additionally, the fracture micromechanisms
strained, except for a very small region close to the stress and approximate hydrogen content are also included on this
concentrator. figure. A growing trend can be observed for both embrittle-
Fig. 11 shows the evolution of sVM, sh, and s22 in the course ment indexes with increasing steel hardness. This trend is
of loading for one of the steel grades (42CrMo4_650). The two consistent with the change appreciated in the fracture
lowest stress profiles correspond to essentially full elastic micromechanisms. MVC was the only failure micro-
loadings (sVM  sys ¼ 880 MPa). For higher applied loads, a mechanism in the grade with the lowest hardness, while
plastic zone develops in the front of the notch and the last MVCþPRHIC and MVCþPRHICþIG were the operative failure
represented stress profile corresponds to the instant of failure micromechanisms in the case of the grades with intermediate
of the specimen. It is also noted that once a plastic zone is hardness and the highest hardness, respectively. Now, it
developed (sVM > sys), a maximum in the Von Mises and in the should be recalled that hydrogen embrittlement is controlled
hydrostatic stress profiles is produced at a certain depth of the by local stress and local hydrogen concentration, and only
notch tip and the location of this maximum displaces to larger occurs when the critical combination of both factors is satis-
depths when loading is increased. fied in some specific microstructure units.
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 9 0 6 8 e9 0 8 2 9079

Fig. 13 e Embrittlement indexes for notched tensile strength and reduction in area corresponding to each of the studied
42CrMo4 grades. Operative fracture micromechanisms and hydrogen content are included. Displacement rate: 0.004 mm/min.

The MVC micromechanism was always the only failure whose extension increased from 100 to 250 mm, as tempering
mechanism observed in the uncharged specimens of all temperature reduces from 650 to 600  C. PRHIC was the failure
grades, while the presence of hydrogen in the pre-charged micromechanism observed in this region. It is related to
specimens of the steel grade with the lowest hardness, hydrogen accumulation in martensite block and packet in-
42CrMo4_700, although maintaining the same operative terphases, giving rise to local micro-plastic tearing on a very
micromechanism, gave rise to larger and shallower dimples fine scale. It is seen that PRHIC requires a greater accumula-
(see Fig. 9 a), as already stated in Ref. [20]. According to the tion of hydrogen in the notch region and, in this case, it was
hydrogen-enhanced localized plasticity mechanism (HELP) due to the existence of higher hydrogen content and higher
[7,9,20,21], the presence of solute or internal hydrogen in- local hydrostatic stresses. The embrittlement produced by
creases the mobility of dislocations enabling them to move PRHIC (failure initiates just ahead of the notch under this
under lower stresses, providing plasticity localization on the mechanism and MVC takes place only afterwards in the
notch region, where shallow dimples are produced and grow, centre of the specimen) was evident in Fig. 13, when the
given rise to very large voids, as the ones shown in Fig. 9 (a). measured embrittlement indexes of the two grades with the
When this was the only operative micromechanism (i.e. in the lowest hardness are compared. Moreover, the area of the
steel grade with the lowest hardness, which is also the one failure surface where PRHIC was detected was larger in the
with the lowest hydrogen content, 1.2e0.7 ppm), the effect of grades with the highest hardness (Fig. 9), the corresponding
hydrogen was nearly negligible, as the embrittlement indexes embrittlement indexes increase accordingly.
were quite low (non-significant for the strength and around Finally, intergranular fracture was observed in the steel
20% for the reduction of area). grades with the highest hardness, also giving rise to the
In all the other grades, ductile features (MVC) are observed highest embrittlement indexes. The notched specimens of
in the centre of the specimen and brittle areas at the notch tip these steels also had higher hydrogen content and were sub-
region [22,23] due to the stress profiles described in Figs. 10e12 ject to higher local hydrostatic stress. The accumulation of
(see also Fig. 9 c, e). Failure initiates at the plastic zone at the hydrogen ahead of the notch front was the highest, giving rise
tip of the notch, where normal and hydrostatic stresses are to decohesion of the prior austenitic grain boundaries
larger and hydrogen accumulates driven by the high hydro- (hydrogen-enhanced decohesion mechanism, HEDE). In these
static stress. By the contrary, in the centre of the specimen cases, intergranular fracture is the first failure micro-
and due to the relatively low level of the local hydrostatic mechanism to take place and gives rise to a substantial
stress, hydrogen diffused away and failure always took place embrittlement, especially in the case of the grade with the
in a ductile manner (MVC). highest hardness.
In the case of the grades with intermediate hardness, the Recent studies have shown that intense slip bands (defor-
existence of PRHIC was detected in the region just ahead of the mation bands) are present beneath intergranular and quasi-
notch. As was observed in Fig. 12, hydrostatic stress reaches cleavage fracture surfaces as a consequence of the enhance-
its peak value at a depth between 80 mm and 150 mm for the ment of dislocation activity by the hydrogen-enhanced local-
different grades. This region suffer the greatest embrittle- ized plasticity (HELP) mechanism [7,24,25]. The reduction of
ment, as it was shown in Fig. 9 c) and e), where quasi-brittle notch area measured in all our tests, except for the steel grade
mechanisms were observed at a narrow peripheral region, with the highest strength under the lowest displacement rate,
9080 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 9 0 6 8 e9 0 8 2

Fig. 14 e Failure micromechanisms ahead of the notch in hydrogen charged quenched and tempered steels under similar
stress concentrator and hydrogenated medium (a) IG, (b) PRHIC and (c) MVC.

is the best evidence of the existence of plasticity in the process structural defects with a large capacity to accumulate hydrogen
region before failure (see Table 5). Hydrogen redistribution atoms can be found at the prior austenitic grain boundaries. In
occurs among the traps in front of the notch in the process this situation, when internal hydrogen is present, hydrogen
region submitted to high hydrostatic stress, causing hydrogen accumulates preferentially at the prior austenitic grain bound-
accumulation on internal boundaries due to hydrogen trans- aries existing in the notch front region submitted to high hy-
port by dislocations, until a crack initiates and propagates by drostatic stress, giving rise to decohesion of these interphases
interface decohesion driven by the action of the local stress when a high enough local stress is applied. As a result, inter-
[7,8,26]. Nagao et al. [25] explain this mechanism as a granular fracture takes place under a much lower tensile stress
concerted action of the HELP and HEDE mechanisms whose than the failure stress of the same specimen without hydrogen.
operation is assisted by the hydrogen deposited on the inter- As the tempering temperature increases, the steel microstruc-
nal boundaries and called it hydrogen-enhanced-plasticity ture relaxes, the distortion of the austenitic grain boundaries
mediated decohesion. progressively diminish and their capacity to accumulate
Regarding the effects of microstructure on the hydrogen hydrogen decreases at the same time. In this case, the in-
embrittlement of quenched and tempered steels, it should be terphases between martensite blocks and packets are the sites
recalled that after low temperature tempering, relaxation of the where hydrogen accumulation is more abundant, fracture
quenched microstructure is still quite low and an abundance of micromechanism changes from IG to PRHIC and embrittlement
 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 9 0 6 8 e9 0 8 2 9081

indexes decrease. It should also be noted that the surface per references
unit volume of all these interphases is much greater than those
related to the austenitic grain boundaries. As the yield strength
of the steels quenched and tempered at higher temperatures [1] Yamabe J, Awane T, Matsuoka S. Investigation of hydrogen
decreases, local normal stress and hydrostatic stress in the transport behavior of various low-alloy steels with high-
process zone diminish, contributing also to improve the pressure hydrogen gas. Int J Hydrogen Energy
embrittlement indexes. Only when the tempering temperature 2015;40(34):11075e86.
[2] Chuang JH, Tsay LW, Chen C. Crack growth behaviour of
is high enough, microstructure relaxation reach a level at which
heat-treated 4140 steel in air and gaseous hydrogen. Int J
hydrogen accumulation in all the aforementioned features is Fatig 1998;20(7):531e6.
low and not sufficient to produce decohesion of any these in- [3] Briottet L, Batisse R, de Dinechin G, Langlois P, Thiers L.
terphases. In this situation, hydrogen distributes more uni- Recommendations on X80 steel for the design of hydrogen
formly in the notch front region and, due to HELP, dislocations gas transmission pipelines. Int J Hydrogen Energy
move easily, plastic deformation in this region increases, the 2012;37(11):9423e30.
[4] Wang M, Akiyama E, Tsuzaki K. Effect of hydrogen on the
fracture micromechanism now becomes ductile (nucleation,
fracture behavior of high strength steel during slow strain
growth and coalescence of microcavities, MVC), and failure
rate test. Corros Sci 2007;49(11):4081e97.
takes place under only a slightly lower level of stress than in the [5] Akiyama E, Wang M, Li S, Zhang Z, Kimura Y, Uno N, et al.
absence of hydrogen. All these features are shown in Fig. 14. Studies of evaluation of hydrogen embrittlement property of
high-strength steels with consideration of the effect of
atmospheric corrosio  n. Metall Mater Trans
Conclusions 2013;44A:1290e300.
[6] Singh DK, Raman RKS, Maiti SK, Bhandakkar TK, Pal S.
Investigation of role of alloy microstructure of AISI 4340 steel
The initial hydrogen content, measured in samples pre-
using circumferentially notched cylindrical specimens.
charged under high temperature hydrogen pressure, Mater Sci Eng A 2017;698:191e7.
increased with decreasing steel tempering temperature due to [7] Nagao A, Smith CD, Dadfarnia M, Sofronis P, Robertson IM.
the fact that hydrogen microstructural trapping is greater in Interpretation of hydrogen-induced fracture surface
distorted, high energy martensitic microstructures. Moreover, morphologies for lath martensitic steel. Procedia Mater Sci
the content of hydrogen strongly trapped is even more 2014;3:1700e5.
[8] Robertson IM, Sofronis P, Nagao A, Martin ML, Wang S,
dependent on the steel microstructure, decreasing with
Gross DW, et al. Hydrogen embrittlement understood. Metall
increasing tempering temperature, mainly due to stress Mater Trans 2015;46B:1085e103.
relaxation and the reduction in dislocation density and in- [9] Gangloff RP, Somerday BP. Edit, “Gaseous hydrogen
terphases (martensite laths, blocks and packets). embrittlement of materials in energy technologies”.
The tensile properties measured on smooth specimens Cambridge, UK: Woodhead Publishing Limited; 2012.
pre-charged in gaseous hydrogen remained practically unaf- [10] Oudriss A, Fleurentin A, Courlit G, Conforto E, Berziou C,
Rebe re
 C, et al. Consequence of the diffusive hydrogen
fected, irrespective of the applied displacement rate.
contents on tensile properties of martensitic steel during the
The embrittlement indexes (strength and reduction in
desorption at room temperature. Mater Sci Eng A
area) obtained using hydrogen pre-charged notch tensile 2014;598:420e8.
specimens increase with steel hardness. This trend is [11] Song EJ, Baek SW, Nahm SH, Baek UB. Notched-tensile
consistent with the change appreciated in the fracture properties under high-pressure gaseous hydrogen:
micromechanisms. MVC was the only failure micro- comparison of pipeline steel X70 and austenitic type 304L
mechanism in the grade with the lowest hardness, while 316L steels. Int J Hydrogen Energy 2017;42:8075e82.
[12] ASTM G146. Evaluation of disbonding of bimetallic stainless
MVCþPRHIC and MVCþPRHICþIG were the operative failure
alloy/steel plate for use in high-pressure, high temperature
micromechanisms in the case of the grades with intermediate
refinery hydrogen service. In: Annual Book of ASTM
hardness and the highest hardness, respectively. Standards, vol. 03.02; 2013.
The use of lower displacement rates always gave rise to [13] Neuber H, Raven FA, Brock JS. Theory of Notch
higher embrittlement indexes, as the accumulation of hydrogen Stresses:principles for exact stress calculation. Berlin: Julius
atoms in the most stressed area of the notched specimen in- Springer; 1937.
creases due to the availability of longer diffusion times. [14] Takagi S, Terasaki S, Tsuzaki K, Inoue T, Minami F.
Application of local approach to hydrogen embrittlement
Finally, these results are explained by the magnitude of the
fracture evaluation of high strength steels. Mater Sci Forum
hydrostatic and normal stresses existing in the vicinity of the 2007;539e543:2155e61.
notch (which are related with the steel yield strength), and by [15] Krauss George. Steels: processing, structure and
differences in local hydrogen accumulation within the performance. 2005.
different microstructural features. [16] Takeda Y, McMahon Jr CJ. Strain controlled vs. stress
controlled hydrogen induced fracture in a quenched
and tempered steel. Metall Mater Trans A
1981;12A:1255e66.
Acknowledgements [17] Wang M, Akiyama E, Tsuzaki K. Effect of hydrogen and stress
concentration on the notch tensile strength of AISI 4135
steel. Mater Sci Eng A 2005;398:37e46.
The authors would like to thank the Spanish Ministry of Econ-
[18] Nibur KA, Somerday BP, Marchi CSAN, Foulk JW,
omy and Competitiveness for the support received to carry out Dadfarnia M, Sofronis P. The relationship between crack-tip
research project MAT2014-58738-C3 (SAFEHIDROSTEEL). strain and subcritical cracking thresholds for steels in high-
9082 i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 3 ( 2 0 1 8 ) 9 0 6 8 e9 0 8 2

pressure hydrogen gas. Metall Mater Trans A properties of X80 pipeline steel. Mater Sci Eng A
2013;44A:248e69. 2017;700:321e30.
[19] Pressouyre GM. Trap theory of Hydrogen embrittlement. [24] Novak P, Yuan R, Somerday BP, Sofronis P, Ritchie RO. A
Acta Metall 1980;28(7):895e911. statistical physical-based, mico-mechanical model of
[20] Matsuo T, Homma N, Matsuoka S, Murakami Y. Effect of hydrogen-induced intergranular fracture in steel. J Mech
hydrogen and prestrain on tensile properties of carbon steel Phys Solid 2010;58:206e26.
SGP (0.078C-0.012Si-0.35Mn, mass%) for 0.1 MPa hydrogen [25] Nagao A, Dadfarnia M, Somerday BP, Sofronis P, Ritchie RO.
pipelines. Trans JSME A 2008;74(744):1165e73. Hydrogen-enhanced-plasticity mediated decohesion for
[21] Murakami Y, Matsuoka S. Effect of hydrogen on fatigue crack hydrogen-induced intergranular and quasi-cleavage
growth of metals. Eng Fract Mech 2010;77(11):1926e40. fracture of lath martensitic steels. J Mech Phys Solid
[22] Sun YW, Chen JZ. Investigation into hydrogen diffusion and 2018;112:403e30.
susceptibility of hydrogen embrittlement of high strength [26] Michler T, San Marchi C, Naumann J, Weber S, Martin M.
0Cr16Ni5Mo steel. J Iron Steel Res Int 2015;22(10):961e8. Hydrogen environment embrittlement of stable austenitic
[23] An T, Zheng S, Peng H, Weng X, chen L, Zhang L. Synergistic steels. Int J Hydrogen Energy 2012;37:16231e46.
action of hydrogen and stress concentration on the fatigue

View publication stats