Hydrogen Embrittlement of Automotive Advanced

High-Strength Steels
GIANFRANCO LOVICU, MAURO BOTTAZZI, FABIO D’AIUTO,
MASSIMO DE SANCTIS, ANTONELLA DIMATTEO, CIRO SANTUS,
and RENZO VALENTINI

Advanced high-strength steels (AHSS) have a better combination between strength and ductility
than conventional HSS, and higher crash resistances are obtained in concomitance with weight
reduction of car structural components. These steels have been developed in the last few decades,
and their use is rapidly increasing. Notwithstanding, some of their important features have to be
still understood and studied in order to completely characterize their service behavior. In par-
ticular, the high mechanical resistance of AHSS makes hydrogen-related problems a great concern
for this steel grade. This article investigates the hydrogen embrittlement (HE) of four AHSS steels.
The behavior of one transformation induced plasticity (TRIP), two martensitic with different
strength levels, and one hot-stamping steels has been studied using slow strain rate tensile (SSRT)
tests on electrochemically hydrogenated notched samples. The embrittlement susceptibility of
these AHSS steels has been correlated mainly to their strength level and to their microstructural
features. Finally, the hydrogen critical concentrations for HE, established by SSRT tests, have
been compared to hydrogen contents absorbed during the painting process of a body in white
(BIW) structure, experimentally determined during a real cycle in an industrial plant.

DOI: 10.1007/s11661-012-1280-8
 The Minerals, Metals & Materials Society and ASM International 2012

I. INTRODUCTION DP steels have a ferrite-martensite mixed microstruc-

ture[5] with a martensite volume fraction ranging from 20
IN the last few decades, the development of new to more than 75 pct in respect to the steel grade. The
classes of high-strength steels allowed for increasing typical tensile strengths of DP steels are between 450 and
both resistance and toughness properties in respect to 1100 MPa, and elongation is between 10 and 25 pct. The
conventional high-strength autobody steels. The trans- microstructure of TRIP steels consists of a ferritic matrix
formation-hardened steels (the so-called advanced high- with dispersed islands of hard phase (bainite and/or
strength steels) along with high-manganese austenitic martensite) and retained austenite.[5] During straining,
steels are examples of new recently developed classes.[1–4] the metastable austenitic phase transforms into martens-
In conventional steels, more traditional hardening ite (TRIP effect), increasing both strength and ductility.
methods are used, such as solid solution, grain refine- Between steels belonging to AHSS, the martensitic steels
ment, or precipitation, whereas in advanced high- have the highest strength levels. Two main groups could
strength steels (AHSS), less conventional ones, based be highlighted, the cold-rolled martensitic steels for cold
on phase transformations, are applied. These methods forming and the boron containing steels optimized for hot
allow for obtaining complex microstructures, where stamping. The typical tensile strength of martensitic steels
mechanical properties of transformed phases are, to can lie between 1000 and 1500 MPa. Despite their limited
some extent, merged. The main classes are dual phase elongation and the high springback effect related to their
(DP), transformation induced plasticity (TRIP), high grades, cold-rolled steels have great potential thanks
martensitic, complex phase steels, and others. They to their lower cost with respect to hot-stamping compo-
differ in terms of type and volume fraction of micro- nents of similar strength. On the contrary, the hot-
structural constituents. stamping process allows for obtaining high-strength
components with low residual stresses and a more
homogeneous microstructure, but with higher costs.
GIANFRANCO LOVICU, Research Assistant, MASSIMO DE Generally, by increasing the mechanical resistance of
SANCTIS and RENZO VALENTINI, Associate Professors, and steels, also their hydrogen embrittlement susceptibility
ANTONELLA DIMATTEO, Young Researcher, are with the increases.[6–8] This general behavior is observed in steels
Dipartimento di Ingegneria Chimica, Chimica Industriale e Scienza
dei Materiali, Università di Pisa, 56126 Pisa, Italy. Contact e-mail: belonging to the same class,[8] although some researchers
[email protected] MAURO BOTTAZZI and FABIO D’AIUTO, stress the role of all the parameters that influence the
Sheets Specialists, are with the Centro Ricerche Fiat Scpa, 10043 hydrogen embrittlement susceptibility, such as microstruc-
Orbassano, To, Italy. CIRO SANTUS, Assistant Professor, is with the tures, trapping state, deformation state, and strength.[9]
Dipartimento di Ingegneria Meccanica, Nucleare e della Produzione,
Università di Pisa, 56126 Pisa, Italy.
Hydrogen degradation could represent one of the
Manuscript submitted January 28, 2011. main limitations to the use of advanced high-strength
Article published online June 30, 2012 steels. In steel, hydrogen can derive from the production

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 43A, NOVEMBER 2012—4075

process, product assembling, and finishing, or from phosphatizing and electrodeposition paint (EDP) pro-
service environment exposure. If the hydrogen content cesses could be responsible for some hydrogen absorp-
reaches a critical value inside the steel, it can induce a tion.[16] The hydrogen concentration absorbed from
strong reduction of mechanical properties, particularly different tested steels has been measured using the
strength and ductility. In the case of conventional barnacle electrode method after exposition of steel
quench and tempered steels with a ultimate tensile specimens to a real industrial process.
strength (UTS) of about 1500 MPa, the critical value to
induce hydrogen embrittlement can be much less than
1 wppm (see, for example, Reference 8). Since the II. MATERIALS AND EXPERIMENTAL
largest parts of car body components are produced by METHODS
cold-forming techniques that can induce high residual
stresses, the effect of very little quantities of hydrogen A. Materials
can be catastrophic. Four different commercial AHSS have been selected in
Although some experimental studies have been carried order to investigate their hydrogen embrittlement suscep-
out on hydrogen embrittlement of AHSS,[10–14] much tibility. In particular, one hot-stamping, two martensitic,
work is still needed. In particular, the hydrogen effect on and one TRIP steels have been analyzed. All steels have
multiphase steels should be better related to the peculiar been received in their commercial coating conditions. Hot-
microstructural features of steels. Above all, when an stamping steel has been received in quenched condition by
austenitic phase is also present, such as for TRIP steels, using the same temperature and cooling rate as for the real
the effect of huge differences on transport and solubility industrial hot-stamping process.
properties of hydrogen within the austenite and the Table I shows the chemical compositions along with
ferritic/martensitic matrix, as well the austenite tendency the thickness and the coating characteristics of each
to transform to martensite during straining and/or during studied steel.
hydrogen charging, are all aspects of the problem that In order to analyze their microstructures, each steel has
needs to be thoroughly investigated. been polished and then etched by employing appropriate
One main experimental technique used to analyze the reagents depending on the expected microstructures. Nital
hydrogen effect on mechanical properties of steels is the reagent has been used for hot-stamping and martensitic
slow strain rate tensile (SSRT) test,[15] i.e., a tensile test steels, while Le Pera etchant has been used to highlight the
performed at a slow strain rate in order to enhance the complex microstructure of TRIP 800 steel.
embrittlement effect of hydrogen. The hydrogen embrit-
tlement mechanism is governed by local hydrogen
transport within the steels. The high-stressed areas (with B. Mechanical Tests
a nonzero hydrostatic component) in the sample are
subject (1) to a lattice distortion that can increase the Preliminary mechanical tests have been performed on
local hydrogen solubility and thus (2) to generate a as-received smooth tensile samples using a MTS 100 kN
chemical potential gradient that can act as a driving tensile machine (MTS Systems Corporation, Eden
force for the hydrogen diffusion. Thus, areas with higher Prairie, MN), in order to measure the mechanical
residual stresses not only are subjected to a higher stress properties of as-received steels.
level but also are richer in hydrogen, dramatically
increasing the failure susceptibility.
C. SSRT
This article analyzes the hydrogen embrittlement
behavior of four different AHSS by means of SSRT Notched tensile samples, with the geometry and
tests on electrochemically charged notched tensile sam- dimensions shown in Figure 1, have been used to carry
ples. The obtained results, along with the fractographic out SSRT tests. The presence of a notch allows for
studies of post-tensile samples, have been deeply inves- obtaining a hydrostatic stress state in its proximity, thus,
tigated in terms of their microstructural and mechanical increasing the embrittling effect of hydrogen. Before
features. Finally, in order to compare the critical hydrogenation, the sample surface had been prepared by
hydrogen concentrations determined by SSRT tests removing the coating layer using SiC paper up to 600 grit.
with the quantity of hydrogen absorbed in real condi- Different electrochemical conditions have been used
tions, particular attention has been put on the pretreat- in order to obtain hydrogen contents ranging in a wide
ment of the painting process. As a matter of fact, interval able to include the embrittlement of materials.

Table I. Thickness, Coating Type, and Chemical Composition of Tested Steels

Chemical Composition

Steel Thickness (mm) Coating Type C (Pct) Mn (Pct) Si (Pct) Cr (Pct) Ti (Pct) Al (Pct) B (wppm)
HS 1500 1.80 Al-Si 0.22 1.25 0.25 0.22 0.04 0.04 >20
M 1400 1.25 electrogalv 0.18 1.51 0.35 0.21 0.04 0.05 >20
M 1200 2.00 electrogalv 0.12 1.70 0.21 0.03 0.03 0.04 >20
TRIP 800 1.12 galv 0.19 2.12 0.63 0.03 0.01 1.02 —

4076—VOLUME 43A, NOVEMBER 2012 METALLURGICAL AND MATERIALS TRANSACTIONS A

 Table II. Main Pretreatment Painting Process
Parameters Registered During the Cycles Used for the
Hydrogen Absorption Analysis
Phosphatizing T = 323 K (50 C) pH = 4.7
Primer V = 220 ‚ 310 V pH = 6
electrodeposition
EDP curing T = 170 ‚ 190 C

dry ice or liquid nitrogen until its hydrogen content has

been measured in the laboratory.
For each sample, two different conditions of coating
layer have been analyzed: the undamaged state and the
Fig. 1—Specimen geometry and dimensions for SSRT tests. The
thickness s depends on the steel strip thickness. case of scratches presence in the protective layer. The
latter condition was chosen in order to simulate the
Aqueous solutions containing NaOH (0.1 M), H2SO4 effect of galvanic coupling between the protective layer
(0.1 N and 1 N) have been used, with the addition of and the underlying metal.
different recombination poisons (As2O3, NaAsO2). The The measurement of superficial absorbed hydrogen
cathodic current densities typically ranged from 1 to has been done by the barnacle electrode method,[20]
50 mA/cm2. Electrochemical charging time has been following the ASTM F1113 standard testing method.[21]
chosen with respect to the saturation time of each steel, The registered current output had been preliminary
depending on specimen thickness and hydrogen diffu- correlated to the hydrogen content using uniformly
sion coefficient,[17] the latter measured by a Devanathan- charged samples, whose hydrogen concentration had
Stackurski electrochemical permeation experiment.[18] been measured using the hot extraction method (using a
Coupons of the same steel have been hydrogenated LECO DH603).
contemporary to the tensile samples to have witness
specimens in order to measure the absorbed hydrogen E. Fractographic Analysis
content before tensile tests. Hydrogen concentration has
been measured using a LECO DH603 hydrogen deter- The post-tensile fractographic analysis has been
minator (LECO Corporation, St. Joseph, MI), based on performed using a JEOL 5600LV Scanning Electron
the hot extraction method. SSRT tests have been Microscope (SEM) (JEOL Ltd., Tokyo, Japan), inter-
performed following the ASTM G129 00 standard test faced with an energy-dispersive X-ray spectrometer. The
method,[19] using a strain rate of about 4 9 105 s1. aim of these observations has been to highlight the
After SSRT tests, one half of each fractured sample correlation between the hydrogen concentration and the
has been used for fractographic analysis, while the other fracture modalities of each analyzed steel.
half has been used to measure the post-tensile hydrogen
concentration by using the LECO DH603.
III. RESULTS
D. Hydrogen Absorbed During Painting Process A. Microstructural Investigation
Hydrogen uptake in autobody components may result Figure 2 shows the optical micrographs of tested
from assembling and/or finishing processes. Because of steels. Hot-stamping and martensitic steels have a
the steel sheet thinness, no residual hydrogen is expected martensitic microstructure. TRIP 800 steel shows a
from the steel production process. Therefore, in this mixed microstructure composed of ferrite, bainite/mar-
context, the painting process could represent its main tensite, and retained austenite, as expected. The Le Pera
source, as a consequence of the cathodic reactions in reagent, used for TRIP steel, allows for highlighting
water solution that takes place in phoshatizing and different phases with different colors. In particular, in
electrophoresis stages.[16] During these reactions, atomic Figure 2(c), ferrite appears light brown, bainite appears
hydrogen can form on the steel surface. The main part dark brown, and martensite and austenite appear white.
of the hydrogen generated on steel surface recombines
to form gaseous hydrogen, but some hydrogen can be
absorbed and diffuse into the steel. B. Mechanical Properties
Some samples of the tested steels have been prepared Main mechanical properties, measured by tensile tests
and connected in a commercial body in white (BIW) on smooth samples, are summarized in Table III. The
structure in order to test the real cycle conditions. Tests measured mechanical properties agree with the expected
have been performed in an industrial plant, during an values for all the tested materials.
usual production cycle. The main pretreatment process Figure 3 shows the engineering stress–strain curves
parameters are summarized in Table II. Three steps of obtained from smooth and notched tensile samples in
the coating process have been analyzed: phosphatizing, the as-received conditions, without hydrogenation.
EDP, and the final curing stages. After each step, three As expected, the mechanical behavior of the tested
coupons of each steel have been extracted and stored in steels is rather different. Hot-stamping and martensitic

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 43A, NOVEMBER 2012—4077

Fig. 2—Microstructures of tested advanced high strength steels. (a) Hot-stamping steel (nital reagent); (b) M 1400 steel (nital reagent); (c) TRIP
800 steel (Le Pera etchant): ferrite in light brown, bainite in dark brown, and martensite and austenite in white; and (d) M 1200 steel (nital
reagent).

Table III. Mechanical Properties of the Analyzed Steels and the post-tensile samples measured by using a region
in the As-Received Conditions close to the fracture surface. Since samples are notched,
the total elongation is short and thus so is the total time of
UTS (MPa) YS (MPa) Elongation (pct) the SSRT test. Thus, the measured differences between
pretensile and post-tensile tests have been always less
HS 1500 1615 1550 6
M 1400 1520 1410 5 than 10 pct. Only three curves for each steel are shown in
M 1200 1305 1205 9 Figure 4: one relative to the as-received (without hydro-
TRIP 800 815 620 26 gen) sample, one for the sample with the maximum tested
hydrogen concentration, and another one for a medium
YS: yield strength.
level.
It can be easily observed that the hydrogen concen-
tration ranges are different for different steels. This is
steels, which have a martensitic microstructure, show because of the different hydrogen solubility in them and
higher mechanical strength and lower ductility. On the their uptake abilities, which are strictly dependent on the
contrary, TRIP 800 steel, characterized by a large microstructure and the electrochemical properties of
strain-hardening effect associated with the presence of each steel.
austenite in its microstructure, has very high ductility In stress–strain diagram, hydrogen acts by reducing
and lower strength. the extension of curve, without any (or with very poor)
modification before the necking point. In the interna-
tional scientific literature, different sensitive parameters
C. SSRT are used as an indicator of hydrogen effect on steels,
Figure 4 shows the engineering stress–strain curves such as the fracture stress or the ductility (elongation or
obtained for notched samples when charged with differ- area) reduction.[10,14,22,23] This last parameter is better in
ent hydrogen contents. The reported hydrogen concen- the case of ductile and low work-hardening steels, where
trations are the average values between the witness the fracture stress does not sensitively change with
samples (hydrogenated together with tensile samples) hydrogen content changes. On the contrary, for ultra
representative of the pretensile hydrogen concentration high-strength steels, the elongation to fracture is not a

4078—VOLUME 43A, NOVEMBER 2012 METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 3—Engineering stress–strain curves for the tested AHSS in the as-received (hydrogen uncharged) conditions. (a) Stress–strain curve for
smooth samples. (b) Stress–strain curve for notched samples.

Fig. 4—Stress–strain curves of tested steels electrochemically charged at different hydrogen contents: (a) hot-stamping steel, (b) M 1400 steel,
(c) M 1200 steel, and (d) TRIP 800 steel.

useful parameter because the hydrogen can induce a tensile specimen further decreases the elongation to
fracture in the elastic range, where the elongation of fracture sensitivity to hydrogen embrittlement (HE).
samples with different hydrogen content tends to show Thus, the authors preferred to limit the analysis to
very similar values. Moreover, the presence of notch in fracture stress parameter.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 43A, NOVEMBER 2012—4079

Fig. 5—Hydrogen embrittlement curves of tested steels: (a) hot-stamping steel, (b) M 1400 steel, (c) M 1200 steel, and (d) TRIP 800 steel. For
each steel, the hydrogen critical concentration for embrittlement and the embrittlement index are reported. Some of the M1400 steel data are
shown as hollow point in order to differentiate them in respect to the others. The meaning of this difference is described in text (detailed in
Section IV–B). N.UTS is the UTS measured for notched specimen.

Figure 5 shows the so-called embrittlement curve of value has been chosen as hydrogen critical concentration
each steel, i.e., the strength vs hydrogen concentra- CHcr.
tion data. Strength values pass from values close to the In Figure 5, the hydrogen critical concentration
as-received steel strength (for low hydrogen concentra- values CHcr related to each steel are reported, along
tions) to a lower plateau at higher hydrogen concentra- with the embrittlement index (EI) calculated as follows:
tions. These two plateaus are separated by a marked
transition region. It is worth noting that this behavior is N:UTSnoH  N:UTSmax H
EI ¼  100
common for all the steel grades, from ductile to high- N:UTSnoH
strength steels.[22,24,25] In order to interpolate the exper- where N.UTSnoH and N.UTSmaxH are the UTS of
imental data the following expression was used[22]: notched specimens values of the uncharged sample and
CH  c correspond with the maximum tested hydrogen concen-
N:UTS ¼ a  b  arctan tration, respectively.
d
The embrittlement index is the percentage of N.UTS
where N.UTS is the UTS of notched specimen; CH is the reduction measured at high hydrogen content in the
hydrogen concentration; and a,b,c, and d are fitting concentration range correspondent to the low strength
parameters. In particular, the c parameter is related to plateau. It is an indication of the maximum hydrogen
the position of the inflection point (the middle of the effect on steel strength.
transition region), d to the transition width, and a and b Results reported in Figure 5 show that the maximum
parameters to the plateau values. EI value is reached for the hot-stamping steel (about
As shown in Figure 5, the use of the ‘‘arctangent’’ 75 pct), whereas lower EI values (about 25 pct) are
function reproduces the experimental data well. measured for the other ones.
In order to identify the concentration at which value Concerning the hydrogen critical concentration for
of the steel is affected by hydrogen embrittlement, the c embrittlement, the lowest value was found for M

4080—VOLUME 43A, NOVEMBER 2012 METALLURGICAL AND MATERIALS TRANSACTIONS A

1400 steel (about 1 wppm), while the other steels show Figure 6, collected data do not show any specific trend
higher values. The critical hydrogen concentration of related to the coating layer wholeness or to the painting
TRIP 800 steel is about 2.5 wppm, and it is about cycle step.
4 wppm for M 1200 and hot-stamping steels.
The embrittlement curve for M 1400 steel exhibits a
further reduction of N.UTS after the low-strength IV. DISCUSSION
plateau. This particular feature, related to a change in
hydrogen damage modality, will be deeply discussed in In Table IV, the main data extracted from SSRT tests
Section IV. are summarized.
It is worth noting that using very similar electro-
chemical charging conditions, the hydrogen concentra-
D. Hydrogen Absorbed During Painting Process
tion is very different from steel to steel. This feature
Figure 6 shows the hydrogen concentration values mainly depends on different microstructural features.
measured from samples subjected to the real painting Also, martensitic steels with similar strength (HS1500
pretreatment. The reported measurement values are the and M1400) showed very different hydrogen uptake
average values calculated from three different samples. ability. This is probably due to their different production
All samples show an absorbed hydrogen concentra- process. HS 1500 is produced by the hot-rolling process,
tion lower than about 0.4 wppm. As clearly visible in while M1400 steel is produced by the cold-rolling

Fig. 6—Hydrogen concentration, measured by the electrochemical Barnacle method, absorbed during the following painting line steps: phospha-
tizing, electrodeposition painting, and final curing. (a) Results for undamaged surface and (b) results for scratched surface.

Table IV. Hydrogen Concentrations and Tensile Properties Measured in Function of Electrochemical Charging
for all the Tested Steels

Charging Conditions
*
Steels Parameters Low High Very High
HS 1500 C (ppm) 1.8 ‚ 5.2 5.1 ‚ 8.6
N.UTS Red.§ (pct) 5 ‚ 65 65 ‚ 78
El. Red.|| (pct) 10 ‚ 55 75 ‚ 90
M 1400 C (ppm) 0.4 ‚ 1.8 1.5 ‚ 3.3 2.5 ‚ 3.6
N.UTS Red.§ (pct) 0 ‚ 25 20 ‚ 42 25 ‚ 48
El. Red.|| (pct) 10 ‚ 15 35 ‚ 50 35 ‚ 60
M 1200 C (ppm) 0.9 ‚ 4.7 4.0 ‚ 5.8
N.UTS Red.§ (pct) 5 ‚ 28 23 ‚ 25
El. Red.|| (pct) 0 ‚ 55 35 ‚ 65
TRIP 800 C (ppm) 1.7 ‚ 3.5 2.6 ‚ 27.8
N.UTS Red.§ (pct) 10 ‚ 20 17 ‚ 27
El. Red.|| (pct) 0 ‚ 75 40 ‚ 85
*(Low): NaOH 0.1 N + As2O3 25 mg/L with current density between 1 and 10 mA/cm2.

(High): H2SO4 0.1 N + As2O3 25 mg/L with current density between 1 and 50 mA/cm2.

(Very high): H2SO4 1 N + NaAsO2 5Æ104 N with current density between 10 and 50 mA/cm2.
§
N.UTS Red. = Reduction of notched ultimate tensile strength in respect to uncharged sample.
||
El. Red. = Reduction of elongation to fracture in respect to uncharged sample.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 43A, NOVEMBER 2012—4081

process. Similar martensitic microstructure, little differ- 6 9 107 cm2/s. Using similar D values, for a 1.7 mm
ences on its fineness, hardening, or chemical composi- in thickness sample, a time of about 6 hours had been
tion can induce heavy differences in hydrogen uptake needed to reach uniform concentration across the
ability. This is confirmed by all the performed tests, both specimen.[17] Finally, the low toughness of a completely
for the electrochemical charging of SSRT specimens and martensitic steel causes the rupture of samples at the
for the hydrogen absorbed during painting, where HS onset of subsurperficial cracks, even if the bulk material
1500 steel showed final hydrogen concentration values does have a hydrogen concentration much lower than
much higher than M1400 and M1200 steels. Obviously, the superficial one.
the presence of austenite phase in TRIP steel makes As shown in Figure 7(a), the fractographic analysis
difficult the comparison with other steels in terms of showed ductile fracture for the as-received sample.
hydrogen diffusivity and solubility. These differences Increasing the hydrogen content, a typical hydrogen
induce also high differences in the hydrogen concentra- induced quasi-cleavage fracture appears (Figure 7(b))
tion range where the decreasing of mechanical proper- with the presence of some fine secondary cracks. At a
ties appears, as visible in Figure 5. hydrogen concentration higher than the critical one, the
In order to highlight the main hydrogen embrittle- secondary cracks enlarge and some intergranular com-
ment properties, the results of each tested steel will be ponent of primary crack can be highlighted, as visible in
discussed in the following. Figure 7(c).
Post-tensile samples have been sectioned in order to
highlight the microstructural characteristic of the crack-
A. Hot-Stamping Steel
initiation sites. Figure 8 shows some of these character-
Although all steel grades can suffer hydrogen embrit- istics on the hot-stamping steel sample charged with
tlement,[9] high-strength steels are usually more prone to 5.2 wppm of hydrogen. Both transgranular and inter-
this problem. As expected, hot-stamping steel, charac- granular fractures have been found as shown in Fig-
terized by a completely martensitic microstructure and ures 8(a) and (b), respectively. The analysis has not
very high strength levels (N.UTS of about 1600 MPa), shown any particular correlation between the crack
shows a very high embrittlement index. The degradation onset and the microstructural features of HS steel.
of mechanical properties is very marked; for example,
the strength of notched specimen with a hydrogen
concentration of about 9 wppm is about 400 MPa, with B. Martensitic Steels
a strength reduction of more than 75 pct. The embrittlement curve of M 1400 steel shows a
The strength reduction of hot-stamping steels similar critical concentration of about 1 wppm and an embrit-
to that presented in this article has been reported in the tlement index of about 25 pct.
literature.[11,14] Lee et al.[11] reported UTS values of In order to highlight the main factor influencing the
about 400 MPa for samples charged to 1.7 wppm. The hydrogen embrittlement susceptibility of martensitic
discrepancy in hydrogen content that has caused a steels, a deep microstructural analysis has been per-
similar EI in the Lee et al. work and the present data can formed on longitudinal section of post-tensile samples in
be explained considering the difference in electrochem- regions close to the failure surface. A great part of
ical charging conditions. They charged 1.7 mm in hydrogen cracks has been found to start from inclusion
thickness tensile samples for only 30 minutes, but this particles. The micrograph in Figure 9 shows a typical
time, according to the hydrogen diffusion coefficient (D) hydrogen crack generated on an inclusion particle for a
value, is not enough to produce a homogeneous sample of M 1400 steel charged at 3.0 wppm of hydro-
hydrogen concentration inside the samples. Thus, close gen. The dangerous shape factor of elongated inclusions
to the steel surface the hydrogen concentration could be can contribute to an increase in the hydrogen embrittle-
much higher than in the center part. In the present work, ment susceptibility.
D was calculated by using a Devanathan-Stackurski[18] Performing energy-dispersive X-ray spectroscopy
electrochemical permeation apparatus, following the microanalysis on inclusions, they have resulted to be
ASTM standard G 148-97.[26] Its value was about constituted mainly by Ca, Si, Al, and Ti particles.

Fig. 7—Micrographs of fracture surface of hot-stamping steel samples charged with different hydrogen contents. (a) Uncharged specimen,
(b) sample electrochemically charged at 2.6 wppm, and (c) sample electrochemically charged at 7.1 wppm.

4082—VOLUME 43A, NOVEMBER 2012 METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 8—Micrographs showing the longitudinal section of post-tensile hot-stamping steel sample (with 5.2 wppm) in the region close to the
fracture surface. (a) Transgranular cracks and (b) intergranular cracks.

elliptical cracks (Figure 11(d)). Deep secondary cracks are

also clearly visible. This behavior has been probably due to
the formation of hydrogen-induced cracks during the
electrochemical charging, which greatly lower the sample
strength. The failure of a sample with a concentration of
about 4 wppm during electrochemical charging confirms
the abovementioned explanation. The hydrogen induced
cracks formed during electrochemical charging obviously
have not the same extension and morphology in different
samples. Thus the fracture stress measured in tensile test,
heavely influenced by the presence of hydrogen induced
cracks, have resulted very scattered. The authors have
decided to not include these data in the embrittlement
curve fitting, since this curve is valid for stress-assisted
hydrogen embrittlement.
In the present explored experimental conditions, the
hydrogen-induced cracks generated during electrochem-
Fig. 9—Example of hydrogen-induced crack started from an inclu- ical charging have been found only for the M
sion particle, in a M 1400 steel electrochemically charged at 1400 steels, although the authors cannot exclude the
2.7 wppm. possibility to induce hydrogen cracking on other steels
at much higher hydrogen concentrations.
Figure 10 shows the fracture surface for M 1400 sam- M 1200 steel shows an embrittlement curve similar to M
ples with different hydrogen contents. The uncharged 1400 steel. The lower strength level determines a lower
sample (Figure 10(a)) presents a ductile fracture, while hydrogen embrittlement susceptibility, shifting the hydro-
hydrogenated samples are subjected to brittle fracture. gen critical concentration to higher values (about 4 wppm).
For hydrogen concentration lower than about 2 wppm, Concerning the general hydrogen embrittlement
samples have shown a mixed transgranular–intergranu- behavior of very high-strength steels (with strength level
lar fracture (Figure 10(b) for a 1.2 wppm charged higher than 1200 MPa) could be important to highlight
sample), while for higher hydrogen content, there has the role of stamping modalities.
been a predominance of intergranular fracture, as shown The hydrogen embrittlement susceptibility of high-
in Figure 10(c) for a sample charged at 3.0 wppm, as grade martensitic steels could be increased by the fact
expected for the increasing of hydrogen activity.[27] that auto body components are produced by the cold
After the high concentration plateau, a further stamping process that gives high residual stresses due to
decrease of the strength level is clearly visible for the springback effect.[28] Although the hot-stamping
M1400 steel, showed in Figure 5(b) as hollow points. steel (comparable in terms of strength and ductility) has
Post-tensile fractographic analysis (Figure 11) showed a much higher embrittlement index, the almost com-
that at a hydrogen concentration higher than about pletely absent of residual stresses derived by the hot-
3 wppm, the fracture surface changes appearance. stamping process can play favorably on its hydrogen
For lower hydrogen concentration, the low magnifica- embrittlement susceptibility.
tion fracture appearance is that typical of notched samples
with a triangular area close to the notch associated with the
C. TRIP Steel
stress intensification region (Figures 11(a) through (c)).
For higher concentration, this morphology has disap- The presence of austenite deeply influences the hydro-
peared and the fracture surface is composed of several gen behavior in steel. Thanks to the very high solubility

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 43A, NOVEMBER 2012—4083

Fig. 10—Micrographs of fracture surface of M 1400 steel samples charged with different hydrogen contents. (a) Uncharged specimen, (b) sample
electrochemically charged at 1.2 wppm, and (c) sample electrochemically charged at 2.7 wppm.

Fig. 11—SEM micrographs showing the fracture surface of M 1400 post-tensile samples with different hydrogen contents. (a) Uncharged speci-
men, (b) 0.9 wppm, (c) 2.3 wppm, and (d) 3.4 wppm.

and low diffusivity of hydrogen in it, austenite acts as a where hydrogen can induce martensitic transformation
sink for hydrogen, lowering its mobility and increasing and related failure (for example, see Reference 29).
the final hydrogen concentration. Thus, in TRIP steel, Probably, this occurs because the embrittlement curve
the hydrogen concentration in austenite is much higher of TRIP 800 steel does not present a marked plateau in
than in the other phase. the low concentration region, as is clearly visible in
During the deformation of TRIP steels, austenite Figure 5.
transforms into martensite (TRIP effect), which is the The microstructural examination of post-tensile sam-
microstructure most susceptible to hydrogen embrittle- ples electrochemically charged with hydrogen has shown
ment. Thus, it is fair to say that just after the application the presence of crack onsets in correspondence of the
of stress, the hydrogen-enriched transformed martensite hard particles, as shown for example in Figure 12. This
quickly cracks, inducing the failure of the specimen. result is in agreement with the above suggested fracture
Moreover, the not stable austenite of TRIP steels could mechanism. Moreover, other research highlighted sim-
be more and more unstabilized by the presence of ilar crack initiation mechanisms. Using EBSD tech-
hydrogen, as happens for austenitic stainless steels niques, Imlau et al.[26] found a high correlation between

4084—VOLUME 43A, NOVEMBER 2012 METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 12—Micrographs showing the longitudinal section of post-tensile TRIP 800 steel sample with different hydrogen contents in the region close
to the fracture surface: (a) 3.5 wppm and (b) 28 wppm.

Fig. 13—Micrographs of fracture surface of TRIP 800 steel samples charged with different hydrogen contents. (a) Uncharged specimen, (b) sam-
ple electrochemically charged at 3.5 wppm, (c) sample electrochemically charged at 9.1 wppm, and (d) sample electrochemically charged at
28 wppm.

the region where cracks form and martensite transfor- above all for concentration higher than the critical one
mation areas. Moreover, Ronevich et al.[14] found some (Figure 13(b) through (d)).
cracks starting from hard particles (martensite islands) Moreover, it is important to note that the ductile
and propagating into ferrite grains. microstructure of TRIP 800 steel and the low tested
Fractographic analysis has shown that fracture grade (compared with the other tested steels) limit the
surfaces become more and more brittle by increasing strength reduction in presence of high hydrogen
the hydrogen content, as shown in Figure 13. The concentration to about 25 pct.
uncharged sample presents a completely ductile fracture Despite the presence of austenitic phase that is usually
(Figure 13(a)). After increasing the hydrogen content, considered to have higher hydrogen embrittlement
the brittle percentage of fracture surface increases, resistance thanks to its higher toughness and its lower

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 43A, NOVEMBER 2012—4085

hydrogen diffusivity, TRIP 800 steel shows some sus- short time of process, was able to absorb hydrogen only
ceptibility to hydrogen embrittlement in respect to its in the surface). No cracks have been found in all of the
low grade. The high hydrogen uptake ability due to tested steels, thus confirming the absence of hydrogen
austenite presence in microstructure, along with the embrittlement associated to the autobody production
tendency of this phase to transform into an embrittled process. Nevertheless, further investigations will be
martensitic phase, make this class of steel suscetible to performed in order to highlight the possible hydrogen
hydrogen embrittlement, at least in case of high hydro- uptake during the car service life.
gen external activity and stress level.

D. Hydrogen Absorbed During Body in White V. CONCLUSIONS

Painting Process
From the above-mentioned tests performed on AHSS,
The hydrogen critical concentrations measured by it is possible to highlight the following conclusions:
SSRT tests are not an exhaustive expression of the
1. Hot-stamping steel has shown a heavy reduction of
hydrogen embrittlement susceptibility of a steel. This
tensile strength (more than 75 pct) for a hydrogen
value has to be correlated to the hydrogen concentration
concentration of about 7 wppm. As expected, its
absorbed in the real conditions during the production
fully martensitic microstructure is highly susceptible
process or service life of the steel components.
to hydrogen embrittlement.
The measurement of hydrogen content absorbed during
2. Martensitic steels with a very high tensile strength
the painting pretreatment processes of body in white has
have shown a hydrogen critical concentration of
been performed in an industrial plant during a real
about 1 and 4 wppm for M 1400 and M 1200 steels,
treatment. In this way, the coupons of the analyzed steels
respectively. In the analysis of crack initiation sites, a
have exactly followed the same process of real cars. As
predominance of hydrogen-induced defects in corre-
already detailed, two types of samples have been used, one
spondence of inclusion particles has been found.
with the coating layer undamaged and another one with
These defects can contribute to the hydrogen embrit-
scratches, in order to simulate damage on the coating layer.
tlement susceptibility of this steel class. Moreover, M
As reported in Figure 6, the measured hydrogen
1400 steel has demonstrated a high hydrogen-induced
contents have been always less than 0.4 wppm in all
cracking susceptibility at a hydrogen concentration
the tested conditions (after phosphatizing, after EDP,
higher than about 3 wppm, with the formation of
and after EDP curing line steps). This means that for all
cracks during the electrochemical charging. M
the tested steels, the absorbed hydrogen is much lower
1200 steel, thanks to its lower resistance, has a higher
than the hydrogen critical concentration. Hot-stamping
hydrogen critical concentration, and it does not show
and TRIP 800 steels showed the highest values of
a tendency to crack in the absence of external stresses.
absorbed hydrogen, while martensitic steels showed
3. Although the presence of the austenitic phase is in
lower tendency to be hydrogenated. Concerning TRIP
its microstructure, the analyzed TRIP 800 steel has
800 steel, its hydrogen uptake ability was derived above
shown some susceptibility to hydrogen embrittle-
all by the austenitic phase present in the microstructure.
ment. Its hydrogen critical concentration is about
The measured hydrogen concentrations show the same
2,5 wppm. This behavior has been correlated to the
differences observed during electrochemical charging of
presence of austenite that transforms into martens-
tensile samples, being much higher for hot-stamping and
ite during deformation. The high hydrogen solubil-
TRIP 800 steels than martensitic steels, highlighting a
ity in this phase can increase the tendency to
different hydrogen uptake ability. No trend is visible
transform and can produce a hydrogen-rich mar-
either in function of the steel grade, or of the cycle steps.
tensitic phase that quickly cracks during straining.
Considering the highest hydrogen concentration mea-
4. The hydrogen concentration absorbed during the
sured for each steel and using the interpolation formula
pretreatment of the painting cycle has shown values
of the embrittlement curve, it is possible to calculate the
much lower than the critical concentrations in all
strength of materials after the painting process. The
the analyzed conditions. Moreover, no hydrogen-in-
strength reduction is less than 1 pct for all the tested
duced delayed cracks have been found on the bend-
steels. Thus, de facto, they do not suffer hydrogen
ed region of the same samples. These results suggest
embrittlement after the BIW production process.
that the tested AHSS can be used to build BIW
Moreover, since BIW is mainly assembled using cold-
components with a good safety factor.
formed components, some coupons of the analyzed
steels have been bended up to reach the 80 pct of the
ultimate real stress. REFERENCES
In order to analyze their hydrogen-delayed cracking 1. International Iron & Steel Institute Committee on Automotive:
susceptibility during painting, these coupons followed Advanced High Strength Steel (AHSS) Application Guidelines,
the same painting line as for the measurement of Version 3, 2006, www.worldautosteel.org.
absorbed hydrogen. Then bended samples were taken 2. C. Federici, S. Maggi, and S. Rigoni: Proc. Conf. on New Devel-
into observation for two weeks in order to register the opments on Metallurgy and Applications of High Strength Steels,
Buenos Aires, Argentina, 2008.
time to eventual cracks formation. The time of two 3. Y. Mukai: Kobelco Tech. Rev., 2005, no. 26, pp. 26–31.
weeks is greater than that needed to make the hydrogen 4. B.C. De Cooman, L. Chen, H.S. Kim, Y. Estrin, S.K. Kim, and H.
concentration zero in the coupon (that, thanks to the Voswinckel: Proc. Conf. on New Developments on Metallurgy and

4086—VOLUME 43A, NOVEMBER 2012 METALLURGICAL AND MATERIALS TRANSACTIONS A

 Applications of High Strength Steels, Buenos Aires, Argentina, 19. ASTM G129 – 00 (reapproved 2006): Standard Practice for Slow
2008. Strain Rate Testing to Evaluate the Susceptibility of Metallic
5. W. Bleck and K. Phiu-On: Mater. Sci. Forum, 2005, vol. 500/501, Materials to Environmentally Assisted Cracking, ASTM Interna-
pp. 97–114. tional, West Conshohoken, PA, 2006.
6. R.A. Oriani, J.P. Hirth, and M. Smialowski: Hydrogen 20. D.A. Berman and V.S. Agarwala: ASTM-STP 962 Hydrogen
Degradation of Ferrous Alloys, Noyes Publications, Park Ridge, Embrittlement: Prevention and Control, L. Raymond, ed., ASTM
NJ, 1985. STP 962 Philadelphia, PA, 1988, pp. 98–104.
7. J.P. Hirth: Metall. Trans. A, 1980, vol. 11A, pp. 861–90. 21. ASTM F 1113 – 87 (reapproved 2005): Standard Test Method for
8. ASM Materials Handbook: Metals Handbook, 9th ed., vol. 13, Electrochemical Measurement of Diffusible Hydrogen in Steels
ASM International, Materials Park, OH, 1987, p. 330. (Barnacle Electrode), ASTM International, West Conshohoken,
9. I.M. Bernstein: Proc. 5th Int. Conf. on the Effect of Hydrogen on PA, 2005.
the Behavior of Materials, A.W. Thompson and N.R. Moody, eds., 22. M. Beghini, G. Benamati, L. Bertini, I. Ricapito, and R. Valentini:
TMS, Warrendale, PA, 1996, pp. 3–11. J. Nucl. Mater., 2001, vol. 288, pp. 1–6.
10. T.B. Hilditch, S-B. Lee, J.G. Speer, and D.K. Matlock: SAE SP, 23. E. Akiyama, S. Li, Z. Zhang, M. Wang, K. Matsukado, K. Tsu-
2003, vol. 1764, pp. 47–56. zaki, and B. Zhang: Proc. 2008 Int. Hydrogen Conf., B. Somerday,
11. S-J. Lee, J.A. Ronevich, G. Krauss, and D. Matlock: ISIJ Int., P. Sofronis, and R. Jones, eds., ASM International Materials
2010, vol. 50 (2), pp. 294–301. Park, OH, 2009, p. 54–61.
12. K.H. So, J.S. Kim, Y.S. Chun, K.T. Park, Y-K. Lee, and C.S. Lee: 24. R. Valentini, A. Solina, S. Matera, and P. De Gregorio: Metall.
ISIJ Int., 2009, vol. 49 (12), pp. 1952–1959. Mater. Trans. A, 1996, vol. 27A, pp. 3773–80.
13. H. Mohrbacher: Proc. Conf. on Steel Product Metallurgy and 25. G. Lovicu, M. De Sanctis, A. Dimatteo, R. Valentini, and P.
Application, Pittsburgh, PA, 2008. Trombetti: 32 Convegno Nazionale AIM, Ferrara, Italy, 2008.
14. J.A. Ronevich, J.G. Speer, and D.K. Matlock: SAE Int. J. Mater. 26. ASTM G148-97 (reapproved 2003), Standard Practice for Evalu-
Manuf., 2010, vol. 3 (1), pp. 255–67. ation of Hydrogen Uptake, Permeation, and Transport in Metals by
15. W.J. Pollock: ASTM-STP 962, Hydrogen Embrittlement: Preven- an Electrochemical Technique, 2003.
tion and Control, L. Raymond, ed., ASTM STP 962, Philadelphia, 27. C.J. McMahon, X. Liu, J. Kameda, and M.J. Morgan: Proc. 2008
PA, 1988, pp. 68–80. Int. Hydrogen Conf., B. Somerday, P. Sofronis, and R. Jones, eds.,
16. I. Krylova: Progr. Org. Coating., 2001, vol. 42, pp. 119–31. ASM International, Materials Park, OH, 2009, pp. 46–53.
17. J. Cranck: The Mathematics of Diffusion, 2nd ed., Oxford Science 28. Die Design Handbook, Society of Manufacturing Engineers
Publications, Oxford, U.K., 1975, pp. 49–51. (SME), Dearborn MI, 1990, pp. 6–9.
18. M.A.V. Devanathan and Z. Stachurski: Proc. R. Soc. London, Ser. 29. S.M. Teus, V.N. Shyvanyuk, and V.G. Gavriljuk: Mater. Sci. Eng.
A, 1962, vol. 270, no. 1340, pp. 90–102. A, 2008, vol. 497, pp. 290–94.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 43A, NOVEMBER 2012—4087