Journal of Alloys and Compounds 720 (2017) 451e459

Contents lists available at ScienceDirect

Journal of Alloys and Compounds

journal homepage: http://www.elsevier.com/locate/jalcom

Mechanisms of hydrogen trapping in austenitic, duplex, and super

martensitic stainless steels
R. Silverstein*, D. Eliezer
Department of Material Engineering, Ben-Gurion University of the Negev, Beer-Sheva, Israel

a r t i c l e i n f o a b s t r a c t

Article history: Hydrogen interaction with structural materials, especially stainless steels, is of great importance due to
Received 23 February 2017 the acute effect that it can have on them. Stainless steels have become very common in many applica-
Received in revised form tions, and in water and high pressure environments in particular, due to their high levels of corrosion
11 May 2017
resistance and broad range of strength. Steel's durability is very much dependent on its microstructure
Accepted 27 May 2017
and interaction with hydrogen. The action of hydrogen can lead to changes in mechanical properties,
Available online 29 May 2017
phase transformation and eventually to environmentally-assisted failure, which is known as hydrogen
embrittlement (fracture). The susceptibility of steels to this hydrogen fracture mechanism is directly
Keywords:
Hydrogen trapping
related to the interaction between traps (defects) and hydrogen. In this research, we study hydrogen
Duplex stainless steel fracture mechanisms through hydrogen interaction with trapping sites by thermal desorption spec-
Super martensitic stainless steel trometry (TDS), and the calculation of hydrogen trapping energies states. Microstructure effects on
Austenitic stainless steel hydrogen were investigated by exploring different stainless steels, including: austenitic stainless steel
Thermal desorption spectrometry (TDS) (AUSS), ferritic-austenitic (duplex) stainless steel (DSS), and super martensitic stainless steel (SMSS). The
objective of this study is to determine the inuences of thermal desorption analysis on the crystal
structure of different stainless steels in order to better understand the trapping mechanisms of hydrogen
in a variety of structure materials. It was found that the AUSS has the greatest stability of austenitic (g)
phasee ~22% higher than DSS and ~45% higher than SMSS. Moreover, the AUSS presented the lowest
hydrogen trapping values of ~31% compared with DSS and ~25% compared with SMSS.
Hydrogen fracture mechanism was found to be highly dependent on the hydrogen trapping states and
even more on the g-phase stability. The hydrogen trapping mechanisms are discussed in detail.
2017 Elsevier B.V. All rights reserved.

1. Introduction Understanding the hydrogen embrittlement phenomenon in

Stainless steels are the most common structural material for a
variety of technologies and applications, particularly for oil, gas,
pipeline and pressure applications e all of these industries combine
hydrogen and mechanical load [1,2]. These materials are very
attractive due to their high level of corrosion resistance and varied
mechanical properties. Problems in maximizing those steels' full
potential may arise from their interaction with hydrogen. In many
applications these steels are exposed to hydrogen, which may lead
to a deleterious effect known as hydrogen embrittlement. The base
condition of this fracture mechanism is highly dependent on the
dislocation process and is determined by hydrogen distribution
combined with the local stress state in the material [3e5].
* Corresponding author.
E-mail address: barrav@post.bgu.ac.il (R. Silverstein).
the variety of stainless steels requires true insight of hydrogen
solubility, diffusivity and distribution within their different micro-
structures, in particular inside phases and defects [6e8]. Since
steel's defects (trapping sites for hydrogen) affect the steel's diffu-
sivity, the crack initiation is highly dependent on the trapping en-
ergy states. It is known from earlier publications [9e14] that traps
with activation energy
60 kJ/mol (also characterized as reversible
traps) will have a major inuence on the material's susceptibility to
hydrogen embrittlement. In this work we study hydrogen trapping
behavior in three different stainless steels: austenitic stainless steel
(316), ferritic-austenitic (duplex) stainless steel (SAF 2205), and
super martensitic stainless steel (SMSS).
The purpose of this work is to study hydrogen thermal
desorption behavior in order to better understand it's trapping
mechanisms in a variety of structural materials.
We study the hydrogen trapping behavior with thermal
desorption spectrometry (TDS), and hydrogen trapping energies

http://dx.doi.org/10.1016/j.jallcom.2017.05.286
0925-8388/ 2017 Elsevier B.V. All rights reserved.
452 R. Silverstein, D. Eliezer / Journal of Alloys and Compounds 720 (2017) 451e459

are analyzed by the Lee and Lee model [15]. The analyzed data was
conrmed by X-ray diffraction (XRD) and microstructural
observations.

2. Experimental procedure

2.1. Microstructure analysis

In this research we used three different stainless steels: duplex

stainless steels (DSS), SAF 2205, austenitic stainless steel (AUSS),
316L, and super martensitic stainless steel (SMSS). These steels
were received in the fully annealed condition, as a 2 mm thick plate.
The chemical compositions of these steels are presented in Table 1.
These chemical compositions were conrmed to belong to the
aforementioned structures according to the attached Scheler di-
agram, Fig. 1.
The hydrogenation technique was electrochemical (cathodic)
charging. The charging was performed at room temperature (RT) in
a 0.5 N H2SO4 (sulfuric acid) water solution and 0.25 g l1 of NaAsO2
(sodium arsenide) with a constant current density of 50 mA cm2 Fig. 1. Scheler diagram for different steel compositions.
for 72 h (h). The microstructure, phase composition, and lattice
parameter of DSS were examined both before and after hydroge-
nation, by means of an X-ray diffractometer (XRD). XRD patterns Fig. 2. This pattern is compared to an uncharged sample. The un-
were measured at RT using a Philips PW 1050/70 diffractometer, charged DSS (black line) displays ferritic-austenitic phases (a/g) in
operating at a power of 40 kV and 30 mA, with Cu anode, and Ni- equal amounts of about 50% for each phase.
ltered generating Cu-Ka radiation (1.54 ). Data collection was The hydrogenation process reveals the signicant changes in the
performed by step scanning of the sample in a 2q range between g-phase compared with those of the a-phase. The charged samples
30 and 100 at steps of 0.02 with 4e10 s per step depending on showed a signicant decrease in g-phase intensity and shifts to
peak intensity, and a scan rate of ~0.3 /min. small 2q values; the latter change indicates a larger lattice param-
The charged specimens were aged at RT and reexamined after eter, which can be explained by hydrogen solubility in that phase.
one month, in order to observe the stability of hydrogen-induced The signicant changes between both phases are related to the
phase transformations and lattice parameters after hydrogen lower diffusion coefcient and the greater solubility of hydrogen in
desorption. the g-phase [16,17]. These changes were seen simultaneously to the
formation of an additional -martensite phase, an hcp solid solution
phase, and g*-phase, an additional fcc phase with a 2% higher
2.2. Thermal desorption spectrometry (TDS)

The characteristics of hydrogen desorption, and trapping states

were investigated by means of TDS. This technique involves accu-
rate measurement of the desorption rate of hydrogen atoms, as
solute or trapped in the material, while treating the sample by non-
isothermal heating at a known rate under UHV, ~10 mPa. In this
work, the samples were heated from RT to 500  C at constant
heating rates of 2  C/min, 4  C/min and 6  C/min. The mass spec-
trometer was operated under the fast multiple mode detection; the
measured intensity channel was set to 2 amu in order to detect
hydrogen desorption. The working procedure, as described else-
where [15], allowed for the identication of different types of traps
that coexist in the specimen.

3. Results and discussion

3.1. Microstructure changes in the presence of hydrogen

3.1.1. Duplex stainless steel

The XRD diffraction pattern obtained from 72 h cathodic Fig. 2. Phase quantities of 72 h cathodic charged: DSS, AUSS and SMSS, after one
hydrogen charged DSS and aged for one month at RT is shown in month at RT.

Table 1
Chemical compositions of the studied stainless steels (DSS, AUSS, SMSS) in wt%.

Sample C S P Mn Si Ni Cr Mo N Cu

SAF 2205 0.025 0.004 0.022 0.411 0.288 6.745 24.012 4.63 0.3 0.072
316L 0.04 0.006 0.04 1.164 0.375 10.693 15.612 2.748 0.039 0.274
SMSS 0.063 0.002 0.021 1.859 0.364 7.002 10.513 3.315 0.009 0.467
 R. Silverstein, D. Eliezer / Journal of Alloys and Compounds 720 (2017) 451e459 453

Fig. 3. SEM observation of DSS samples (a) uncharged at lower magnication, (b) uncharged at higher magnication, (c) 72 h cathodic charged, showing the formation of needle
shape -martensite phase after one month at RT, (d) surface cracking in 72 h cathodic charged DSS- non-etched, and (e) etched-red arrows pointing at cracking sites. (For
interpretation of the references to colour in this gure legend, the reader is referred to the web version of this article.)

lattice parameter, Fig. 2. Regarding, a-phase reections, they
exhibited peaks' broadening, which indicates the formation of
smaller grain size. These observations were supported by previous
works on cathodic charged DSS alloys [12,18e20]. XRD results are
supported by the SEM micrographs, Fig. 3. The uncharged sample,
Fig. 3a and b, presents two phases with elongated grains and an
average grain size of ~10 mm. Supported by energy dispersive
diffraction (EDS) measurements, it was shown that the high weight
percent (wt%) areas of chromium and molybdenum, also known as
a-phase stabilizers [2], were ascribed to a phase regions, and high
wt% of nickel and nitrogen, also known as g-phase stabilizers [2],
were ascribed to g-phase regions.
Supported by the XRD pattern in Fig. 2, the 72 h hydrogen
charged DSS and aged for one month at RT (Fig. 3 c), showed a'-
martensite phase, pointed out with a red arrow. During aging of the
hydrogenated specimens, the -martensite became stable, while
the aphase reections became broadened. This change can be
0
explained by the next / g / a phase's transition, as described in
our previous work [7,12], as well as in other works [21,22]. This
means that part of the -martensite phase changes into a'-
martensite with a body centered tetragonal structure (bct). The
formation of a'-martensite is probably due to elastic stresses during
hydrogen desorption [22,23].
The surface cracking observation of 72 h charged DSS and aged
for one month at RT, Fig. 3d, shows severe cracking. Most cracking
seems to be intergranular inside g-phase and some of them seem to
be in the interface of a-g phases, Fig. 3e. These crackings are
probably due to the lower hydrogen diffusivity inside the austenitic
phase, Dgz3.3  1016 m2/s [16], thus causing the hydrogen con-
centration gradient beneath the surface along with higher
hydrogen concentration on the surface itself. This concentration
gradient implies large tensile stress on the surface layer, and has a
direct effect on the formation of -martensite, in addition to surface
cracking [24].
3.1.2. Austenitic stainless steel
In order to compare the microstructural changes of hydrogen on
austenitic stainless steel (AUSS), an XRD analysis was performed on
AUSS with the same charging condition as DSS (72 h cathodic
charging and aged for one month at RT, Fig. 4). These results were
compared to an uncharged sample, indicated by the black line in
Fig. 4. The uncharged sample indicates the appearance of small
traces of a phase, even though this phase is not seen in the SEM
micrograph (Fig. 5a). The SEM micrograph of uncharged austenitic
stainless steel (AUSS, Fig. 5a), shows a single austenitic fcc g-phase.
The AUSS consists of equiaxed grains with an average grain size of
~35 mm. The appearance of annealing twins present in each
austenitic grain are common and are usually found in annealed fcc
crystals [25].
The comparison between XRD spectra of DSS to AUSS, Figs. 2 and
4, respectively, indicates that g-phase experiences the same phe-
nomenon: peaks' broadening and a shift to smaller angles (can be
seen at 0 h aging, the green line in Fig. 4). This behavior indicates a
~2% increment in lattice parameter compared with the uncharged
sample. In addition, the formation of -martensite phase can also be
seen. This appearance is also seen in the SEM micrographs, Fig. 5b,
pointed out with red arrows, which can also indicate the formation
of a'-martensite due to elastic stresses during hydrogen desorption.
The -martensite phase's reections in DSS demonstrated higher
intensity (~70% higher than AUSS's peaks), meaning a greater sta-
bility of this phase in DSS. We estimated the -martensite quantities
using the Rietveld method [26,27], whose analysis is based on the
454 R. Silverstein, D. Eliezer / Journal of Alloys and Compounds 720 (2017) 451e459

Fig. 4. XRD pattern of the uncharged and cathodic hydrogen charged 316L for 72 h and aged for different time intervals.

Fig. 5. SEM observations of AUSS samples (a) uncharged, (b) 72 h cathodic charged, showing the formation of needle-shaped -martensite phase after one month at RT, and (c)
surface cracking in 72 h cathodic charged AUSS.

fact that intensity of different peaks in an XRD depends on the a steel containing a larger content of -martensite phase will
concentration of this phase in the material. The Rietveld method contribute to less severe damage. The description of g-phase sta-
revealed -martensite quantities of ~7 wt% and ~35 wt% in AUSS bility can be seen in Table 2. This table shows that the phase sta-
and DSS, respectively. This difference is probably related to the bility of g-phase in AUSS is the greatest, leading to a lower
alloying elements in DSS, as we have already shown in our previous hydrogen-induced phase transition (g / ) and, therefore, a
work [7]. higher severity of surface cracking.
Surface cracking observation of 72 h hydrogen charged AUSS
and aged for one month at RT, Fig. 5c, shows severe cracking, due to Table 2
the lower hydrogen diffusivity inside the g-phase. A comparison Calculated activation energies for hydrogenated for 96 h SAF 2205.
with DSS reveals much more severe intergranular cracking in AUSS.
Sample Phase concentration (wt%) S stability factor
The reason belongs to the content of g and phases in those steels:
A lower diffusivity for hydrogen will cause a larger hydrogen con- g a /a

centration gradient inside the steel. The -martensite phase pre- SMSS 30 70 8 17
sents ~7 orders higher diffusion constant, D,Hz5  1010 m2/sec SAF 2205 (DSS) 50 50 35 24
AUSS (316L) 95 5 7 32
[16], compared with g-phase, Dg z 3.3  1016 m2/s [16]. Therefore,
 R. Silverstein, D. Eliezer / Journal of Alloys and Compounds 720 (2017) 451e459
Fig. 6. XRD pattern of the uncharged and cathodic hydrogen charged SMSS for 72 h
and aged for different time intervals.
3.1.3. Super martensitic stainless steel (SMSS)
The XRD diffraction pattern, Fig. 6, shows a comparison between
uncharged SMSS and 72 h cathodic charged with hydrogen and
aged for different aging times at RT. It is clearly seen that imme-
diately after charging, indicated by green line in Fig. 5, the residual
g-phase peaks shift to smaller angles. As for DSS and AUSS, the
SMSS also showed additional reections which were formed by
hydrogen-induced g-phase transition. The XRD spectra show
-martensitic phase with hcp crystal structure. The Rietveld
method revealed -martensite quantities of ~8 wt% in SMSS. No
changes were seen in the original a'-martensitic peaks. Two very
important conclusions can be deduced. First, the g/ hydrogen-
induced phase transition will always appear in steels containing
hydrogen and g-phase, regardless of the g-phase content in a steel.
Second, the stability of g-phase plays an important role in the phase
transition; higher stability will lead to a lower amount of
-martensitic phase.
In line with the XRD results of uncharged SMSS, the micro-
structure of SMSS presented in the SEM micrograph, Fig. 7a, con-
sists mainly of a'- martensite with bct crystal structure and a small
quantity of retained g-austenite ~30 wt%, analyzed by the Rietveld
method [26,27]. The martensite morphology is lath with an average
martensitic grain size of ~10 mm. The surface cracking of 72 h
hydrogen charged SMSS, can be seen in Fig. 6b. The SEM micro-
graph indicates on that there were few micro-cracks on the surface
of the sample. Supported by the surface cracking of DSS and AUSS,
Figs. 3d and 5c, it may be concluded that the sample with the
Fig. 7. SEM observations of SMSS: (a) uncharged, and (b) surface cracking at 72 h cathodic charged.
455
highest -martensite phase content, or with the lowest hydrogen-
induced g-phase stability, will show the least damage.
In order to study the g-phase stability, we used the following
equation, supported by the works of Rozenak and Eliezer
[22,28,29], and others [30]. The g-phase stability of stainless steels
is given by the value of S:
S Ni 0:68Cr 0:55Mn 0:45Si 27C N ; (1.1)
where S is the austenite stability factor and Ni, Cr, Mn, Si, C and N are
the stabilizing elements in wt%.
A lower value of S is related to a greater transformation of g/,
and therefore, to a higher amount of /a'martensite phase. Ac-
cording to a 'Scheler diagram', Fig. 1, in order to stabilize a steel
with a specic phase composition, a certain amount of stabilizers
should be added to the steel, i.e. AUSS, DSS and SMSS will have
different alloying elements, Fig. 1.
The data in Table 2 presents phase quantities estimated by the
Rietveld method. It can be seen that AUSS shows the higher S value
(22% higher than DSS and 45% higher than SMSS), therefore the
phase transition to -martensite is lowere 7 wt% in AUSS compared
with 35 wt% at DSS. Generally, the transition from g to -martensite
can be explained by internal stresses which are created during
hydrogen absorption. According to our previously published work
[12], as well as the work of others [22,28,29,31e33], these stresses
can provide a signicant driving force for g / martensite
transition.
A very important conclusion from these results is that high g-
phase stability accelerates hydrogen embrittlement, which estab-
lishes the intensive amount of cracks on the steel's surface.
3.2. Thermal desorption spectrometry (TDS) of hydrogenated
stainless steels
3.2.1. Duplex stainless steel (DSS)
The hydrogen trapping mechanisms were examined by thermal
desorption spectrometry (TDS). An example of a 72 h charged DSS
sample after the examination of thermal desorption measurements
is presented in Fig. 8 a. All samples were heated from RT to 450  C,
at three non-isothermal heating rates: 2  C/sec, 4  C/sec and 6  C/
sec. In order to solve hydrogen trapping behavior and interaction,
we used Lee and Lee's model [15,34]. This model predicts the
hydrogen evolution rate from trap sites; hydrogen desorption rate
from trapping sites is given by:


dx Ea
A1  Xexp  ; (1.2)
dt RT
where X is the hydrogen content that escapes a trap, A is the
456 R. Silverstein, D. Eliezer / Journal of Alloys and Compounds 720 (2017) 451e459

Fig. 8. Experimental results at various ramp rates for 72 h cathodic hydrogen charged DSS samples: (a) hydrogen desorption rate as a function of temperature, and (b) deter-
mination of the activation energies for the three tted hydrogen desorption peaks.

reaction rate constant, Ea is the activation energy for releasing
hydrogen from its trapping site, R is the gas constant and T is the
absolute temperature.
By referring to the shift of the peak temperature with different
heating rates (Fig. 8a), the activation energy for hydrogen desorp-
tion from a certain trap can be calculated. By imposing a linear
change of temperature, 4, the activation energy of the rate deter-
mining desorption process (Ea) is obtained [15,34]:
phase. Therefore, the last irreversible trapping value 62.0 0.5 kJ/
mol is probably related to a'-martensite. The role of irreversible
trapping sites is very crucial in preventing the hydrogen embrit-
tlement phenomenon. This evidence was also seen in previous
work of us on DSS alloys [7,12,13]. Since trapping sites control the
availability of hydrogen to potential trapping sites, it is very
important that the activation energy with hydrogen will be as high
as possible [39].

 . 
dln 4 Tc2 Ea 3.2.2. Austenitic stainless steel (AUSS)
 ; (1.3) As for the DSS, the activation energies calculations of AUSS,
d1=Tc RT
Fig. 9 and Table 3, show two rst values, 19.0 0.5 kJ/mol and
where, Tc is the peak temperature and 4 is the heating rate. 32.0 2.0 kJ/mol, which are lower than 60 kJ/mol and are classied
Thus by plotting ln(4/Tc2) versus 1/Tc,Fig. 8b, where Tc is the as reversible traps. The higher value in AUSS's trapping sites,
critical temperature for hydrogen evaluation from a trap site, this 43.0 0.5 kJ/mol, is also ascribed to reversible trapping sites. The
should yield straight lines with the slope -Ea/R (R is the gas con- reason for the change, compared with DSS, in the last value is
stant) [15]. The slope of ln4 versus 1/T was tted to experimental probably related to g-phase's stability or hydrogen-induced phase
data with a 95% condence limit on the slope and multiplied by R to transition. Lower gphase stability, such as in the case of DSS, will
obtain Ea values. result in higher hydrogen absorption content and higher trap
The calculated activation energies from tted linear curves, activation energy. In the case of AUSS, the content of hydrogen-
Fig. 8b and Table 3, reveal the rst two values to be lower than
60 kJ/mol and one value which is highere 62 0.5 kJ/mol. The rst
two values can be classied as reversible traps, due to their lower
value [9,35,36]. The last value is classied as an irreversible trap-
ping site. According to different works on steels [11,16,37], revers-
ible trapping sites can be attributed to either an elastic stress eld
of dislocation with a range of 0e20 kJ/mol, a screw's dislocation
core or grain boundaries with a range of 20e30 kJ/mol and high
angle boundaries, or austenitic-ferritic interface with a range of
40e55 kJ/mol. In accordance with some of our previously published
work [7,8,12,13,38], a trapping value close to 60 kJ/mol can be
ascribed to second phases, such as -martensite or a'-martensite

Table 3
Calculated activation energies for cathodic charged for 72 h samples: DSS, AUSS, and
SMSS.

Sample Ea [kJ/mol] Ea [kJ/mol] Ea [kJ/mol]

DSS 20.0 0.5 34.0 0.5 62.0 0.5

AUSS 19.0 0.5 32.0 2.0 43.0 0.5
Fig. 9. Activation energies for the three tted hydrogen desorption peaks of 72 h
SMSS 21.0 0.5 32.0 0.5 58.0 0.5
cathodic hydrogen charged AUSS.
 R. Silverstein, D. Eliezer / Journal of Alloys and Compounds 720 (2017) 451e459
Fig. 10. Activation energies for the three tted hydrogen desorption peaks of 72 h
cathodic hydrogen charged SMSS.
Fig. 11. (a) TDS spectra at a 6  C/min heating rate of 72 h cathodic hydrogen charged
samples: SMSS, DSS, and AUSS. (b) Integration of TDS spectra (a) resulting in the
hydrogen desorbed content for each sample.
Fig. 12. Inuence of heating rate on hydrogen evaluation content vs. temperature in cathodic hydrogen charged for 72 h samples: (a) DSS, (b) AUSS, and (c) SMSS.
457
induced -martensite phase is 80% lower compared with DSS. This
difference explains the change in the last value of AUSS's trapping
energy. It can be said that phase transformation and steel's stabi-
lizers can signicantly affect steel trapping states. This data is also
supported by our previously published work [12].
The activation trapping states completely support the hydrogen
embrittlement mechanism, Figs. 3d, 5c, and 7b. It can be seen from
Table 3, that the steel with the lower trapping states levels, such as
AUSS, will mostly promote cracking. This phenomenon can be
explained as follows: At lower trapping states levels, hydrogen will
be easily available to potential cracking sites, establishing the
fracture and pathway.
3.2.3. Super martensitic stainless steel (SMSS)
The activation measurements of 72 h cathodic hydrogen
charged SMSS is presented in Fig. 10 and Table 3. These calculations
predicted values of reversible trapping sites, and one trapping site
with a close value to 60 kJ/mol. These values were 21.0 0.5 kJ/mol,
32.0 0.5 kJ/mol and 58.0 0.5 kJ/mol. It can be seen that the great
differences between AUSS and SMSS are in the last trapping site's
value. The SMSS presents a 25% larger value compared with AUSS.
These differences can be explained with s stability factor, Table 2,
which indicates ~47% higher values of s-stability factor in AUSS
compared with SMSS. Changes can also be seen between DSS and
SMSS, both samples presented high activation value ascribed to
martensite phase, and small differences (~6%) were seen between
trapping states. These differences can be explained by difference in
the amount of -martensite phase, and its distribution. This phase
probably had a higher effect on trapping state than a'- martensite.
Even though, a'-martensite content was signicantly higher in
SMSS ~70 wt%, its -martensite content was signicantly lower 8 wt
% compared with 35 wt% DSS. Supported by the above indication
SMSS with the lowest stability of g-phase and among the highest
trapping states' levels will show less severe cracking.
The results in Table 3 are well manifested by the changes applied
in the TDS spectra at 6  C/min of cathodic charged for 72 h SMSS,
DSS and AUSS samples.
It can be seen that each peak position, Fig. 11, is in line with its
suitable activation energy, Table 3; higher temperature's peak
levels indicate on higher activation energy. Moreover, the
-martensite phase trapping site appears only in SMSS and DSS
where the g-phase stability is lower and the amount of martensite
is higher. An additional interesting observation is the spectra in-
tensity, which indicates desorbed hydrogen content, Fig. 11b. It can
458 R. Silverstein, D. Eliezer / Journal of Alloys and Compounds 720 (2017) 451e459
be clearly seen that -martensite phase signicantly affects the
amount of hydrogen absorption/desorption, due to its higher
hydrogen diffusion constant. SMSS presents the highest desorbed
hydrogen content, ~50% higher than DSS and ~90% higher than
AUSS, due to higher martensite content.
In order to better understand the desorption process and trap-
ping mechanisms, the inuence of heating rate on hydrogen eval-
uation content in DSS, AUSS, and SMSS are presented in Fig. 12.
Signicant changes can be seen between DSS and AUSS steels.
Fig. 12a and b, respectively. These changes are related to the
appearance of -martensite as a trapping site at DSS. The DSS
presents a certain jump at ~200  C, whereas the AUSS does not
show any jump. The reason is probably due to the -martensite
trapping site. Since AUSS does not show any sign for an
-martensite trapping site, the jump, which is related to hydrogen
desorption from the -martensite trap site, does not appear.
Changes related to heating rates are the same in all steels. The
2  C/min heating rate presents the lowest desorption rate while the
6  C/min presents the highest desorption rate, Fig. 12a and b. These
results are in line with the TDS spectra of DSS (Fig. 8a).
Regarding SMSS, Fig. 12c, the big difference between DSS and
SMSS is related to the amount of martensite content compared with
this of g (70 wt % in SMSS and 35 wt % in DSS). It is known from
previously published works that hydrogen's solubility in a-phase is
signicantly higher than in g-phase [16,17]. Therefore, SMSS will
show greater hydrogen content evaluation. In addition, it can be
seen that SMSS and AUSS show lower desorption temperatures
compared with DSS; ~220  C for SMSS, ~160  C for AUSS and
~300  C for DSS. These differences were ascribed to the activation
energy values which were the highest in DSS. Therefore, the
following order of maximum temperature desorption can be
written as follows: DSS>SMSS>AUSS. This order is also valid for the
activation energy values.
4. Summary and conclusions
Hydrogen trapping mechanisms of SAF 2205 duplex stainless
steel, 316L austenitic stainless steel and super martensitic stainless
steel were investigated using thermal desorption spectrometry,
and microstructural observations.
In all of the above mentioned steels, the dominant phase
0 SAF2507 duplex stainless steels, in: Hydrogen Transport and Cracking in
transformation was the hydrogen induced g / / a or g / .
These indications were supported by XRD and SEM results. We
concluded that the next transformation g / will always appear in
steels containing hydrogen and g-phase, regardless of the g-phase
content in the steel.
It was shown using Lee and Lee's model that trapping sites were
attributed to stress dislocation elds at 0e20 kJ/mol and to higher
activation energies at 30e40 kJ/mol to grain boundaries. The DSS
and SMSS presented additional trapping site close to 60 kJ/mol at
~58 kJ/mol for DSS and ~62 kJ/mol for SMSS, which was related to
the massive formation of ~35 %wt and ~70 %wt -martensite phase
contents in DSS and SMSS, respectively. In this paper, we have
shown that the stability degree of g-phase signicantly affects the
trapping mechanisms and, therefore, affects the hydrogen embrit-
tlement mechanism. The hydrogen charged AUSS showed the
greatest stability factor (S) among the rest of the steels: ~22% higher
than DSS and ~45% higher than SMSS. This determination affects
the amount of formed -martensite phase, which was 80% lower in
AUSS compared with DSS and 12% lower compared with SMSS. This
determination is in line with hydrogen trapping behavior. It was
concluded that the activation energy levels were the highest in DSS
and the lowest in AUSS.
Hydrogen embrittlement was proved to be signicantly affected
by the g-phase stability; higher stability of g-phase will produce
severe damage and promote the fracture mechanism, due to a
lower diffusion constant in that phase. Regarding these insights, it
can be said that susceptibility to the hydrogen fracture mechanism
will be the highest in AUSS and lowest in SMSS, where g-phase
stability is lower and trapping states are higher. Moreover, it can be
concluded that the main mechanism affecting the hydrogen frac-
ture mechanism is g-phase stability rather than trapping states.
References
[1] J.R. Davis, Stainless Steels, third ed., Ohaio: ASM international, 1999.
[2] P. Lacombe, B. Baroux, G. Beranger, Stainless steels, Phys. Les Ulis (1993)
16e55.
[3] I.M. Robertson, P. Sofronis, A. Nagao, M.L. Martin, S. Wang, D.W. Gross,
K.E. Nygren, Hydrogen embrittlement understood, Metall. Mater. Trans. B 46
(2015) 1085e1103.
[4] R.K. Dayal, N. Parvathavarthini, Hydrogen embrittlement in power plant
steels, in: R. Baldev, R. Bhanu Sankara (Eds.), Sadhana Academy Proceedings in
Engineering Sciences, vol. 28, Universities press, 2003, pp. 431e451.
[5] R.A. Oriani, A decohesion theory for hydrogen-induced crack propagation, in:
Stress Corrosion Cracking and Hydrogen Embrittlement of Iron Base Alloys,
1977, pp. 351e358.
[6] V. Olden, C. Thaulow, R. Johnsen, Modelling of hydrogen diffusion and
hydrogen induced cracking in supermartensitic and duplex stainless steels,
Mater. Des. 29 (2008) 1934e1948.
[7] R. Silverstein, D. Eliezer, B. Glam, S. Eliezer, D. Moreno, Evaluation of hydrogen
trapping mechanisms during performance of different hydrogen fugacity in a
lean duplex stainless steel, J. Alloys Compd. 648 (2015) 601e608.
[8] R. Silverstein, D. Eliezer, Hydrogen trapping energy levels and hydrogen
diffusion at high and low strain rates (~105s1 and 107s1) in lean duplex
stainless steel, Mater. Sci. Eng. A 674 (2016) 419e427.
[9] G.M. Presouyre, Trap theory of hydrogen embrittlement, Acta Metall. 28
(1980) 895e911.
[10] A. Mcnabb, P.K. Foster, A new analysis of the diffusion of hydrogen in iron and
ferritic steels, Trans. Metall. Soc. AIME 227 (1963) 618e627.
[11] I. Maroef, D.L. Olson, M. Eberhart, G.R. Edwards, Hydrogen trapping in ferritic
steel weld metal, Int. Mater. Rev. 47 (2002) 191e223.
[12] R. Silverstein, D. Eliezer, Hydrogen trapping mechanism of different duplex
stainless steels alloys, J. Alloys Compd. 644 (2015) 280e286.
[13] R. Silverstein, O. Sobol, T. Boellinghaus, W. Unger, D. Eliezer, Hydrogen
behavior in SAF 2205 duplex stainless steel, J. Alloys Compd. 659 (2016)
2689e2695.
[14] Y. Yagodzinskyy, O. Todoshchenko, S. Papula, H. Ha, Hydrogen solubility and
diffusion in austenitic stainless steels studied with thermal desorption spec-
troscopy, Steel Res. Int. 82 (1) (2011) 20e25.
[15] S. Lee, J. Lee, The trapping and transport phenomena of hydrogen in nickel,
Metall. Trans. A 17 (1986) 181e187.
[16] A. Turnbull, R.B. Hutchings, Analysis of hydrogen atom transport in a two-
phase alloy, Mater. Sci. Eng. A 177 (1994) 161e171.
[17] A. Turnbull, E. Beylegaard, R. Hutchings, Hydrogen transport in SAF2205 and
Metals, 1994, pp. 268e279.
[18] R. Silverstein, D. Eliezer, B. Glam, D. Moreno, S. Eliezer, Inuence of hydrogen
on microstructure and dynamic strength of lean duplex stainless steel,
J. Mater. Sci. 49 (2014) 4025e4031.
[19] R. Bar, E. Dabah, D. Eliezer, T. Kannengiesser, T. Boellinghaus, The inuence of
hydrogen on thermal desorption processes in structural materials, Procedia
Eng. 10 (2011) 3668e3676.
[20] R. Silverstein, D. Eliezer, B. Glam, D. Moreno, Dynamic strength of duplex steel
in the presence of hydrogen, in: SteelyHydrogen2014 Conference Pro-
ceedings, 2014, pp. 662e666.
[21] S. Teus, V. Shyvanyuk, V. Gavriljuk, Hydrogen-induced g/3 transformation
and the role of 3-martensite in hydrogen embrittlement of austenitic steels,
Mater. Sci. Eng. A 497 (1e2) (Dec. 2008) 290e294.
[22] P. Rozenak, L. Zevin, D. Eliezer, Hydrogen effects on phase transformations in
austenitic stainless steels, Mater. Sci. 19 (1984) 567e573.
[23] N. Narita, C.J. Altstetter, H.K. Birnbaum, Hydrogen-related phase trans-
formations in austenitic stainless steels, Metall. Trans. A 13A (1982)
1355e1365.
[24] P. Rozenak, D. Eliezer, Quantitative X-ray phase analysis of sensitized type 316
stainless steel after cathodic hydrogen charging, Mater. Sci. Eng. 67 (1984)
1e4.
[25] S. Mahajan, C.S. Pande, M.A. Imam, B.B. Rath, Formation of annealing twins in
fcc crystals, Acta Matter 45 (1997) 2633e2638.
[26] H.M. Rietveld, A prole renement method for nuclear and magnetic struc-
tures, J. Appl. Crystallogr. 2 (1969) 65e71.
[27] L.B. McCusker, R.B. Von Dreele, D.E. Cox, D. Loue r, P. Scardi, Rietveld rene-
ment guidelines, J. Appl. Crystallogr. 32 (1999) 36e50.
[28] P. Rozenak, D. Eliezer, Phase changes related to hydrogen- induced cracking in
austenitic stainless steel, Acta Met. 35 (1987) 2329e2340.
[29] P. Rozenak, D. Eliezer, Nature of the g and g*phases in austenitic stainless
steels cathodically charged with hydrogen, Metall. Trans. A 19 (1988)
 R. Silverstein, D. Eliezer / Journal of Alloys and Compounds 720 (2017) 451e459
2860e2862.
[30] S. Floreen, J.R. Mihalisim, High strength stainless steel by deformation in low
temperature, in: Advances in the Technology of Stainless Steels and Related
Alloys, American Society for Testing Materials, 1965, pp. 17e25.
[31] P. Rozenak, D. Eliezer, Effects of ageing after cathodic charging in austenitic
stainless steels, J. Mater. Sci. 19 (1984) 3873e3879.
[32] D. Eliezer, D.G. Chakrapani, C.J. Altstetter, E.N. Pugh, The inuence of austenite
stability on the hydrogen embrittlement and stress- corrosion cracking of
stainless steel, Metall. Trans. A 10 (1979) 935e941.
[33] E. Minkovitz, M. Talianker, D. Eliezer, TEM investigation of hydrogen induced
-hcp-martensite in 316L-type stainless steel, J. Mater. Sci. 16 (1981)
3506e3508.
[34] A. Turnbull, R.B. Hutchings, D.H. Ferriss, Modelling of thermal desorption of
hydrogen from metals, Mater. Sci. Eng. 238 (January) (1997) 317e328.
459
[35] G.M. Pressouyre, I.M. Bernstein, A quantitative analysis of hydrogen trapping,
Metall. Trans. A 9A (November) (1978) 1571e1580.
[36] E. Barel, G. Ben Hamu, D. Eliezer, L. Wagner, The effect of heat treatment and
HCF performance on hydrogen trapping mechanism in timetal LCB alloy,
J. Alloy. Compd. 468 (2009) 77e86.
[37] E. Abramov, D. Eliezer, Trapping of hydrogen in helium-implanted metals,
J. Mater. Sci. Lett. 7 (2) (Feb. 1988) 108e110.
[38] R. Silverstein, D. Eliezer, Effects of residual stresses on hydrogen trapping in
duplex stainless steels, Mater. Sci. Eng. A 684 (2017) 64e70.
[39] P. Sofronis, M. Dadfarnia, P. Novak, R. Yuan, B. Somerday, I.M. Robertson,
R.O. Ritchie, T. Kanezaki, Y. Murakami, A combined applied mechanics/ma-
terials science approach toward quantifying the role of hydrogen on material
degradation, in: Proceedings of the 12th International Conference on Fracture
(ICF-12), 2009, pp. 1e10.