Journal of Materials Science & Technology  2019 , 35 (6): 1081-1087 https://doi.org/10.1016/j.jmst.2018.12.019

Hydrogen-assisted fracture features of a high strength ferrite-pearlite steel

Yuefeng Jiangab, Bo Zhanga*, Dongying Wangc, Yu Zhoua, Jianqiu Wanga, En-Hou Hana, Wei Keaa

a CAS Key Laboratory of Nuclear Materials and Safety Assessment, Institute of Metal Research, Chinese Academy of Sciences, Shenyang 110016, China
b School of Materials Science and Engineering, University of Science and Technology of China, Shenyang 110016, China
c Shenyang Blower Works Group Corporation, Shenyang 110869, China

Corresponding authors:   * Corresponding author at: 72 Wenhua Rd, Shenyang, 110016, China.E-mail address: bxz011@imr.ac.cn (B. Zhang).

Received: 2018-08-30

Revised:  2018-11-14

Accepted:  2018-11-30

Online:  2019-06-20

Copyright:  2019 Editorial board of Journal of Materials Science & Technology Copyright reserved, Editorial board of Journal of Materials Science & Technology

More

Abstract

Up to now, the exact reason of hydrogen-induced fracture for ferrite-pearlite (FP) steel is still not fully understood. This study presents detail observations of the feature beneath the fracture surface with the aim to reveal the hydrogen-induced cracking initiation and propagation processes. Slow strain rate tensile (SSRT) testing shows that the FP steel is sensitive to hydrogen embrittlement (HE). Focused ion beam (FIB) was used to prepare samples for TEM observations after HE fracture. The corresponding fractographic morphologies of hydrogen charged specimen exhibit intergranular (IG) and quasi-cleavage (QC) fracture feature. Pearlite colony, ferrite/pearlite (F/P) boundary and the adjacent ferrite matrix are found to be responsible for the initial HE fracture and the subsequent propagation. With increasing of the stress intensity factor, fracture mode is found to change from mixed IG and QC to entire QC feature which only occurs at the ferrite matrix. No crack is observed at the ferrite/cementite (F/C) interface. This may be mainly due to the limited pearlite lamella size and relatively low interface energy.

Keywords: Ferrite-pearlite steel ; TEM ; Fracture ; Hydrogen embrittlement

0

PDF (2380KB) Metadata Metrics Related articles

Cite this article Export EndNote Ris Bibtex

Yuefeng Jiang, Bo Zhang, Dongying Wang, Yu Zhou, Jianqiu Wang, En-Hou Han, Wei Kea. Hydrogen-assisted fracture features of a high strength ferrite-pearlite steel[J]. Journal of Materials Science & Technology, 2019, 35(6): 1081-1087 https://doi.org/10.1016/j.jmst.2018.12.019

1. Introduction

Ferrite-pearlite (FP) steels are widely used in petroleum pipeline, fasteners, bolts for the construction of bridges and many other industrial applications due to high strength, simple heat-treatment process and sufficient toughness [1], [2], [3], [4], [5]. Nevertheless, in practical applications, FP steels are very vulnerable to hydrogen embrittlement (HE) which causes a severe loss in ductility. It also leads to catastrophic and unpredictable failure of structural components in the hydrogenous environment. Hydrogen in the steels can exist in dissociated form and concentrate at grain boundaries, vacancies and interstitial sites. The mechanism of HE supposes that high concentrations of dissolved hydrogen decrease the cohesive force between the atomic bonds of the lattice, at grain boundaries and interfaces (hydrogen-enhanced decohesion (HEDE)), also resulting in localized plastic deformation or cleavage, leading to the coalescence of microcracks and finally to fracture (hydrogen-enhanced localized plasticity (HELP)) [6], [7], [8], [9]. These mechanisms suggest that the initiation and propagation of hydrogen-assisted cracking depend on the microstructure of the steel and the distribution of hydrogen.

It is generally recognized that microstructures play important roles in the initiation process of HE. Shibanuma et al. [10] reported that cracks initiated at the pearlite colony due to the rapid transportation of hydrogen. Park et al. [11] observed that HE cracks initiated at the grain boundary triple junction due to the reduced threshold stress for crack nucleation by hydrogen. Kim et al. [12] found crack nucleation at the ferrite/cementite (F/C) interface in pearlite as hydrogen can reduce the strength of interface bonding. Martin et al. [13] studied the HE of FP steel and showed that HE cracks nucleated at the slip band intersections in the ferrite matrix due to severe plastic deformation accelerated by hydrogen. According to these studies, the initiation of HE can occur at pearlite colony, ferrite boundary, F/C interface and ferrite matrix.

After crack nucleation, the crack propagation is promoted by the reversible hydrogen. Nemoto et al. [14] found the crack propagation across the ferrite grains. Similar results are reported by Robertson et al. [15], they found that ferrite matrix is the HE propagation path due to the severe plastic deformation of ferrite and a large number of hydrogen trapped in dislocations. However, Briottet et al. [16] and McEniry et al. [17] found that the F/C interfaces acted as trapping sites which were the HE propagation path. In this case, the fracture feature is quasi-cleavage (QC). They explained that hydrogen enrichment at the F/C interfaces can easily lead to stress concentration due to the presence of high strength cementite. With the change of stress state, the interaction of hydrogen and microstructure can be different and become more complex. On this basis, the hydrogen-induced fracture need a further study. Recently, focused ion beam (FIB) technique has been successfully used to make sample normal to the fracture surface from the site-specific location for a transmission electron microscopy (TEM) observation [18,19]. FIB-TEM observation can provide the evolution of dislocations during the HE fracture, which can help to reveal the path of crack propagation and consequent microstructure evolution.

Up to now, detailed views of the fracture feature and observations of hydrogen distribution in nanoscale are still challenging tasks. In this study, HE resistance of a FP steel was evaluated through a slow strain rate test (SSRT) in the air and under the electrochemical hydrogen charging condition. The fracture feature was observed by scanning electron microscopy (SEM) and TEM. Hydrogen binding energies were measured by thermal desorption spectrum (TDS). FIB-TEM samples were extracted from QC propagation region with the aim to obtain more details of the crack propagation feature and microstructure beneath the fracture surface. This is helpful for a better understanding of the HE mechanism.

2. Experimental

2.1. Materials

Commercial AISI 4140 ferrite-pearlite steel was used in this work. The chemical composition of the steel is given as follows (in wt%): 0.38 C, 0.24 Si, 0.68 Mn, 0.003 S, 0.012 P, 1.07 Cr, 0.02 Ni, 0.21 Mo and balance Fe. The steel rods with 35 mm nominal diameter and 200 mm length were austenitized at 850 °C for 180 min followed by air cooling.

2.2. Microscopic observations

A XL30-FEG field emission SEM operating at an accelerating voltage of 15 kV was used for the microstructural observations. Prior to SEM observation, the surface of the specimens was etched with 3% nital after polishing. In order to identify different type of boundaries in FP steels (ferrite boundary and ferrite/pearlite (F/P) boundary etc.), the microstructure observation was also performed using a JEOL 2100 F field-emission TEM with an energy-dispersive X-ray (EDX) unit operating at 200 kV. XRD phase analysis was carried out using CuKα radiation at 50 kV.

2.3. Thermal desorption spectrum analysis

The TDS device consists of a gas chromatograph, a thermal conductivity detector with a heating unit and a recorder. Standard-mixture gas (Ar + 50 ppm H2) within the gas chromatograph was used for the hydrogen content calibration in the specimen. TDS was conducted in the temperature range from 25 to 800 °C at different heating rates of 100, 200 and 300 °C/h. The sampling interval of the carrier gas to travel from the point of injection to gas chromatograph was 5 min to obtain desorption profiles. The occluded hydrogen content in the specimens (5 mm in diameter) was measured immediately. Hydrogen in steel is prone to be trapped at various structural defects such as dislocations and grain boundaries. To identify the hydrogen trapping sites, the activation energy for hydrogen desorption was calculated based on the work of Lee et al. [20,21] by an equation of the original formula of Kissinger [22],

$\frac{\partial ln(\phi / T_{p}^{2})}{\partial (1/T_{p})}$=- EaR(1)

where ∅ is the heating rate in K/s, Tp is the peak temperature in K, Ea is the activation energy for hydrogen desorption in kJ/mol, and R is the gas constant. After performing TDS analysis, the corresponding temperature peaks in the hydrogen desorption curves were measured. The slope of ($\phi / T_{p}^{2}$)versus (1/Tp) plot allows to obtain the value of Ea that corresponds to a specific hydrogen trapping site.

2.4. SSRT under electrochemical hydrogen charging

SSRT was carried out at a strain rate of 1 × 10-5 s-1 during which the tested specimen is electrochemically charged with hydrogen in 1 M NaOH aqueous solution containing 1 g L-1 thiourea at a cathodic current density of 10 mA/cm2. The dimensions of rod-like specimens were based on ASTM Standard E8 recommendations. HE susceptibility of steels (Iδ) was evaluated by the degree of elongation loss in the hydrogen environment and calculated by the following equation:

Iδ=(1-$\frac{δ_{H}}{δ_{Air}}$)×100%, (2)

where δH and δAir are the elongations of the specimens in H and air environments, respectively.

2.5. Fracture feature observations

The fracture surfaces of the hydrogen-charged specimens were observed using a XL30-FEG field emission and a JEOL 6060 L V SEM equipped with a FEI Dual Beam 235 FIB instrument, respectively. These new modalities include the use of a dual-beam system and FIB milling described in Ref. [23]. The sub-cracks were observed with the aim to identify the crack initiation sites. In order to explore the path of crack propagation, FIB was used to extract specimens normal to the fracture surface from site-specific locations. It also enables a microscale observation of the microstructure and the crack propagation feature beneath the fracture surface.

3. Results

3.1. Microstructure of the FP steel

The SEM images of the microstructure for the FP steel are presented in Fig. 1. Ferrite and pearlite colony can be clearly seen in Fig. 1(a) as marked with F (dark) and P (bright), respectively. Pearlite colony with lamellar (layered or plate-like) structure composed of alternating layers of F/C interfaces are observed by SEM as shown in Fig. 1(b). In order to have detailed views of the pearlite colony structure, TEM images were obtained for a further study. A typical TEM micrograph is shown in Fig. 1(c), where a F/P boundary and several intersections between F/C interface and F/P boundary can be clearly seen. The interlamellar spacing of two adjacent F/C interfaces is about 200 nm. Fig. 1(d) shows two pearlite colonies with different orientations. The orientation of pearlite colony is usually related to the original austenite. XRD results show that there is no obvious retained austenite presented in the steel since only diffraction peaks for the bcc Fe are observed as shown in Fig. 2.

Fig. 1.   Typical microstructure micrographs of FP steel showing ferrite and pearlite structure: (a) SEM micrograph of ferrite (dark) and pearlite (bright), (b) SEM micrograph of pearlite colony with lamellar F/C interfaces, (c) TEM micrograph of F/P boundary with several intersections between F/C interface and F/P boundary and (d) TEM micrograph of two pearlite colonies with different orientations.

Fig. 2.   XRD spectrum of the FP steel after heat treatment.

3.2. Activation energy for hydrogen desorption

TDS curves were measured at three heating rates. The TDS curve at the heating rate of 100 °C/h is shown in Fig. 3(a). Two peak desorption rates are identified which locate at 140 °C and 350 °C, respectively. The temperature associated with the peak desorption rate increases with increasing the heating rate. The activation energies for hydrogen desorption can be determined according to the variation of the peak temperature with the heating rate. The plots of $ln(\emptyset/T_{p}^{2})$ versus 1/Tp for the two hydrogen desorption peaks are shown in Fig. 3(b). The calculated activation energies of hydrogen desorption are 16.7 and 67.8 kJ/mol for the low and high temperature peaks, respectively. For low-temperature peak, Watanuki et al. [24,25] reported that the low activation energy for desorption is associated with H trapping in the reversible trapping sites, such as lattice, grain boundaries and dislocations. For the FP steel, F/P boundary, pearlite colony and dislocations in the ferrite matrix can be the reversible H trapping sites, which have the low H desorption activation energy. The high-temperature peak usually corresponds to the irreversible hydrogen trapping sites. Up to now, the nature of this peak is not clearly understood. It is known that any interstitial elements can increase the concentration of vacancies, which can be designated as superabundant ones in the tentative thermodynamical equilibrium with the dissolved interstitial elements [26]. Experimental observations and ab initio theoretical calculations have demonstrated that hydrogen can enhance vacancy formation [27,[26], [27], [28], [29]. The observed high desorption activation energy is probably due to hydrogen-vacancy interaction and the formation of stable hydrogen-monovacancy complexes [28]. Therefore, the hydrogen-induced vacancies can be the irreversible hydrogen trapping sites in this steel.

Fig. 3.   Measurement of the hydrogen desorption activation energy: (a) a typical hydrogen desorption curve of the FP steel with heating rate of 100 °C/h, and (b) the plots of $ln(\emptyset/T_{p}^{2})$ versus 1/Tp for the two hydrogen desorption peaks.

3.3. HE susceptibility and fracture feature

Typical SSRT curves of AISI 4140 tested in air and under hydrogen charging condition are presented in Fig. 4. The specimen tested in the air has δair of 10.7%, while the hydrogen-charged specimen (with δH of 4.6%) fractures much earlier than the uncharged specimen. Based on the SSRT results, the calculated HE susceptibility of the hydrogen-charged steel is 57%, indicating that the FP steel is sensibility to HE. Fig. 5(a-d) shows the fracture surfaces of the uncharged and charged specimens. As can be seen in Fig. 5(a), the center of fracture surface of the uncharged sample exhibits a ductile fracture involving numerous dimples by homogeneous microvoid coalescence. The hydrogen charged specimen shows a mixed fracture mode with IG and QC feature at the edge of the fracture surface (Fig. 5(b)). This result indicates that the hydrogen-induced crack initiates at the edge of the fracture surface and then simultaneously propagates through the grain boundary and inner the grain. In order to identify the initiation sites of the cracks, we also examined the subcracks. It clearly shows that the hydrogen-induced crack nucleates (a microcrack) at pearlite colony boundary, F/P boundary and the adjacent ferrite matrix (Fig. 5(c)). As the stress increases, more microcracks nucleate and then coalesce to cracks mixed with IG and QC feature (indicated with arrows). When the crack reaches near the center of the specimen, the macroscopic fracture feature changes to the QC fracture mode with some “featureless” flat regions, as shown in Fig. 5(d). In this case, IG mode was not observed.

Fig. 4.   Typical SSRT curves for FP steels at a constant strain rate of 1 × 10-5s-1 at 298 K in the air and in hydrogen charging environment at current density of 10 mA/cm2.

Fig. 5.   SEM images of the cracking FP steel after SSRT: (a) fracture surface center of the uncharged sample with numerous dimples, (b) the edge of fracture surface of the hydrogen charged samples showing a mixed fracture mode with IG and QC feature, (c) subcracks observation showing hydrogen induced crack nucleation (a microcrack) at pearlite colony boundary, F/P boundary and the adjacent ferrite matrix with IG and QC feature, and (d) QC fracture mode with some “featureless” flat regions in the center of fracture surface.

3.4. FIB-TEM observations of the QC fracture path

In order to have a detailed view of the QC feature, the TEM foils of the hydrogen-charged specimen were extracted beneath the fracture surfaces by FIB lift-out technique. The specific sites for TEM foils with three tear ridges are shown in Fig. 6(a). One tear ridge is presented in the bright-field TEM micrograph (Fig. 6(b)) where Pt layer (deposited to preserve the fracture surface) can be clearly seen. There is a high density of dislocations in the tear ridge which indicates that severely localized plastic deformation occurs before the final fracture. In order to identify the phases around the crack propagation region, SAED analysis and elemental mapping (TEM/EDX) of the cracking zone were performed. SAED analysis (Fig. 6(b)) reveals that the propagation zones is the ferrite matrix with a body-centered cubic structure. Elemental mapping shows that there is rarely C atoms uniformly distributed in the zone (Fig. 6(c)). Thus, the tear ridge is caused by the fracture of the ferrite matrix.

Fig. 6.   TEM observations of the microstructure beneath the QC fracture surface. (a) the specific site for TEM foils with three tear ridges, (b) bright-field image and corresponding SAED pattern of one tear ridge with Pt layer, and (c) EDX elemental mapping of C.

4. Discussion

In this study, it has been shown that for the specimen tested in air, the fracture surface exhibits ductile fracture (Fig. 5(a)). While for the hydrogen-charged specimen, the fracture surface shows a mixture of IG and QC feature (Fig. 5(b)). The cracks were found to initiate at pearlite colony boundary, F/P boundary and the adjacent ferrite matrix (Fig. 5(c)). With the increase of crack length, these cracks propagate along pearlite colony boundary, F/P boundary and through the adjacent ferrite matrix, finally all cracks pass through the ferrite matrix. Based on SEM, TEM observations and TDS analysis, the cracking process of FP steel in H environment can be understood as follows.

At the early stage of crack initiation, cracks tend to nucleate at some weak sites, such as pearlite colony, F/P boundary, ferrite matrix, F/C interface (in pearlite) and segregated impurities, etc [30,31]. As for pearlite colonies, there are many random boundaries with misorientation angles higher than 15° [32,33]. The atoms on the high angle boundary are severely displaced from their equilibrium positions, thus giving rise to high grain boundary energy [34]. In addition, the pearlite colonies are the favorite trapping site for the diffusible hydrogen as measured by TDS. As a result, HE cracks can easily initiate at the pearlite colony.

Apart from pearlite colony, the F/P boundary and the adjacent ferrite matrix are also the crack initiation sites as shown in Fig. 5(c). These can be understood based on void formations at F/P boundary and in the ferrite due to the synergistic interplay of hydrogen and dislocation (HELP mechanism). During the tensile test, dislocations can form and slip in the relatively soft ferrite. It is well known that hydrogen can reduce the repulsive interactions of dislocations and enhance local plasticity [19]. The enhanced motion of dislocations makes more dislocations impinging on the F/P boundary and the intersections (Fig. 1(c)), which can lead to the stress concentration. As a result, void formations at intersections of the F/P boundaries can occur, which gives rise to crack initiation on the F/P boundary [35]. After the microcrack nucleation at F/P boundary, there is a plastic zone near the crack tip. In this plastic zone, a high density of dislocations can form in the ferrite matrix due to the localized plasticity. Localized plasticity and hydrogen concentration can cause stress concentration at the slip band intersections. As a result, the slip band intersections can also be the favorable sites for crack initiation.

It has been reported that QC fracture occurs along the F/C interface in the pearlite colony [12,16,17]. In this study, no crack was observed to initiate at the F/C interface. The probable reason for this is that the F/C interface is a low energy interface [36]. Since, the (101)c//{112}f and (001)c//{215}f habit planes enable a low energy interface to form due to the good rigid atomic and lattice interfacial matching [37]. Based on the Griffith-Orowan fracture criterion [35], the effective decohesion work γγeff is the sum of the energy to break the interface γc and the plastic work γp accompanying the decohesion event as

γeff=2γcp (3)

A high γeff indicates a high resistance against cracking. Compared with pearlite colony and F/P boundary, the F/C interface has the highest effective decohesion work γeff since it has the lowest interface energy (i.e., highest energy γc to break the surface). Therefore, it has the highest resistance against fracture. Meanwhile, the interlamellar spacing of two adjacent F/C interfaces is about 200 nm in the pearlite. Consideration of the relationship between the resolved shear stress required for expansion of a dislocation loop and the dislocation bowing follows the Orowan relation [38]:

δτy= GbL(4)

where Δτy is the increase in yield stress, G is the shear modulus, b is the Burgers vector of dislocation. It is hard for dislocation emission in the interspacing between two adjacent F/C interfaces due to the high strength and limited gap (low L) inside the pearlite lamella (less than 200 nm). Therefore, crack propagation along the F/C interface is retarded.

It is well-known that the threshold stress intensity factor (KIH) decreases linearly with the increase of logarithm of the total hydrogen concentration CH [39], [40], [41]:

KIH=A-BlnCH (5)

For a 4340 steel, Gerberich et al. [42] found that:

A-$\frac{ kE }{ασ_{ys}}$ ($\frac{ RT }{\beta{\bar{V}}_{H}}$lnCH-$\frac{σ_{ys}}{2}$) ,B=$\frac{ kERT }{α\beta {\bar{V}}_{ { H σ}_{ys}}}$ (6)

where CH is the critical hydrogen concentration of HE, $\bar{V}$H is the partial molar volume of hydrogen, R is the gas constant, T is the absolute temperature, k, α and β are number constants, E is the Young’s modulus and σys is the yield strength. According to TDS analysis (Fig. 3(a)) and reports in the literature [24,25], the pearlite colony boundary, F/P boundary and dislocations (in ferrite) are the reversible hydrogen trapping sites. The CH in these locations can be saturated during the hydrogen charging process, as a result the threshold stress intensity factors KIH become low. Therefore, these locations are prone to HE crack initiation and propagation. The observed IG and QC fracture at the preliminary stage of crack propagation supports the above hypothesis.

With the increase of K, fracture mode changes from a mixture of IG and QC to an entire QC feature as observed in the center of the fracture surface (Fig. 5(d)). The plastic zone size ahead of crack tip increases with increasing of K [43]. In the case where the plastic zone is large enough to contain more part of the ferrite matrix. The dislocation density (Fig. 6(b)) and stresses can be very high. As a result, the cracking of pearlite colony and F/P boundary can be replaced by transgranular cracking in the ferrite matrix due to the activation of multiple planar slips [44]. This effect can be remarkable if high concentrations of hydrogen atoms are presented within the soft phase (ferrite), since hydrogen can accelerate the formation of the dislocation structure and promote coalescence of cracks along the intense slip band [13]. For the same reasons discussed above, pearlite is not involved in the high K fracture process.

5. Conclusions

The feature of hydrogen induced cracking initiation and propagation has been studied for a FP steel. The main conclusions are drawn as follows.

(1) Pearlite colony boundary, F/P boundary and the adjacent ferrite matrix are the HE crack nucleation sites. The high angle grain boundary instability in the presence of hydrogen and the stress concentration lead to the crack initiation on the pearlite colony boundary. Voids formation at the intersections of the F/P boundaries and slip band in the ferrite matrix result in the crack initiation on the F/P boundary and the adjacent ferrite matrix.

(2) There is no crack observed at the F/C interface. The interface with low energy and the limited gap for dislocation multiplication are the probable reasons.

(3) At low K, the IG and QC features are co-existence at the preliminary stage of crack propagation. With the increase of K, cracking mode changes to an entire QC fracture. The cracking of pearlite colony boundary and F/P boundary can be replaced by cracking of slip-band intersections in the ferrite matrix due to the large plastic zone and the activation of multiple planar slips facilitated by hydrogen.

Acknowledgments

This work was financially supported by the Joint Funds of the National Natural Science Foundation of China (Grant No. U1608257).

The authors have declared that no competing interests exist.


/