Effect of hydrogen charging intensities and times on hydrogen embrittlement of Q&P980 steel

Q&P steel has good development prospects because of its excellent mechanical properties, but with the improvement in strength grade, hydrogen-induced delayed fracture (HIDF) is almost inevitable. In this paper, slow strain rate tensile tests and deep-drawn cup tests of Q&P980 steel under different hydrogen charging strengths and times were carried out, and the microstructure and fracture morphology were analysed by SEM. The results show that the plastic loss of Q&P980 steel was more obvious with increasing hydrogen charging intensity and hydrogen charging time, and a good elongation of 6.63% is still retained under the hydrogen content of 2.134 ppm. The deep-drawn cup samples were placed in acidic distilled water and alkaline and acidic solutions, and only a deep-drawn ratio of 1.9 showed HIDF in the three solutions. Specifically, 12 cracks were observed after soaking in HCl solution for two days. The main reason is that the martensite, austenite island and ferrite phase interface of Q&P980 steel increase stress during deformation and with the transformation-induced plasticity (TRIP) effect, resulting in hydrogen segregation at the phase interface and crack initiation leading to HIDF.


Introduction
Retaining carbon-rich metastable retained austenite (RA) in a martensitic matrix by a quenching and distribution (Q&P) process has been proven to be useful for multiphase high-strength steels without reducing their plasticity.This ingenious design is based on the transformation-induced plasticity (TRIP) of retained austenite during deformation, and the TRIP effect increases the work hardening rate and plasticity of highstrength steels [1][2][3].Q&P steel, carbide-free bainitic (CFB) steel, medium manganese (MMn) steel, etc, as the third generation advanced high-strength steel (AHSS), have good comprehensive mechanical properties and low alloy content, while meeting the needs of lightweight and collision safety, which have great potential in the lightweight development of the automotive industry [4,5].Hydrogen-induced delayed fracture (HIDF) has threatened the engineering applications of high-strength steels for decades [6,7].Hydrogen is the smallest atom in production and use and can be used in various ways in steel, such as pickling, electroplating, galvanizing and welding [6].Improving the hydrogen embrittlement resistance of steel has always been the research focus of advanced high-strength steel (AHSS).Usually, microdefects and microstructural optimization are used to improve hydrogen traps to suppress HIDF, such as dislocations, carbides, retained austenite, and chemical heterogeneity [8][9][10][11][12].Xu et al found that NbC can become an effective hydrogen trap, significantly improving the hydrogen embrittlement resistance of AHSS treated by the Q&P process [13].When the atomic concentration of carbides is the same, the hydrogen capture ability of carbides decreases in the order of NbC > TiC > VC [14].Austenite is also considered an effective hydrogen trap and is beneficial for reducing HE sensitivity due to its low diffusion rate and high solubility [15,16].Preliminary studies have shown that Q&Ptreated steel has a higher hydrogen embrittlement sensitivity than traditional quenched and tempered (Q&T)treated steel [17].Li et al [18] found in the U-shaped bending experiment of 1180 MPa AHSS that the block-like retained austenite in the Q&P steel reduced its resistance to hydrogen-induced delayed fracture.Zhu et al's study showed that [19] thin film retained austenite has higher HE resistance than block retained austenite.Further research on the origin of cracks indicates that hydrogen-induced cracks in Q&P steel initiate at the martensite/ austenite interface, while no cracks are observed in the ferrite phase.The fracture characteristics before and after hydrogen charging transition from ductile dimple fracture with ductile micro-pores to a 'quasi-cleavage' zone and mixed morphology of intergranular cracking [20,21].In addition, by introducing the ferrite phase to reduce the hydrogen embrittlement sensitivity of low-carbon steel, ferrite can reduce the stress concentration caused by martensitic transformation and delay crack propagation during deformation [22].Wang et al [23] studied the HIDF of two types of Q&P steel and found that hydrogen-induced cracks nucleate in sensitive areas such as retained austenite or strain-induced fresh martensite and tend to propagate along the original austenite grain boundaries and martensite grain boundaries.Ferrite in the steel can be used to passivate hydrogen-induced cracks and interrupt the continuity of the original austenite grain boundaries and martensite interfaces.Li et al [24] showed that almost all hydrogen-induced cracks originate from segregation bands and ultimately lead to fracture, and the order of local hydrogen concentration is as follows: segregation bands > retained austenite > original austenite grain boundaries > bulk grain boundaries > martensite > ferrite.Therefore, further research should be conducted on the influence of the stability of retained austenite with different morphologies and positions on hydrogen embrittlement (HE) sensitivity to help elucidate the HE mechanism.
In this work, we carried out slow strain rate tensile experiments and deep-drawn cup static experiments on Q&P980 steel under different hydrogen charging intensities and times and analysed the microstructure and fracture morphology to study the HIDF mechanism of martensite, ferrite and retained austenite.

Experimental material and methods
The materials used in this study were Q&P980 steel produced by a plant.The specific chemical composition is given in table 1, where the C content is 0.2 (wt%) and alloying elements are relatively small, mainly austenitic stability elements such as Mn and Si, which are typical characteristics of the third-generation advanced highstrength steel element addition.The starting temperature of hot rolling was 1250 °C, from 250 mm rolling to 3.5 mm, the final rolling temperature was 880 °C, and coiling occurred at 650 °C.After acid cleaning and rust removal, the samples were cold rolled to 1.6 mm.The continuous annealing homogenization temperature was 800 °C, the quenching temperature was 250 °C, and the quenching and partitioning temperature was 400 °C.
Before the tensile test, the sample was subjected to electrochemical continuous hydrogen charging in a 0.25 mol l −1 H 2 SO 4 aqueous solution containing 500 mg l −1 of NH 4 SCN, and a hydrogen charging current density of 10 ∼ 70 mA cm −2 .Different charging times and hydrogen charging intensities were given.The hydrogen content was measured using a heating dehydrogenation analysis unit (TDS) produced by the Japanese R-DEC company.The tensile properties of Q&P980 steel were tested at room temperature by slow strain rate testing (SSRT), with a strain rate of 5 × 10 −5 s −1 .The dog bone shaped specimen used for tensile testing has a gauge distance of 25 mm and a width of 6 mm.The microstructure and fracture morphology were observed using a scanning electron microscope (SEM) (TESCAN VEGA 3); the samples for SEM were polished and etched with 4% nitric acid.For most metals and alloys, hydrogen charging can lead to a significant decrease in plasticity [25].Hydrogen-induced relative plastic loss (I δ ) was generally used as a measure of hydrogen embrittlement sensitivity, which was calculated as shown in equation (1) [26]: where 0 d and H d represent the elongation of the unfilled and hydrogen-charged states, respectively.
Based on the actual forming process of high-strength steel parts, the delayed cracking phenomenon of Q&P980 steel was studied by using a deep-drawn cup static experiment.First, the plates were wire-cut into 3 round billets each with diameters of 75, 85 and 95 mm.The billets were drawn deep into cup-shaped samples using the plate forming testing machine, and the surface was polished with 200 grid sandpaper.The test draw ratios were 1.5, 1.7 and 1.9, and the samples are shown in figure 1.Then, 0.1 mol l −1 NaOH aqueous solution and 0.1 mol l −1 HCl aqueous solution were prepared, and the control group was distilled water.The deep-drawn cups with different draw ratios were placed in acidic, alkaline and neutral environments for delayed fracture tests, and the number of cracks and crack growth morphology were recorded every day.

Microstructure of Q&P980 steel
Figure 2 shows the SEM image, x-ray diffraction (XRD) pattern, inverse pole figure (IPF) and phase distribution of Q&P980 steel.The multiphase structure of martensite and ferrite could be observed by SEM, and there were some granular island structures at the interface of the martensite and ferrite phases, which were generally retained austenite, and some white bright film tissues could be observed between the martensite laths, which may be thin-film retained austenite.   of retained austenite was 15.2% by XRD analysis, and this metastable austenite was the main source of plasticity improvement in the TRIP effect.Figure 2(d) was composed of the superposition of the phase map and the band contrast map, where the red was the retained austenite structure, which was mainly distributed at the phase boundary.

HE sensitivity of Q&P980 steel under different hydrogen charging intensities and times
Figures 3(a) and (b) stress-strain curves of Q&P980 steel with hydrogen charging current and charging time as variables, respectively.It can be seen that with increasing current density and charging time, the sample elongation (δ) and tensile strength decreased significantly, showing a high hydrogen embrittlement sensitivity.Table 2 shows the characterization chart of hydrogen-induced relative plastic loss (I d ) of samples corresponding to figure 2. The control group was not filled with hydrogen.Sample A was filled with hydrogen for 5 min with a current density of 10 mA cm −2 ; sample B was filled with hydrogen for 5 min with a current density of 30 mA cm −2 ; sample C was filled with hydrogen for 5 min with a current density of 50 mA cm −2 ; sample D was filled with hydrogen for 5 min with a current density of 70 mA cm −2 ; and sample E was filled with hydrogen for 10 min with a current density of 10 mA cm −2 .The hydrogen-induced relative plasticity loss of sample D was the most serious, and all samples retained a good elongation of more than 6%.On the contrary, the tensile strength varies less with the hydrogen charging current density and time, but only decreases by 63 MPa as the hydrogen charging current density increases from 0 to 70 mA cm −2 , which may be mainly due to the high yield strength (∼780 MPa) and strong strain hardening ability of the test steel.Figure 4 shows the hydrogen escape rate-temperature curve and hydrogen content of Q&P980 steel under different hydrogen charging process conditions.It can be seen from the figure that the hydrogen charging sample had a strong hydrogen escape peak near 130 °C and a weak hydrogen escape peak near 400 °C.For the unfilled test, the hydrogen content was 0.221 ppm, which mostly came from the smelting and heat treatment process of the steel.When the hydrogen charging current density of Q&P980 was 10 mA cm −2 , the hydrogen content was 1.397 ppm after 5 min of hydrogen charging, and the hydrogen content was 1.777 ppm after the hydrogen charging time was further extended for 10 min.When the hydrogen charging current density was increased to 70 mA cm −2 , the hydrogen content was significantly increased to 2.134 ppm. Figure 5(a) shows the relationship between hydrogen-induced relative plastic loss and hydrogen charging current density during a hydrogen charging time of 5 min.It can be seen that the I δ of the sample gradually increased with increasing hydrogen charging current density.When the current density increased from 0 mA cm −2 to 10 mA cm −2 , the I d increase of the sample was the largest, and then with the continuous increase of the current density, the growth    6(a), the macro section of the HE fracture was flat, no tearing cracks were found, and there was no obvious necking phenomenon.The crack source of the fracture surface after hydrogen charging appears at the edge of the sample, and during the fracture process, the crack propagates from one end to the other, which is significantly different from that of the nonhydrogen charged sample.The crack growth region of the fracture of sample E consisted of a large number of dimple morphologies and a small number of quasi-cleavage morphologies (figure 6(f)), while the central part of the crack source presented obvious cleavage brittle fracture characteristics (figure 6(h)), but the brittle fracture area was relatively small in general, which is consistent with the results observed by Liu et al [21] on the HE fracture of Q&P steel.
To further observe the formation of cracks, the crack expansion side of the hydrogen-filled sample was ground, polished and etched by nitrate alcohol to observe crack initiation.Figure 7 shows SEM images of the fracture side of the hydrogen-filled sample.We statistically analysed the hydrogen fractures under four different  hydrogen charging conditions and found that crack nucleation mainly occurred at three interfaces: the front granular austenite of the island structure and the trifurcated interface of ferrite and martensite; the granular austenite interface; and the martensite and ferrite interface.However, Zhu et al [19] and Wang et al [23] only found crack sources at the martensite/austenite interface of Q&P steel.

HDIF of Q&P980 deep-drawn cup
The deep-drawn cup was soaked in different solutions for 14 days, and the crack occurrence was recorded with a camera regularly every day, with only a 12 h interval recorded on the first day.As shown in figure 8, no crack was found in the deep-drawn cup with draw ratios of 1.5 and 1.7 in any solution, and cracks were found only on the cup with a draw ratio of 1.9 (figures 8(a)-(c)).The number of cracks on the deep-drawn cup in 0.1 mol l −1 HCl aqueous solution was much larger than that in distilled water and 0.1 mol l −1 NaOH aqueous solution, the number of cracks reached 5 after soaking for 12 h, and 12 was recorded after two days, then remained same (figure 8(d)).The cracks generated on the deep-drawn cup first appear at the edge of the cup mouth, then expanded longitudinally to the bottom and did not penetrate the bottom of the cup, but expanded along the circumferential direction of the deep-drawn cup, as shown in figure 9.

Discussion
Hydrogen exists in the following forms in metals: H, H + , H − , H 2 , metal hydride, CH 4 gas mass, etc Among them, molecular hydrogen is generally present in the defects inside the metal, such as voids, small cracks and grain boundaries, and H 2 accumulates at the internal defects to produce internal pressure.The activity of  molecular hydrogen formed in the metal is very poor, which generally cannot be diffused through the lattice to the surface; even if heated, it is difficult to remove, so it is easy to produce hydrogen bulges.However, in this work, no hydrogen bulge was found due to the short hydrogen charging time, and the maximum hydrogen content in the sample was only 2.134 ppm, almost half of that in reference [27].There are currently two popular hydrogen embrittlement theories.One is the weak bond theory proposed by Troiano [28], which was later improved by Oriani [29].It is believed that the essence of the weak bond theory is that the 3d 5 4s 2 conduction band of iron, which is not fully filled, overlaps with the hydrogen atom 1s 1 , causing the repulsive force between the iron atoms in the matrix to be greater than the attractive force.Another theory is the hydrogen-enhanced local plasticity (HELP) theory.Lynch [30] believes that any fracture process is the result of local plastic deformation, and hydrogen can promote plastic deformation; therefore, under lower external stress, the plastic deformation in the local deformation zone at the crack tip can reach a critical state, leading to delayed fracture.In other words, hydrogen can generate critical plastic deformation, thereby reducing the critical stress field intensity factor K IC required for crack propagation.
It can be seen from the analysis in figure 6 that the nucleation location of hydrogen-induced cracks was closely related to retained austenite, a large number of micro-cracks nucleated at the relevant interface of retained austenite, and few micro-cracks and pores were observed in martensite and ferrite.The volume expansion effect of martensitic transformation during quenching leads to high residual internal stress at the interface between martensite and retained austenite [31].In addition, the TRIP effect of retained austenite during deformation may increase the degree of stress concentration, so hydrogen tends to segregate [32] at the interface of martensitic austenite with higher stress and form micro-cracks and voids under the combined action of stress and hydrogen.Xiong et al [33] showed that Paralympic stability was seriously related to the impact of morphology.The TRIP effect in the deformation process of Q&P steel first occurred in the massive Paralympic with high carbon content, while the thin film retained austenite with low carbon content remained stable when the engineering strain was as high as 12%.In previous studies on HE, the microscopic characteristics of quasi cleavage and brittle inter-granular fracture were often explained by hydrogen-enhanced decohesion (HEDE), while the ductile microporous aggregation fracture mode was explained by the synergistic effect of plastic mediated HELP mechanism [34].These comprehensive HE characteristics have been observed in different grades of steel, including TRIP steel [35][36][37][38][39].In this experiment, after yielding Q&P980 multiphase steel, the TRIP effect occurs in granulated austenite, forming brittle twin martensite and inducing hydrogen-induced crack nucleation at the interface of retained austenite, martensite and ferrite (as shown in figure 7).This can be explained as inter-granular brittle fracture or quasi cleavage brittle fracture controlled by the HEDE mechanism, as shown in figure 6(h) [40].Ferrite belongs to the soft phase, and plastic deformation occurs during the crack propagation process to alleviate the stress concentration phenomenon at the crack tip, resulting in crack passivation [23,41].At the same time, due to the low hydrogen content in ferrite, no micro-cracks or micropores were formed in ferrite, which has also been confirmed in DP and Q&P steels containing ferrite [42,43].Due to the low degree of hardening of the ferrite phase and relatively small plastic strain, micro-cracks generated at the ferrite martensite interface can be eliminated without further propagation.Therefore, the location where brittle cracks occur is limited to the interface or hard martensite; Therefore, the main failure mode during crack propagation is controlled by the ductile fracture mechanism, characterized by dimples and a small portion of quasi cleavage features (figure 6(f)).Previous studies have also shown that when hydrogen concentration is low, the HELP mechanism can lead to an increase in micro-void coalescence (MVC) fracture [40].
Based on the above experimental results and analysis, the nucleation and growth process of hydrogeninduced cracks in Q&P980 composite steel were as follows: After the formation of the main crack, micro-crack nucleation was induced near the retained austenite region near the crack tip, and a large number of micro-pores were generated in the martensitic structure.Under the continuous action of stress, the micro-cracks in the retained austenite tended to merge to form secondary cracks, and the main cracks expanded forwards through the combination of secondary cracks.When the main crack met the martensite structure during the merging process, it preferentially expanded along the interface between the martensite slat and the grain, forming a quasicleavage fracture morphology (figure 6(f)).When the ferrite structure was encountered in the merging process, it was passivated due to the plastic deformation of the ferrite.However, as the deformation continued, plastic instability was induced due to the limited deformation ability of the ferrite, and the main crack eventually broke through the ferrite grain (figure 7(d)).
In the experiment, an interesting phenomenon was found.By observing the macroscopic morphology of tensile fracture of unfilled and hydrogen-filled SSRT, it was found that, as shown in figure 6(e), the fracture diverged from the central position to the specimen surface, but after hydrogen charging, the fracture occurred first on the specimen surface.This result showed that during the hydrogen charging process, although hydrogen diffusion can theoretically reach the central position, the hydrogen content on the surface of the sample was greater than that at the central position, resulting in the fracture source usually occurring on the surface, which is similar to the phenomenon observed by Wang et al [42].
For the deep-drawn cup experiment, two factors are mainly involved: the environment and the amount of material deformation.For the draw ratio of 1.9, rust appeared on the surface when soaked in distilled water, and one crack was found after the first day without increasing in number, as shown in figure 8(a).As shown in figure 9b, when soaked in 0.1 mol l −1 NaOH aqueous solution, the surface of the deep-drawn cup remained bright, and after three days, one crack appeared, and the number did not increase.After soaking in 0.1 mol l −1 HCl aqueous solution for 0.5 days, 5 cracks were generated on the deep-drawn cup, and with the extension of soaking time, 12 cracks were finally generated, as shown in figure 8(c).In environmental terms, if hydrogen gas touches a metal surface, the molecules will be adsorbed on the metal surface and then further decompose into atomic hydrogen on the surface (chemisorption) [44].Even at room temperature and pressure, there will be a certain amount of atomic hydrogen on the surface of the metal, which can enter the interior of the metal by desorption [16].In the metal, the atomic interaction force is in equilibrium, but on the surface, the atomic coordination number is smaller than in the body, so the atomic interaction force is unbalanced, and the metal has a larger surface energy.This unbalanced interaction force can attract heterogeneous atoms to the surface (physical adsorption), resulting in a decrease in the surface energy of the metal (system energy decrease).Therefore, the metal has a tendency to adsorb heterogeneous atoms.H 2 molecules adsorbed on metal surfaces can decompose into atoms in the presence of a thermal activation energy and then adsorb onto the surface through covalent forces.
Because there are not enough hydrogen ions in distilled water and alkaline solutions, very little hydrogen enters the interior of the metal, and in acidic solutions, hydrogen atoms generated by hydrogen evolution corrosion in the anode can easily enter the interior of the metal.Chemisorption requires less activation energy, so there are fewer H atoms entering the interior of the metal, and most of them spill out as H 2 .However, precisely due to this small number of residual H atoms, the number of cracks produced by the material changed greatly.In figure 8(d), the Q&P980 deep-drawn cup was soaked in acid solution for 3 to 14 days, and the number of cracks was 12 times that in water and alkali solution.In terms of deformation, the deformation of Q&P980 steel generated residual stress.With the increase in the shape variable, the residual stress increased, and the number of cracks increased.Therefore, it can be considered that the residual stress has a proportional relationship with crack initiation; that is, there is a safe draw ratio, and it is believed that the maximum safe draw ratio will continue to increase with increasing pH value, as shown in figure 8(d).In addition, when the Q&P980 steel material is deformed, a large number of dislocations will be generated in the matrix, and a small amount of retained austenite will produce the TRIP effect to form fresh martensite.There are a large number of dislocations in these fresh martensites, which will also provide a favourable place for the entry and migration of hydrogen.The cracks generated on the deep-drawn cup did not extend through the bottom of the cup after spreading to the bottom but spread along the deformation region, as shown in figure 9, which also indicates that deformation promotes the growth of cracks figure 9.

Conclusions
In this work, we carried out SSRT and deep-drawn cup static experiments on Q&P980 steel under different hydrogen charging intensities and times, and the microstructure and HIDF mechanism were investigated.The main conclusions are as follows: 1.With the increase in hydrogen charging strength and the extension of hydrogen charging time, the hydrogeninduced plastic loss of Q&P980 steel is more obvious, but it still retains a good elongation of more than 6%.
2. In the hydrogen environment, the internal stress and high-density dislocation existing in the martensite and austenite microstructure will lead to hydrogen enrichment, which easily produces cracks.The retained austenite inside is ineffective as a hydrogen trap, but when the TRIP effect occurs, a large number of hydrogen atoms will be released, which will improve the HE sensitivity of the material.Although ferrite has always played a role in inhibiting crack initiation and propagation, during the stretching process, micro-cracks formed by the interface separation between the ferrite and martensite can easily cause hydrogen atoms to aggregate, leading to HE.
3. The test of the deep-drawn cup shows that the stress concentration caused by deformation and the increase in dislocation density will lead to the occurrence of HIDF, and crack initiation is more likely with increasing hydrogen content in the solution.

Figure 2 (
b) shows the XRD results with a scanning angle range of 45 ∼ 95°, it can be found that the diffraction intensity of ferrite/martensite peaks in (200)α and (211)α is the highest, while the diffraction intensity of austenite peaks in (200)γ, (220)γ, and (311)γ is relatively weak.The volume fraction

Figure 1 .
Figure 1.Physical picture of cup-shaped samples.

Figure 2 .
Figure 2. SEM microstructure (a), XRD diffraction pattern (b), IPF diagram (c) and phase diagram (d) of the Q&P980 original structure after EBSD characterization treatment.The red in the phase diagram is retained austenite.

Figure 3 .
Figure 3. (a) Engineering stress-strain curves for different charging current densities; (b) engineering stress-strain curves for different charging times.

Figure 5 (
b) shows the change relationship between Iδ and hydrogen charging time at a hydrogen charging current density of 10 mA cm −2 , which also shows a trend of rapid increase at first and then a slow growth rate.Hydrogen was biased on the surface of the sample during hydrogen charging, which will be discussed later.Figures 6(a)-(d) shows the SEM morphology of SSRT tensile fracture of the control sample, where figure 6(d) is the enlarged image of the rectangular region in figure 6(c).It can be seen from the figure that the fracture had an obvious necking phenomenon (figure 6(a)), and the surface of the fracture had undulation.The crack source was located in the centre of the fracture, and the crack spread from the centre to the periphery when the sample was fractured.Many cracks with sizes of 100-20 μm can be observed in the crack source area parallel to the RD and TD directions.The crack source (figures 6(c) and (d)) and crack growth region (figure 6(b)) were both distributed with small and dense dimples, showing obvious ductile fracture characteristics.Figures 6(e) ∼ (h) show the typical hydrogen embrittlement fracture of hydrogen-filled sample E, where figure 6(h) is the enlarged image of the cyan region in figure 6(g).As shown in figure

Figure 4 .
Figure 4. Hydrogen desorption rate curve and hydrogen content of test steel under different hydrogen charging process conditions.

Figure 5 .
Figure 5.The relationship between hydrogen damage and (a) charging current density; (b) charging time.

Figure 7 .
Figure 7. SEM microstructure of the fracture side of the Q&P980 hydrogen charged sample: (a), (b), (c) and (d) correspond to samples A, C, D and E, respectively.

Figure 8 .
Figure 8. Variation in the number of cracks in deep-drawn cups with different draw ratios for three kinds of solutions: (a) distilled water, (b) 0.1 mol l −1 NaOH aqueous solution, and (c) 0.1 mol l −1 HCl aqueous solution.(d) The number of cracks in the deepdrawn cup in different solutions when the deep-drawn ratio is 1.9.

Figure 9 .
Figure 9.For different time periods the drawing cup (a) top view and (b) side view in 0.1mol/L HCl solution in 0.1mol/L HCl solution.

Table 2 .
Characterization of hydrogen damage with different charging conditions.
rate gradually became flat.