Effect of electrochemical hydrogen charging on the mechanical behavior of a dual-phase Ti–4Al–2V–1Mo–1Fe (in wt.%) alloy
电化学氢气充电对双相 Ti–4Al–2V–1Mo–1Fe （单位：wt.%） 合金力学行为的影响

1. Introduction 1. 引言

Due to their low density, specific strength, excellent corrosion resistance and high temperature resistance, titanium (Ti) and its alloys are widely used in aerospace, industrial, marine and biomaterials fields [[1], [2], [3], [4], [5]]. However, since Ti is a metal with high susceptibility to hydrogen [6], Ti-based materials are susceptible to hydrogen embrittlement (HE) in hydrogen containing environments [7,8], which is similar to many other metallic materials such as steels [9,10], magnesium (Mg) alloys [[11], [12], [13], [14], [15], [16]] and zirconium (Zr) alloys [17,18]. Generally, the existing forms of hydrogen in metals are either in solution, trapped at crystal defects or in the formed hydrides, and could subsequently affect their service properties [[19], [20], [21], [22], [23], [24], [25], [26], [27]]. For Ti and its alloys, the mechanism of HE is usually related to the formed hydrides [7,[22], [23], [24], [25], [26], [27]].
钛 （Ti） 及其合金由于其低密度、比强度、优异的耐腐蚀性和耐高温性，被广泛应用于航空航天、工业、海洋和生物材料领域 [[1]， [2]， [3]， [4]， [5]]。然而，由于钛是一种对氢高度敏感的金属[6]，因此钛基材料在含氢环境中容易发生氢脆（HE）[7,8]，这与许多其他金属材料相似，如钢[9,10]、镁（Mg）合金[[11]、[12]、[13]、[14]、[15]、[16]] 和锆 （Zr） 合金 [17,18]。通常，金属中现有的氢形式要么在溶液中，要么被困在晶体缺陷中，要么在形成的氢化物中，随后可能影响它们的服役特性 [[19]， [20]， [21]， [22]， [23]， [24]， [25]， [26]、[27]]。 对于钛及其合金，HE 的机理通常与形成的氢化物有关 [7，[22]、[23]、[24]、[25]、[26]、[27]]。

Since HE is one of the most harmful failure modes, HE issues of Ti-based materials have drawn much attention in the past few decades [6,23,[27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37], [38], [39], [40], [41]]. Among them, the effect of hydrogen on the mechanical behavior has always been one of the researching focuses [[28], [29], [30], [31], [32], [33], [34], [35], [36], [37], [38], [39], [40], [41]]. As for the hydrogen charging methods, most researchers used the way of gaseous hydrogen charging [[28], [29], [30], [31], [32], [33], [34], [35], [36]], whilst few researchers adopted the electrochemical hydrogen charging [[37], [38], [39], [40], [41]]. It is worth to be noted that the reaction conditions of two methods are different. Gas phase hydrogen charging usually occurs in hydrogen gaseous environment with high temperature and high pressure [[28], [29], [30], [31], [32], [33], [34], [35], [36]], whilst the electrochemical charging usually occurs in some electrolyte environments (such as marine) due to the cathodic protection or in galvanic coupling with heterogeneous metals [42]. In real service conditions, Ti and its alloys are more likely to undergo electrochemical hydrogen charging, and the relative failure cases have been widely reported [43,44]. Thus, it is of a great importance to study the effect of electrochemical hydrogen charging on the mechanical behavior of Ti-based materials. Previous studies mainly focused on the HE behavior of both the α-Ti and (α + β) dual-phase Ti alloys. Liu et al. [37] reported that the failure elongation ratio (ε_f) and the impact toughness of a Ti–Al–Zr (α-Ti) alloy were decreased gradually after being charged at 10 mA/cm² for 0–72 h, which was closely related to the formed hydrides at grain boundaries. Liu et al. [38] reported that for a pure-Ti being electrochemically charged at 75 mA/cm² for 0–100 h, the fracture mode transformed from ductile (0 h) to the ductile/brittle mix-mode (5–100 h), and the cleavage feature on the fracture was formed by the brittle cracking of δ-TiH_x and ε-TiH₂ hydrides. However, in previous work, it still lacks of the direct evidence for establishing the relationship between the cracking and formed hydrides. As for the (α + β) dual-phase alloys, Roy et al. [39] carried out the slow strain rate tensile testing for a Ti–Ni–Mo–Fe alloy under different charging potentials, and found that the ε_f was decreased by about 44% under −1168 mV (vs. Ag/AgCl), which was ascribed to the brittle cracking of the hydrides formed at the crack tip. However, the hydride-induced cracking process was not captured. Takumi et al. [40] reported that with the applied potential being changed from 200 mV (vs. SHE) to −1000 mV (vs. SHE), the ε_f of the Ti–6Al–4V alloy was decreased from 33% to 17%. For the sample being charged at −1000 mV (vs. SHE), γ-TiH hydride was detected by XRD but cannot be observed by using microscopy. Jeong et al. [41] reported that when the tensile test was performed in a 0.6 M NaCl solution at - 0.89 V (vs. SCE), the ε_f of an annealed Ti–6Al–4V alloy was reduced by 35% and the fracture near the outer surface was composed of cleavage and quasi-cleavage planes. However, no hydrides were detected or observed during the test. Based on the description mentioned above, the electrochemical hydrogen charging could lead to the mechanical degradation especially the ductility of both α-Ti and (α + β) dual-phase Ti alloys. Moreover, lots of studies demonstrated that the ductile degradation was due to the formation and cracking of brittle hydrides [[37], [38], [39], [40]]. However, in previous work, the phenomenon of hydride-induced cracking was usually judged by the brittle characteristics of fracture surfaces [[37], [38], [39]], whereas no direct observations to the cracking of hydrides were actually reported. Thus, it is hard to conclude that the formation of brittle fracture surface is mainly attributed to the hydride cracking. Additionally, whether the hydride cracking is the only cause for the degraded ductility still needs further investigation. Besides, in the early researches, the specific fracture process of the charged Ti alloys has been scarcely reported. Therefore, it is still unclear about whether the hydride cracking could influence the subsequent fracture behavior and ductility of the charged samples.
由于 HE 是危害最大的失效模式之一，因此 Ti 基材料的 HE 问题在过去几十年中引起了广泛关注 [6,23，[27]， [28]， [29]， [30]， [31]， [32]， [33]， [34]、[35]、[36]、[37]、[38]、[39]、[40]、[41]]。其中，氢对力学行为的影响一直是研究重点之一 [[28]， [29]， [30]， [31]， [32]， [33]， [34]， [35]， [36]、[37]、[38]、[39]、[40]、[41]]。 至于充氢方式，大多数研究人员采用气态充氢的方式[[28]， [29]， [30]， [31]， [32]， [33]， [34]， [35]， [36]]，而很少有研究人员采用电化学氢气充电 [[37]， [38]， [39]， [40]， [41]]。值得注意的是，两种方法的反应条件不同。气相充氢通常发生在高温高压的氢气环境中[[28]， [29]， [30]， [31]， [32]， [33]， [34]， [35]，[36]]，而由于阴极保护或与异质金属的电偶耦合，电化学充电通常发生在某些电解质环境（如海洋）中[42]。 在实际使用条件下，Ti 及其合金更容易发生电化学氢气充电，相对失效的情况已被广泛报道 [43\u201244]。因此，研究电化学氢电荷对 Ti 基材料机械行为的影响具有重要意义。以前的研究主要集中在 α-Ti 和 （α + β） 双相 Ti 合金的 HE 行为上。Liu等[37]报道，Ti-Al-Zr（α-Ti）合金在以10 mA/cm²充电0-72 h后，失效伸长率（ε）和冲击韧性逐渐降低，这与晶界形成的氢化物密切相关。Liu等[38]报道，对于以75 mA/cm²电化学充电0-100 h的纯Ti，断裂模式从延展性（0 h）转变为韧性/脆性混合模式（5-100 h），断裂上的解理特征是由δ-TiH_x和ε-TiH₂的脆性开裂形成的氢化物。然而，在以前的工作中，它仍然缺乏建立开裂和形成的氢化物之间关系的直接证据。对于（α + β）双相合金，Roy 等[39] 对 Ti-Ni-Mo-Fe 合金在不同充电电位下进行了慢应变速率拉伸测试，发现在 -1168 mV 下ε降低了约 44%（与 Ag/AgCl 相比），这归因于裂纹尖端形成的氢化物的脆性开裂。 然而，氢化物诱导的开裂过程没有被捕获。Takumi等[40]报道，随着施加的电位从200 mV（vs. SHE）变为−1000 mV（vs. SHE），Ti-6Al-4V合金的ε从33%降低到17%。对于以 −1000 mV 充电（相对于 SHE）的样品，XRD 检测到 γ-TiH 氢化物，但无法使用显微镜观察到。Jeong等[41]报道，当在-0.89 V的0.6 M NaCl溶液中进行拉伸试验时（与SCE相比），退火的Ti-6Al-4V合金的ε降低了35%，外表面附近的断裂由解理面和准解理面组成。然而，在测试过程中没有检测到或观察到氢化物。根据上述描述，电化学氢充电会导致机械退化，尤其是 α-Ti 和 （α + β） 双相 Ti 合金的延展性。此外，大量研究表明，延展性降解是由于脆性氢化物的形成和开裂 [[37]， [38]， [39]， [40]]。然而，在以前的工作中，氢化物诱导开裂的现象通常是通过断裂表面的脆性特性来判断的[[37]， [38]， [39]]，而实际上没有直接观察到氢化物开裂的报道。 因此，很难得出脆性断裂表面的形成主要归因于氢化物开裂的结论。此外，氢化物开裂是否是延展性降低的唯一原因仍需要进一步研究。此外，在早期研究中，带电 Ti 合金的具体断裂过程鲜有报道。因此，氢化物开裂是否会影响带电样品随后的断裂行为和延展性仍不清楚。

In this work, through investigating and comparing the mechanical behaviors of an as-cast dual-phase Ti–4Al–2V–1Mo–1Fe (Ti4211) alloy after being performed electrochemical hydrogen charging for different time, the effect of formed hydrides on the mechanical properties is studied. Moreover, the fracture modes of the differently charged samples are characterized.

2. Materials and experimental methods

Samples with a cross section of 10 mm × 10 mm and a thickness of 2 mm for microstructural characterization were cut from the cast ingot and sequentially ground with 300# to 3000# silicon carbide abrasive paper and polished to a 1 μm finish. Then, the polished surfaces were etched by using a Kroll reagent (3 mL HF, 9 mL HNO₃ and 88 mL H₂O) for 15 s and observed by using scanning electron microscopy (SEM; XL30-FEG ESEM). Phase analysis was conducted on an X-ray diffractometer (XRD; D/Max 2400) with Cu Kα radiation (wavelength: 0.154056 nm) under 40 kV and 40 mA, over a range of 30–80° at a step size of 0.02° and a scanning rate of 4°/min.

Samples for electrochemical hydrogen charging were mounted with epoxy resin, leaving an exposure area of 1 cm², and then respectively performed electrochemical hydrogen charging for 4, 8 and 24 h in a 3.5 wt % NaCl solution. In order to reflect the effect of hydrides formation on the mechanical degradation, an accelerated hydrogen charging process with a constant cathodic charging current density of 100 mA/cm² was employed. Meanwhile, this cathodic charging current density was also used for investigating the HE issues of other Ti alloys [25,46]. The hydrogen charging equipment was an electrochemical workstation (CorrTest, CS350). A classical three electrode cell with Pt counter electrode and saturated calomel reference electrode (SCE) was used. During the process of hydrogen charging, a thermostat water bath was used to maintain the temperature at 25 ± 0.5 °C. After being charged, samples were quickly cleaned with ethanol. Surface and the etched cross-sectional morphologies of the differently charged samples were characterized by using SEM. Moreover, phase analysis was conducted by using XRD.

Tensile samples with a gauge length of 25 mm, a cross-sectional area of 4 mm (in width) × 2 mm (in thickness) were cut from the cast ingot. Before the testing, samples were abraded, polished, etched and mounted with epoxy resin, and only one side of the gauge sections (with an area of 1 cm²) was exposed. Then, samples were pre-charged for 4, 8 and 24 h. Uniaxial tensile testing for differently charged samples were carried out by using the tensile testing machine (Care, IBTC-5000) at a strain rate of 1 × 10⁻⁴ s⁻¹ at room temperature (25 ± 2 °C). During the tensile testing, the overall surfaces and laterals of the samples were in-situ observed by using optical microscopy (OM; Keyence VHX-900F). After the tensile testing, the side surfaces near the fracture sites were observed by using OM. Moreover, the overall fracture surfaces and typical fracture characteristics were observed by using SEM.

3. Results

3.1. Microstructural characterization

Fig. 1 shows the microstructures of differently charged Ti4211 alloys. It can be seen from Fig. 1(a) that the uncharged alloy is composed of the large equiaxed α-Ti phases and the fine β-Ti phases. The average size of α-Ti phases is 10 μm, whereas that of the β-Ti phases is 5 μm. Fig. 1(b) shows that for the sample being charged for 4 h, the hydrogen-induced cracks with the average length of 5 μm appear in the interior of α-Ti phases and along α-Ti/β-Ti interfaces and most of them are blocked by β-Ti phases. For the sample being charged for 8 h, the quantities of cracks are further increased and most of them appear along the α-Ti/β-Ti interfaces, as shown in Fig. 1(c). With the charging time being prolonged to 24 h, cracks gradually connect with each other to form network structure, and their average length exceeds 10 μm, as shown in Fig. 1(d).

1. Download: Download high-res image (2MB)
2. Download: Download full-size image

Fig. 1. Microstructures of the differently charged samples: (a) 0 h, (b) 4 h, (c) 8 h and (d) 24 h.

It was generally accepted that the occurrence of hydrogen-induced cracking was owing to the unbearable volume expansion caused by the formed brittle hydrides [47]. To further disclose the existence and type of hydrides, the analyzed XRD results of differently charged samples are shown in Fig. 2. It can be seen that the uncharged alloy is composed of α-Ti and β-Ti phases. After being charged for 4 h, the diffraction peaks of δ-TiH_x appear. With the prolongation of charging time, the peak intensities of δ-TiH_x increase, whilst those of α-Ti phase decrease. For the sample being charged for 24 h, some peaks of the α-Ti phase become wider because they are overlapped by δ-TiH_x peaks [48]. Moreover, for the charged samples, the diffraction angles corresponding to the peaks of β-Ti phase slight decrease due to the expansion of lattice parameter caused by absorbed hydrogen [49].

1. Download: Download high-res image (359KB)
2. Download: Download full-size image

Fig. 2. XRD patterns of the differently charged samples.

Cross-sectional morphologies of the differently charged samples are shown in Fig. 3. Fig. 3(a–c) are respectively the overall morphologies of the samples being charged for 4 h, 8 h and 24 h. It can be seen that the areas near and far away from the hydrogen charging direction exist significant differences. In order to reveal the microstructures of different areas, high-magnified observations of the sample being charged for 24 h are shown in Fig. 3(d–f). Fig. 3(d) shows that α-Ti phases in the area near the hydrogen charging direction have been completely transformed into needle-like hydrides, and the area is designated as hydride layer. Moreover, the hydrogen-induced crack with a length of 8 μm can be observed in the interior of the hydrides. Fig. 3(e) shows the microstructure of the middle part defined as the hydrogen-affected area. In this part, α-Ti phases cannot be completely transformed into hydrides. Meanwhile, needle-like hydrides are preferentially formed in the interior of α-Ti phases and at α-Ti/β-Ti interfaces. The remaining area is not affected by hydrogen and can be designated as the unaffected matrix, as shown in Fig. 3(f).

1. Download: Download high-res image (4MB)
2. Download: Download full-size image

Fig. 3. Cross-sectional morphologies of the differently charged samples: (a) 4 h, (b) 8 h, (c) 24 h, images (d), (e) and (f) are high-magnified observations to the squared areas in image (c).

Based on the description mentioned above, the cross-sections of the charged samples can be roughly divided into three parts according to different microstructures, i.e. the hydride layer, hydrogen-affected area and unaffected matrix. With prolonging the charging time, both of the thicknesses of the hydride layers and the hydrogen-affected areas are increased. For the samples being charged for 4 h, 8 h and 24 h, the average hydride layer thicknesses are respectively 160 μm, 230 μm and 380 μm, and those of the hydrogen-affected areas are respectively 240 μm, 340 μm and 600 μm. Moreover, it is worth noting that in the hydride layer and hydrogen-affected area, no hydrides can be observed in the interior of β-Ti phases, which indicates that the β-Ti phase has the lower susceptible to HE than the α-Ti phase and α-Ti/β-Ti interface.

3.2. Mechanical properties

To reveal the influence of hydrogen charging on the mechanical properties of the alloy, tensile stress-strain curves of the differently charged samples are shown in Fig. 4.

1. Download: Download high-res image (260KB)
2. Download: Download full-size image

Fig. 4. Tensile stress-strain curves of the differently charged samples.

The determined values of ε_f, yield strength (σ_0.2), ultimate strength (UTS) and time to failure (t_f) are listed in Table 2. It can be seen that the σ_0.2 and UTS values of the sample being charged for 4 h are slightly lower than those of the uncharged sample. However, the ε_f is only 2.6%, which is about 79% lower than that of the uncharged sample (12.3%). With prolonging the charging time, the mechanical properties of the samples are continuously decreased. When the charging time is 24 h, the alloy exhibits the characteristic of brittle fracture and the ε_f is only 0.6%. Moreover, the t_f values are also decreased gradually with the increased charging time.

Table 2. Mechanical properties of the differently charged samples.

Hydrogen charging time/h	εf/%	σ0.2/MPa	UTS/MPa	tf/s
0	12.3 ± 0.5	832 ± 36	867 ± 29	1598 ± 64
4	2.6 ± 0.2	752 ± 18	784 ± 26	338 ± 26
8	1.8 ± 0.3	717 ± 7	721 ± 6	234 ± 39
24	0.6 ± 0.1	–	578 ± 3	78 ± 13

3.3. Failure analysis

Fig. 5 shows the quasi in-situ observations to the tensile process of the sample being charged for 8 h. Fig. 5(a) and (b) are respectively the overall surfaces before and during the tensile testing. Compared with the unstressed sample (Fig. 5(a)), obvious cracking occurs on the hydrogen charging surface after applying a tensile strain of 0.5%, whilst no cracks are formed in the area far away from the charging direction, as shown in Fig. 5(b). Besides the main crack for inducing the ultimate fracture of samples, secondary crack can also be observed, which can be attributed to the brittle cracking of the formed hydrides [50]. To disclose the locations of the cracks, high-magnified images to the lateral surface are shown in Fig. 5(c–f). Fig. 5(c) shows that the lateral morphology could be roughly divided into three parts. The area near and far away from the charging direction is respectively the hydride layer (Fig. 5(d)), hydrogen-affected area (Fig. 5(e)) and unaffected matrix (Fig. 5(f)). Moreover, the thicknesses of hydride layer and hydrogen-affected area are close to those of the cross-section of the sample being charged for 8 h (Fig. 3(b)). According to Fig. 5(c–f), it can be concluded that the cracking of hydride layer will be earlier than that in the matrix during the tensile testing.

1. Download: Download high-res image (3MB)
2. Download: Download full-size image

Fig. 5. Quasi in-situ observation of the sample being charged for 8 h during the tensile test: images (a) and (b) are respectively the overall surface at 0% strain and 0.5% strain, image (c) is high-magnified observations of the squared area in image (b), images (d), (e) and (f) are high-magnified observations of squared areas in image (c).

To reveal the effect of hydrogen charging on the fracture behavior, the overall fracture side and fracture surfaces of the differently charged samples are shown in Fig. 6. It can be observed from Fig. 6(a) that the fracture side surface of the uncharged sample exhibits a significant necking, which is the typical ductile characteristic. Fig. 6(b) exhibits that the necking degree of the sample being charged for 4 h is obviously smaller than that of the uncharged sample, indicating that the degradation of the ductility occurs. Moreover, secondary cracks with the length of several millimeters can be observed. Compared with the sample being charged for 4 h, the necking degree of the sample being charged for 8 h is further reduction, and its top fracture surface tends to be flat, as shown in Fig. 6(c). Fig. 6(d) is the sample being charged for 24 h, the sample shows an obviously flat fracture surface and almost no necking, which are the typical brittle fracture characteristics. Therefore, it further indicates that the ductility decreases gradually with the prolonged charging time.

1. Download: Download high-res image (2MB)
2. Download: Download full-size image

Fig. 6. Overall fracture side and fracture surfaces of the differently charged samples: images (a), (b), (c) and (d) are respectively the fracture side surfaces of the samples being charged for 0 h, 4 h, 8 h and 24 h, images (e), (f), (g) and (h) are the corresponding overall fracture surfaces. The high-magnified observations to the squared areas A ~ D in images (e) and (h) are shown in Fig. 7.

1. Download: Download high-res image (2MB)
2. Download: Download full-size image

Fig. 7. High-magnified fracture surfaces of the differently charged samples: image (a) is the uncharged sample, images (b), (c) and (d) are different areas of the sample being charged for 24 h.

Overall fracture surfaces of the differently charged samples are shown in Fig. 6(e–h). Fig. 6(e) is the uncharged sample, similar to the fracture side surface (Fig. 6(a)), it shows an obvious necking. As for the charged samples, the necking degree is reduced gradually with the increase of the charging time, as shown in Fig. 6(f–h). Moreover, according to the different characteristics, fractures of the charged samples could be divided into three parts. The area near the hydrogen charging top surface (area 1) shows the typical brittle characteristic, whilst the middle part (area 2) and the side far away from the top surface (area 3) present ductile characteristics. Moreover, with the increase of the charging time, the thicknesses of brittle areas (area 1) are increased. For the sample being charged for 4 h, its average thickness is about 160 μm. As for the samples being charged for 8 h and 24 h, their average thicknesses are respectively 230 μm and 380 μm. It is worth noting that the thicknesses of brittle areas of the differently charged samples are close to those of the hydride layers (Fig. 3), which demonstrates that the brittle characteristic on the fractures are formed by the brittle cracking of the hydride layers. To further reveal the effect of hydrogen charging on the fracture mechanism, high-magnified observations to the fractures of the uncharged sample and the sample being charged for 24 h (the squared areas A ~ D) are performed.

Fig. 7 shows the high-magnified fracture surfaces of the samples being charged for 0 h and 24 h. Fig. 7(a) reveals that the fracture surface of the uncharged sample is composed of dimples, which proves that the fracture mode of the alloy is ductile fracture. As for the sample being charged for 24 h, however, the area near the hydrogen charging direction shows quasi-cleavage feature, as shown in Fig. 7(b). The area is formed by the brittle cracking of the hydride layer and can be called as brittle area. Fig. 7(c) shows the middle part of the sample being charged for 24 h, a mix morphology of dimples and quasi-cleavage features can be observed, which proves that the brittle and ductile fracture occur simultaneously. Moreover, dimples in this part are radiate, which is the typical characteristic of tearing dimple [51]. Due to the transition of brittle to ductile fracture occurs, this part can be called as transition area. Fig. 7(d) exhibits the fracture morphology of the area far away from the hydrogen charging direction. Similar to the uncharged sample (Fig. 7(a)), it is also consisting of dimples. Thus, it can be inferred that the area is formed by the ductile fracture of unaffected matrix, which can be called as ductile area. Based on the description mentioned above, the fracture mode of the alloy changes from ductile to a mix mode of ductile and brittle after being charged.

4. Discussion

According to Fig. 4, ε_f values of the charged samples are decreased with the increase of the charging time. To disclose the mechanism for the ductile degradation, the fracture process of charged samples is analyzed. Fig. 8 is the schematic diagram of the fracture process (the cross-sectional morphology) of the sample being charged for 4 h. Fig. 8(a) shows that the uncharged alloy is composed of α-Ti and β-Ti phases. Fig. 8(b) is the sample being charged for 4 h, needle-like hydrides are preferentially formed in the interior of α-Ti phases and at α-Ti/β-Ti interfaces, whilst no hydrides are formed in the interior of β-Ti phases. This is because one of the conditions for the hydride formation is that the content of hydrogen in a certain phase lattice exceeds its limit solubility [25,26]. At room temperature, the body-centered cubic (BCC) structured β-Ti phase has a higher limit solubility of hydrogen than that of the hexagonal-closed packed (HCP) structured α-Ti phase [42]. Thus, hydrides are not easily formed in the β-Ti phases. Moreover, according to the different microstructures, the cross-section can be divided into three parts, i.e. hydride layer, hydrogen-affected area and unaffected matrix. It was reported that Ti-hydride was brittle and had a lower stress-intensity threshold for crack propagation than Ti matrix [24]. Thus, it can be deduced that the hydride layer is more prone to be cracked than other areas. In addition, it can be seen from Fig. 8(b) that several hydrogen-induced cracks appear in the interior of the hydride layer. Under the applied tensile stress, the cracks will merge with each other and propagate, which could further increase the pre-cracking possibility of the hydride layer. Based on the above circumstances, during the tensile testing, the whole hydride layer can be brittle cracked earlier than the rest of alloy, as shown in Fig. 8(c). The brittle cracking of the hydride layer will lead to the occurrence of the brittle area on the fracture surface. Then, since hydrides are only formed on the hydrogen charging side, the preferential cracking of the hydride layer is equivalent to form a prefabricated crack on this side, resulting in a stress concentration effect [51]. Thus, after the hydride layer being cracked, the sample will be subject to a tearing stress perpendicular to tensile direction, resulting in the rapid propagation of the crack tip. With the continuous tearing stress, the crack will first propagate into hydrogen-affected area, as shown in Fig. 8(d). When the crack propagates to the hydrides (either in the interior of α-Ti phases or at α-Ti/β-Ti phases), brittle cracking occurs and the quasi-cleavage feature will be formed on the fracture surface. Compared with the hydrides, the uncharged α-Ti and β-Ti phases are ductile phases [24]. Therefore, after two phases being torn, the fracture surface will present tearing dimples. When the crack propagates from the hydrogen-affected area to the unaffected matrix, it will decrease the remaining area of tensile samples. In the meantime, the subsequently plastic deformation occurred in remaining area contributes to the further tensile strain of tensile samples. Eventually, when the remaining cross-sectional area cannot further bear the applied stress, the ductile fracture will occur, as shown in Fig. 8(e).

1. Download: Download high-res image (1MB)
2. Download: Download full-size image

Fig. 8. Schematic illustrations of the fracture process of the charged Ti4211 alloy: images (a) and (b) are respectively the microstructures of the samples being charged for 0 h and 4 h, images (c), (d) and (e) are the fracture processes of the sample being charged for 4 h.

Based on the description mentioned above, the ductile degradation of the charged Ti4211 alloy can be attributed to the stress concentration effect caused by the preferential cracking of the hydride layer. Due to the stress concentration effect, the charged sample can be easily fractured. As a result, the ε_f value is decreased. With prolonging the charging time, the thicknesses of hydride layers are increased. It was reported that when the total thicknesses of the samples were fixed, the cracking of the thicker hydride layer would form the deeper prefabricated crack and lead to the stronger stress concentration effect [51]. Consequently, the tearing crack will propagate more rapidly and lead to the continuous decreases of the ε_f values.

5. Conclusions

The effect of electrochemical hydrogen charging on the mechanical properties and fracture behavior of an as-cast Ti4211 alloy was investigated. The main conclusions are drawn as follows:

• 1)

  After being electrochemically charged for 4 h, 8 h and 24 h, δ-TiH_x hydrides were formed in the interior of α-Ti phases and at α-Ti/β-Ti interfaces, which could cause the hydrogen-induced cracking. Meanwhile, α-Ti phases near the hydrogen charging direction could be completely transformed into hydrides.

• 2)

  The ε_f values of the samples being charged for 0 h, 4 h, 8 h and 24 h were respectively 13.2%, 2.6%, 1.8% and 0.6%. The ductile degradation of the charged samples was due to the stress concentration effect caused by the preferential cracking of hydride layers.

• 3)

  The fracture surfaces of differently charged samples could be divided into three parts, i.e. brittle area, transition area and ductile area. Among them, the brittle area was formed by the cracking of the hydride layer.

• 4)

  The fracture mode was transformed from ductile into ductile/brittle mixed mode for the samples being charged for longer than 4 h. With the increase of the charging time, the fracture mode of the alloy is gradually close to brittle fracture.

CRediT authorship contribution statement

S. Wang: Formal analysis, Writing - original draft, All authors have reviewed the manuscript. . D.K. Xu: Conceptualization, Writing - review & editing, Conceived the research and provided guidance, All authors have reviewed the manuscript. Z.Q. Zhang: Writing - review & editing, Contributed to the scientific discussions, All authors have reviewed the manuscript. Y.J. Ma: Methodology, Material preparation, All authors have reviewed the manuscript. Y.X. Qiao: Contributed to the scientific discussions, All authors have reviewed the manuscript.