The Effect of Hydrogen Embrittlement on Fracture Toughness of Cryogenic Steels

Abstract

This study investigates the effect of hydrogen embrittlement on the fracture toughness of 9% Ni steel and STS 316L stainless steel under cryogenic conditions ranging from −80 °C to −253 °C. Hydrogen charging was performed using electrochemical methods, and hydrogen uptake was quantitatively analyzed using thermal desorption spectroscopy (TDS). Fracture toughness was evaluated using crack tip opening displacement (CTOD) testing per ISO 12135, both without hydrogen (WO-H) and with hydrogen (W-H). The results showed a gradual decrease in CTOD values with decreasing temperature in both steels under hydrogen-free conditions, with ductile fracture maintained even at −253 °C. In contrast, hydrogen-charged specimens exhibited significant toughness degradation at intermediate subzero temperatures (−80 °C to −130 °C), particularly in 9% Ni steel due to its BCC crystal structure. However, at −160 °C and below, the effect of hydrogen embrittlement was suppressed mainly owing to the reduced hydrogen diffusivity. Scanning electron microscopy (SEM) analysis confirmed the transition from ductile to brittle fracture with decreasing temperature and hydrogen influences. At −253 °C, fully brittle fracture surfaces were observed in all specimens, confirming that at ultra-low temperatures, fracture behavior is dominated by temperature effects rather than hydrogen. These findings identify a practical temperature limit (approximately −160 °C) below which hydrogen embrittlement becomes negligible, providing critical insights for the design and application of structural materials in hydrogen cryogenic environments.

1. Introduction

Hydrogen has emerged as a promising clean-energy carrier in response to global efforts to achieve carbon neutrality. As hydrogen infrastructure continues to expand into pipelines, refueling stations, storage vessels, and mobility platforms, the materials used in these environments must endure exposure to hydrogen without compromising their mechanical integrity. A major challenge in this context is hydrogen embrittlement (HE), a phenomenon in which hydrogen degrades the mechanical properties of metals, leading to unpredictable and catastrophic failures. Hydrogen embrittlement manifests as a decrease in ductility, fatigue life, and fracture toughness, often occurring well below the yield or ultimate strength of a material. This is particularly critical in high-strength steels, titanium alloys, and nickel-based superalloys, where small concentrations of hydrogen can dramatically increase the susceptibility to cracking [1]. Of all the mechanical properties affected, fracture toughness, which quantifies the resistance to crack initiation and propagation, is among the most critical when assessing material reliability under hydrogen exposure. As liquid hydrogen application technology expands to cryogenic applications, understanding how HE influences fracture behavior becomes increasingly urgent. Fractures in hydrogen pipelines, storage tanks, and the structural components of ships can have extreme safety and economic consequences. Therefore, ensuring sufficient fracture toughness in the presence of hydrogen is vital for design and material selection.

Numerous studies have investigated the relationship between the hydrogen content and fracture toughness across different classes of materials. Olden et al. [2] performed extensive fracture toughness testing on hydrogen pre-charged steels and demonstrated a strong inverse relationship between $K_{IC}$ values and hydrogen charging time. In addition, their results also exhibited a more substantial degradation effect in higher-strength alloys. Yamabe et al. [3] studied the behavior of tempered martensitic steels under diffusible and trapped hydrogen conditions. Their analysis using J-integral methods revealed that hydrogen-trapping sites, such as carbide–matrix interfaces, could delay but not prevent embrittlement under long-term exposure. Lynch [4] presented a comprehensive synthesis of HE mechanisms and correlated various failure modes with metallurgical features, such as grain size, retained austenite content, and phase distribution. Robertson et al. [5] recently emphasized the use of in situ microscopy and atomistic simulations to track hydrogen–dislocation interactions, enabling a more accurate modeling of hydrogen damage at the crack tip. Chen et al. [6] highlighted the need for unified standards that account for modern high-strength alloys and realistic service conditions. Despite decades of research, accurately predicting the impact of hydrogen on fracture toughness remains a challenge, particularly in multiphase or ultra-high-strength steels, where microstructural heterogeneity complicates the embrittlement behavior [7].

In addition, hydrogen embrittlement has a profound effect on the crack tip opening displacement (CTOD) and tensile behavior, both of which are widely used to assess ductility and structural integrity under service-like loading conditions. Multiple studies have reported a notable reduction in the CTOD values under hydrogen-charged conditions, even in materials that exhibit high ductility in air [8]. The presence of hydrogen near the crack tip increases the local stress triaxiality, reduces the cohesive strength, and promotes the nucleation of microvoids or intergranular cracks at lower levels of plastic strain. For example, Yamabe et al. [3] observed that the reduction in the CTOD due to hydrogen exposure was significantly more pronounced in high-strength steels with tempered martensite than in those with ferrite–bainite structures. Depending on the charging time and microstructural trapping capacity, the degradation ratio often exceeds 50%. This reduction in the CTOD implies an increased risk of crack initiation at relatively low external loads, which is particularly critical for pressure vessels and pipelines operating under fluctuating loads.

In parallel, tensile tests conducted under hydrogen pre-charged or in situ exposure conditions have shown marked decreases in the elongation at break, area reduction, and ultimate tensile strength (UTS). The most typical manifestation is the transition from ductile necking-type fractures to brittle intergranular or quasi-cleavage fracture surfaces. These changes are particularly evident in notched specimens, where hydrogen-assisted cracking is often initiated in stress-concentrated zones [9]. The extent of hydrogen-induced degradation in tensile properties depends on several factors: hydrogen concentration, charging method (gaseous vs. electrochemical), strain rate (with slower strain rates generally leading to more severe embrittlement), and microstructural features such as grain boundary character, dislocation density, and inclusion distribution. Researchers, such as Takai [7], have emphasized the time-dependent nature of hydrogen embrittlement, demonstrating that lower strain rates facilitate greater hydrogen diffusion into stressed regions, resulting in an earlier onset of cracking.

The degradation of these mechanical properties under HE conditions reinforces the need for robust material qualification, environmental compatibility testing, and the evaluation of hydrogen-specific fracture mechanics. In particular, previous studies have focused on HE characteristics in terms of strength under tensile loads. Moreover, there are minimal studies on the initiation and propagation of cracks, including CTOD, under compressive loads that can evaluate fracture toughness. In this study, the effect of hydrogen embrittlement on the fracture toughness of body-centered cubic (BCC)-structured materials (Ni steels) and face-centered cubic (FCC)-structured materials (austenitic stainless steels, STS) was investigated. These materials are commonly used to fabricate storage tanks in liquefied hydrogen carrier systems. It is well established that BCC and FCC metals exhibit markedly different behaviors in response to hydrogen, particularly at cryogenic temperatures (approximately −165 °C), resulting in varying degrees of susceptibility to hydrogen embrittlement and associated degradation in mechanical performance. Two primary objectives are pursued to address this drawback.

(1)

The influence of temperature on the fracture toughness of ferritic (BCC) and austenitic (FCC) materials was quantitatively assessed through CTOD, δ, tests conducted at varying cryogenic temperatures;

(2)

The degradation in fracture toughness due to HE was examined following electrochemical hydrogen charging, with an emphasis on the correlation between hydrogen behavior (HE due to diffusion) and fracture resistance.

(3)

The temperature range over which HE occurred, based on the characteristics of the ferritic (BCC) and austenitic (FCC) materials, was experimentally determined, and the threshold temperature at which the resistance to HE was maintained is identified.

2. Materials and Test Methods

2.1. Material Properties

To evaluate the stable applicability in a liquid hydrogen environment, cryogenic mechanical property tests, including a tensile test and fracture toughness test, were conducted at −253 °C. In addition, tensile tests were carried out over a range of temperatures to investigate the variation in mechanical properties as a function of temperature. The tests were performed on materials used for liquefied natural gas (LNG) storage tanks, which are Ni steel, a representative cryogenic steel with a BCC structure, and STS steel, a representative cryogenic steel with an FCC structure. In addition, the composition of the materials was 9% nickel steel (9% Ni steel) and STS 316L steel, which are widely used in cryogenic applications because of their excellent mechanical performance at low temperatures. The chemical compositions and room-temperature mechanical properties of both steels are summarized in

Table 1. Chemical composition of 9% Ni and STS 316L steels (wt. %).

Steel	C	Si	Mn	P	S	Ni
9% Ni	0.05	0.23	0.64	0.003	0.0006	9.24
STS 316L	0.015	0.51	1.87	0.36	0.03	10.68

and

Table 2. Mechanical properties of 9% Ni and STS 316L steels.

Material	Yield Stress, (MPa)	Tensile Stress (MPa)	Elongation (%)	Charpy Impact Test, 25 °C, (J)
9% Ni	652	704	24	349, 340, 335
STS 316L	229	610	71	432, 415, 395

, respectively.

As presented in Table 2, both 9% Ni steel and STS316 steel exhibit excellent low-temperature impact toughness and have been widely applied in LNG storage tank applications. Therefore, they are considered sufficiently applicable as candidate materials for the fabrication of liquefied hydrogen storage tanks. However, this requires not only the validation of their mechanical performance at ultra-low temperatures but also an understanding of their behavior in hydrogen-rich environments, particularly concerning hydrogen embrittlement (HE) [10]. Generally, austenitic stainless steels such as STS 316L exhibit superior resistance to HE compared to materials with a BCC structure, such as 9% Ni steel. This is primarily attributed to the lower hydrogen diffusivity and higher trapping capacity associated with FCC microstructures. Consequently, STS 316L may offer a higher margin of safety in applications involving the simultaneous exposure to cryogenic temperatures and hydrogen. Nonetheless, for the design and safe operation of future liquefied hydrogen systems, it is essential to conduct comparative assessments of the mechanical properties and HE susceptibilities of both materials under relevant service conditions.

2.2. Tensile Test Method

To evaluate the cryogenic mechanical properties of the materials, tensile tests were conducted on 9% Ni steel and STS 316L stainless steel base metals according to the ASTM E8/E8M-23 standard test method [11]. Ultra-low-temperature tensile testing was performed using a cryostat-integrated universal testing machine (Cryo H&I, Pyeongtaek-si, Republic of Korea) specifically designed for subzero applications. Figure 1 shows the cryogenic tensile testing system employed in this study, capable of reaching temperatures as low as −260 °C. The cooling chamber was equipped with three 4 K-capacity cryo-coolers to minimize thermal disturbances during testing. It requires approximately 90 min to reach −253 °C prior to test initiation.

The testing system had a maximum load capacity of 300 kN, with a variable crosshead speed ranging from 600 to 0.005 mm/min. All the cryogenic tensile tests in this study were conducted at a constant crosshead speed of 0.005 mm/min to ensure high measurement precision at low temperatures. The dimensions and geometries of the tensile test specimens are shown in Figure 2. The specimens had a gauge length of 25 mm and a parallel section diameter of 7 mm. The test temperature was monitored using a diode sensor positioned at the edge of the specimen to ensure accurate thermal measurements. The experimental and calculated tensile properties of both materials are listed in

Table 3. Experimental and calculated tensile properties of 9% Ni and STS 316L steels.

Temperature	Materials	Yield Strength, Rp0.2 (MPa)	Tensile Strength, Rm (MPa)	Strength Ratio, Rp0.2/Rm
−10 °C	STS 316L	287	706	0.40
−20 °C	STS 316L	301	749	0.40
−30 °C	STS 316L	315	791	0.39
−50 °C	STS 316L	341	854	0.39
−80 °C	9% Ni *	751	826	0.91
	STS 316L	385	1002	0.38
−100 °C	STS 316L	443	1059	0.41
−130 °C	9% Ni *	852	919	0.93
−140 °C	STS 316L	469	1255	0.37
−160 °C	9% Ni	925	1046	0.88
−253 °C	9% Ni	991	1858	0.65
	STS 316L	631	1771	0.35

* Tensile properties temperature calculated according to ISO 15653 [12].

. As expected, both the 0.2% yield strength ($R_{P0.2}$) and the ultimate tensile strength ($R_m$) increased as the temperature decreased. However, the increase in yield strength was more significant than that in tensile strength, leading to a higher strength ratio ($R_{P0.2}$/$R_m$) at lower temperatures. Specifically, the strength ratio increased from 0.91 to 0.65 for 9% Ni steel and from 0.41 to 0.35 for STS 316L as the temperature decreased from −80 °C to −253 °C, indicating a greater contribution of yield strength to the overall mechanical response under cryogenic conditions.

In 9% Ni steel, to provide input data for the CTOD evaluations at various cryogenic temperatures, yield and tensile strengths at −80 °C and −130 °C were estimated based on empirical correlations specified in ISO 15653 (International Organization for Standardization) [12], using Equations (1) and (2). However, a noticeable reduction in the strength ratio was observed at −253 °C, suggesting a potential change in deformation behavior at ultra-low temperatures. The tensile properties at −160 °C and −253 °C were experimentally determined and used for subsequent fracture toughness analysis.

$$R_{P0.2}\left(T\right)=R_{P0.2}\left(atroomtemperature\right)+\left({\displaystyle \frac{105}{491+1.8T}}\right)-189$$

(1)

$$R_m\left(T\right)=R_m\left(atroomtemperature\right)[0.7857+0.2423exp(-{\displaystyle \frac{T}{170.646}}\left)\right]$$

(2)

where T denotes the fracture test temperature $(T>-196°\mathrm{C})$.

The tensile properties of STS316L were obtained directly from tensile tests conducted at each experimental temperature. It should be noted that ISO 15653 is applicable only to ferritic structures and is therefore not suitable for materials with an austenitic structure, such as STS316L

2.3. Fracture Toughness Test Method

Fracture toughness was evaluated for 9% Ni and STS 316L steels using CTOD measurements based on three-point bending (3PB) tests at cryogenic temperatures, including −253 °C. The 3PB specimens were designed with a deep through-thickness notch and a ratio of the initial crack length to the specimen width of a/W = 0.50, where a denotes the nominal pre-crack length and W denotes the specimen width. Given the importance of the fatigue pre-crack quality in fracture toughness testing, pre-crack growth was carefully monitored and controlled to ensure precise advancement to the target length during the pre-cracking procedure. CTOD evaluations were conducted in accordance with ISO 12135 [12]. A 500 kN universal testing machine (Cryo H&I, Pyeongtaek-si, Republic of Korea) equipped with a cryogenic cooling chamber was used to perform the tests. The crosshead displacement rate is maintained at 1 mm/min. Tests were conducted at various cryogenic temperatures. Since the two materials exhibit hydrogen embrittlement at different temperature ranges, fracture toughness evaluations were conducted under distinct temperature conditions. For 9% Ni steel, the tests were carried out between −80 °C and −160 °C, whereas for STS316L, they were performed between −10 °C and −140 °C. CTOD tests were conducted at the same temperatures as the tensile tests for both steels, and the yield and tensile strengths obtained from the tensile tests were applied to calculate the strength ratio for CTOD. The crack mouth opening displacement (CMOD) was measured using a 10 mm clip gauge mounted across the notch mouth. To achieve ultralow temperatures, two cryocoolers capable of reaching 4 K were used to cool the test chamber. These coolers minimize the thermal impact of heat generated during the bending process. A duration of approximately 90 min was required to reach −253 °C. A silicon diode sensor maintained the temperature within ±2 °C of the target. Following the ISO 12135 guidelines, the fracture toughness measurements commenced after holding the test temperature for 1 min per millimeter of specimen thickness (total hold time of 40 min). The CTOD test specimens of the 9% Ni and STS 316L steels are shown in Figure 3. Given its BCC microstructure, 9% Ni steel exhibits relatively high hydrogen diffusivity and susceptibility to hydrogen uptake [13]. Therefore, as shown in Figure 3a, standard CTOD specimens were used for hydrogen charging and subsequent fracture testing with a 20-mm-thick steel plate. In contrast, STS 316L exhibits lower hydrogen diffusivity and limited through-thickness penetration. Therefore, as shown in Figure 3b, a modified CTOD specimen geometry was developed to enable more efficient hydrogen charging in STS 316L. The CTOD values for both specimens were calculated using Equation (3) from ISO 12135.

$$\delta ={\left[\left({\displaystyle \frac{S}{W}}\right){\displaystyle \frac{F}{{\left(B{B}_{N}W\right)}^{0.5}}}{g}_1\left({\displaystyle \frac{a_0}{W}}\right)\right]}^2{\displaystyle \frac{\left(1-{\nu }^2\right)}{m{R}_{p0.2}E}}+\tau {\displaystyle \frac{0.4(W-a_0)V_p}{0.6a_0+0.4W+a_0}}$$

(3)

here, $\tau =\left[-1.4{\left({\displaystyle \frac{R_{p0.2}}{R_m}}\right)}^2+2.8\left({\displaystyle \frac{R_{p0.2}}{R_m}}\right)-0.35\right]\left\{0.8+0.2exp\left[-0.019\left(B-25\right)\right]\right\}$,

$$m=4.9-3.5{\displaystyle \frac{R_{p0.2}}{R_m}}$$

where S denotes the span between the outer loading points in the three-point bend test, F denotes the applied force, B denotes the specimen thickness, B_N denotes the specimen net thickness between the side grooves, a₀ denotes the final crack length, ν denotes Poisson’s ratio, E denotes Young’s modulus, V_p denotes the CMOD, R_p0.2 represents the 0.2% offset yield strength perpendicular to the crack plane at the test temperature, and R_m denotes the ultimate tensile strength perpendicular to the crack plane at the test temperature.

2.4. Hydrogen Charging Method

In this study, hydrogen charging was conducted using the cathodic electrolytic method, an electrochemical technique compliant with ISO 16573-1 [14]. This method involves the electrochemical generation of hydrogen atoms at the cathode, which is connected to the specimen, thereby allowing atomic hydrogen to diffuse into the metal substrate. The hydrogen charging conditions detailed in

Table 4. Hydrogen embrittlement test conditions with 9% Ni and STS 316L steels.

Items	9% Ni	STS 316L
Test method	Cathodic electrolytic method	Cathodic electrolytic method
Test current density	50 A/m2	100 A/m2
Test hydrogen electrolyte	3% NaCl + 0.3% NH4SCN	3% NaCl + 0.3% NH4SCN
Test time	48 h	72 h
Test temperature	19 °C (Room temperature)	80 °C

were implemented at a current density of 50–100 A/m² within a broader operational range of 0–20 A/m², as recommended by ISO 16573. The electrolyte used for the charging process was a mixture of 3% NaCl and 0.3% NH₄SCN. In addition, hydrogen was introduced at room temperature (19 °C) with 48 h for 9% Ni steel. However, in the case of STS 316L, which possesses an FCC structure, hydrogen charging is more challenging than that in the case of 9% Ni steel with a BCC structure. To ensure sufficient hydrogen absorption, STS 316L specimens were charged under elevated temperature conditions of 80 °C for a duration of 72 h after introducing a fatigue pre-crack into the specimens. Generally, hydrogen penetration is limited to a depth of approximately 0.05 mm from the surface in FCC-structured materials [15,16]. However, in this study, since fracture toughness was evaluated, it was not feasible to prepare specimens with an extremely small thickness, such as 0.1 mm. Therefore, 2 mm-thick plate specimens of STS316L were prepared in this study to enable effective hydrogen charging and fracture toughness testing across the thickness range. In the case of BCC-structured materials, hydrogen diffusion is relatively facile compared to that in FCC-structured materials; therefore, the specimen was maintained at room temperature for 48 h following hydrogen charging to ensure sufficient hydrogen uptake. In contrast, for the FCC-structured material, which exhibits lower hydrogen diffusivity and solubility, the specimen was maintained at an elevated temperature of 80 °C for 72 h to facilitate effective hydrogen charging. Of the two materials investigated, 9% Ni steel is relatively susceptible to hydrogen permeation and desorption at room temperature. To minimize hydrogen loss, the specimens were immediately stored in liquid nitrogen (−196 °C) after hydrogen charging and maintained there until testing. Prior to testing, the cooling chamber of the CTOD apparatus (Cryo H&I, Pyeongtaek-si, Republic of Korea) was pre-cooled to the target temperature. The hydrogen-charged specimens were then transferred directly from liquid nitrogen into the chamber and stabilized at the test temperature for approximately 20 min (1 min/thickness) before fracture toughness (CTOD) testing. The storage time in liquid nitrogen was kept consistent under all conditions, and fracture toughness evaluations were performed following the same procedure. As shown in Figure 4, the specimen was immersed in the electrolyte solution and connected to the negative terminal (cathode), and a platinum mesh was employed as the counter electrode (anode). Platinum mesh was selected because of its high electrical conductivity, chemical stability, and resistance to oxidative decomposition [17]. A photograph of the hydrogen-charging setup is shown in Figure 4, in which the three specimens were charged simultaneously under identical electrochemical conditions. Susceptibility to HE can be attributed to several known mechanisms. In hydrogen-enhanced decohesion (HEDE), hydrogen atoms are preferentially adsorbed at defect sites, such as grain boundaries or inclusions, reducing the cohesive strength and facilitating brittle fracture initiation. Hydrogen-enhanced localized plasticity (HELP) and hydrogen increase dislocation mobility, promote localized plastic deformation, and enable crack nucleation and propagation. Under certain thermodynamic and microstructural conditions, hydrogen-induced phase transformations can trigger phase transformations that reduce the material’s mechanical integrity. Adsorption-induced dislocation emission (AIDE) and hydrogen adsorption on metal surfaces may stimulate the release of dislocations, further weakening the microstructure [10,17,18,19,20,21,22,23]. Hydrogen in metallic materials exists primarily in two states. The first is diffusible hydrogen that can freely migrate through a metal lattice. It is characterized by a low binding energy (activation energy of 10–60 kJ/mol) and is typically released at temperatures below 400 °C. The second is nondiffusible (Trapped) hydrogen, which resides in microstructural trap sites, such as dislocations, vacancies, inclusions, and grain boundaries and exhibits a higher binding energy (>60 kJ/mol) and a desorption temperature exceeding 400 °C. Owing to its strong interactions with trap sites, nondiffusible hydrogen can remain in the metal for extended periods, significantly contributing to its long-term embrittlement behavior [24,25].

3. Results and Discussion

3.1. Quantitative Analysis of Hydrogen Content

In this study, the amount of hydrogen absorbed into the material was quantitatively analyzed using TDS to measure the amount of hydrogen absorbed into the test piece immediately after hydrogen charging [26]. TDS is a technique that measures the quantity of gas desorbed from a specimen as it is gradually heated, typically over a temperature range of 25 °C to 700 °C. Desorbed hydrogen was detected using a mass spectrometer or thermal conductivity detector, and the total amount of hydrogen was determined by integrating the signals over time. The results are expressed as the weight fraction of hydrogen relative to the specimen mass (wppm). Using the TDS method, the hydrogen concentrations in the CTOD specimens subjected to the electrochemical hydrogen charging procedure were 0.422 weight parts per million (wppm) for the 9% Ni steel and 0.840 wppm for the STS 316L steel. The results of the hydrogen uptake analysis for both steels, conducted using TDS, are presented in Figure 5. In materials with the BCC crystal structure, hydrogen diffusion is relatively rapid owing to the shorter diffusion paths and easier interstitial migration. This promotes localized stress concentration, which can significantly enhance susceptibility to HE [27]. Previous studies have shown that in BCC materials, ductility and fracture toughness can be significantly degraded at hydrogen concentrations as low as approximately 0.5 wppm, a level sufficient to initiate brittle fracture mechanisms [18]. Therefore, the hydrogen content introduced in this study was considered sufficient for investigating the embrittlement behavior and its influence on the fracture toughness of BCC-structured cryogenic steels. The diffusion behavior of hydrogen in metallic materials is strongly influenced by their crystallographic structures. In contrast, materials with FCC structures, such as austenitic stainless steels (e.g., STS 316L), exhibit significantly lower hydrogen diffusivity. The FCC lattice was more densely packed with smaller interstitial spaces, which hindered the movement of hydrogen atoms. Consequently, hydrogen penetration in FCC materials tends to be shallow and slow, often limited to the surface or near-surface regions, even after extended charging durations. Although this low diffusivity provides higher resistance to HE, hydrogen can become trapped once absorbed, and localized embrittlement may still occur under specific conditions. This fundamental difference in hydrogen diffusion rates between the BCC and FCC structures is a key factor in material selection for hydrogen-related environments and in designing effective hydrogen charging and embrittlement resistance testing protocols.

3.2. Effect of Temperature on Strength and Fracture Toughness

CTOD evaluations were conducted to investigate the effect of temperature on the fracture behavior of 9% Ni steel and STS 316L without hydrogen charging, as summarized in

Table 5. Effect of temperature on fracture toughness with 9% Ni and STS 316L steels without hydrogen charging.

Specimens	Test Temp.	a0, (mm)	a0/W	Material Properties (MPa)			CTOD, δ (mm)
				${\mathit{\sigma }}_{\mathit{Y}\mathit{S}\mathit{P}}$	${\mathit{\sigma }}_{\mathit{T}\mathit{S}\mathit{P}}$	E	Indiv.	Avg.
9% Ni	−80 °C	12.51	0.63	751	826	206,000	0.73	0.73
	−100 °C	11.54	0.58	767	862	206,000	0.75	0.69
		12.38	0.62	767	862	206,000	0.62
	−130 °C	12.51	0.63	852	919	206,000	0.62	0.62
	−160 °C	12.52	0.63	925	1046	206,000	0.49	0.49
	−253 °C	21.99	0.55	991	1528	206,000	0.04	0.03
		22.32	0.56	991	1528	206,000	0.03
		22.63	0.57	991	1528	206,000	0.02
STS 316L	−10 °C	11.21	0.56	229	610	228,500	0.60	0.64
		11.20	0.56	229	610	228,500	0.68
	−20 °C	12.04	0.60	229	610	228,500	0.65	0.65
	−30 °C	11.07	0.55	341	854	228,500	0.56	0.56
	−50 °C	10.88	0.54	341	854	228,500	0.65	0.65
	−80 °C	11.04	0.55	443	1059	228,500	0.67	0.64
		10.62	0.53	443	1059	228,500	0.61
	−140 °C	10.77	0.54	484	1269	228,500	0.45	0.45

. To assess the consistency of fatigue crack introduction, the ratio of fatigue crack length to specimen width (a₀/W) was confirmed to be within the range of 0.53 to 0.63, indicating minimal deviation among specimens. In the absence of hydrogen charging (WO-H), the CTOD values decreased with temperature, which was attributed to the increased atomic binding force and reduced ductility under cryogenic conditions. Despite the decrease in CTOD values to 0.73 mm, 0.69 mm, 0.62 mm, and 0.49 mm, the specimens exhibited excellent fracture resistance even under ultra-low temperature conditions. In addition, the strength increases under low-temperature conditions. In the case of STS 316L, the overall trend was consistent with that observed for 9% Ni steel. However, the increase in strength was more pronounced at ultralow temperatures, resulting in a greater enhancement in both the tensile and yield strengths. Figure 6 shows the load–displacement (P–V) curves and fracture surfaces at each temperature under hydrogen-free conditions. At all test temperatures, partial brittle fracture features were observed; however, the overall fracture morphology was predominantly ductile. Although CTOD evaluation is typically conducted with three repetitions per temperature, this study performed one or two repetitions to obtain baseline data on the fracture behavior under hydrogen influence under consistent experimental conditions. Within the temperature range examined in this study, no brittle fractures were observed up to the maximum load, and the fracture surfaces exhibited a ductile (soft) morphology. Figure 7 presents the load–displacement (P–V) curves and corresponding fracture surfaces of the STS 316L specimens under hydrogen-free conditions. The STS 316L material consistently displayed ductile fracture features at all tested temperatures, with no evidence of brittle cracking after reaching the maximum load. For both the 9% Ni and STS 316L steels evaluated in this study, all tested temperature ranges exhibited ductile fracture behavior. No evidence of brittle fracture was observed, indicating that the temperature did not induce brittle failure under the tested conditions. Although a gradual decrease in the CTOD values was observed with decreasing temperature, this reduction was attributed to the natural decline in toughness typically associated with lower temperatures. Notably, the decrease was not abrupt, suggesting that the material retained a stable fracture resistance even under cryogenic conditions.

3.3. Effect of Hydrogen on Fracture Toughness

Table 6. Effect of temperature on fracture toughness with 9% Ni and STS 316L steels with hydrogen charging.

Specimens	Test Temp.	a0, (mm)	a0/W	Material Properties (MPa)			CTOD, δ (mm)
				${\mathit{\sigma }}_{\mathit{Y}\mathit{S}\mathit{P}}$	${\mathit{\sigma }}_{\mathit{T}\mathit{S}\mathit{P}}$	E	Indiv.	Avg.
9% Ni	−80 °C	12.73	0.64	751	826	206,000	0.32	0.32
	−100 °C	12.02	0.60	767	862	206,000	0.31	0.31
		12.55	0.63	767	862	206,000	0.31
	−130 °C	12.09	0.60	852	919	206,000	0.38	0.39
		12.61	0.63	852	919	206,000	0.39
	−160 °C	12.30	0.65	925	1046	206,000	0.43	0.43
STS 316L	−10 °C	13.82	0.69	229	610	228,500	0.56	0.57
		10.82	0.54	229	610	228,500	0.59
	−20 °C	11.07	0.55	229	610	228,500	0.60	0.55
		13.08	0.65	229	610	228,500	0.52
		10.42	0.52	229	610	228,500	0.52
	−30 °C	11.22	0.56	341	854	228,500	0.56	0.56
	−50 °C	10.79	0.54	341	854	228,500	0.63	0.62
		11.07	0.55	341	854	228,500	0.61
	−80 °C	11.08	0.55	443	1059	228,500	0.57	0.55
		10.91	0.54	443	1059	228,500	0.54
	−140 °C	11.40	0.57	484	1269	228,500	0.41	0.41

presents the results of the fracture toughness evaluations conducted to investigate the influence of HE behavior on 9% Ni and STS 316L steels. To ensure consistency with the hydrogen-free specimens (WO-H), the fatigue crack length-to-specimen width ratio (a₀/W) was examined. Because all test specimens were fabricated under identical mechanical conditions, the hydrogen-charged specimens (W-H) exhibited stable fatigue crack insertion.

The CTOD tests for the 9% Ni steel hydrogen-charged specimens were performed at the same temperatures as those for the hydrogen-free condition (−80 °C, −100 °C, −130 °C, and −160 °C). The corresponding CTOD values were 0.32 mm, 0.31 mm, 0.39 mm, and 0.43 mm, respectively. Unlike the expected reduction in the fracture toughness with decreasing temperature, the hydrogen-charged specimens did not follow this trend. Despite the low-temperature conditions, an increase in the fracture toughness was observed. This behavior is attributed to the relatively more active influence of HE at higher subzero temperatures rather than to the temperature effect itself, leading to a degradation in the fracture resistance. In 9% Ni steel, at −80 °C, where hydrogen activity is relatively high, the CTOD value was the lowest, indicating a more pronounced HE influence. This suggests that both temperature and hydrogen effects are superimposed at this intermediate-low temperature. Interestingly, the CTOD value at −160 °C—where the fracture toughness is typically low due to cryogenic brittleness—was the highest among the hydrogen-charged specimens. This anomaly is likely because of the significant reduction in hydrogen diffusivity at ultralow temperatures, which limits the mobility and influence of hydrogen on the fracture process. Under such conditions, temperature dominates the fracture behavior, whereas the effect of hydrogen becomes negligible. As supported by previous studies [28], a decrease in hydrogen diffusivity at low temperatures reduces the potential for hydrogen-assisted decohesion or embrittlement.

A similar phenomenon was also observed in the STS 316L stainless steel; however, the temperature range at which this behavior appeared was higher than that of the 9% Ni steel. Although STS 316L exhibits limited hydrogen permeability, it retains absorbed hydrogen for extended periods owing to its low diffusivity. In contrast to 9% Ni steel, which exhibited minimal influence from hydrogen at cryogenic temperatures, a significant reduction in fracture toughness due to hydrogen was observed in STS 316L within the relatively higher temperature range of −10 °C to −80 °C. This behavior is attributed to differences in the temperature range at which hydrogen becomes mobile within the microstructure of each material. Consequently, it is suggested that the hydrogen charging temperature conditions should be tailored according to the crystal structure of the base material, with specific consideration given to the different diffusion behaviors of BCC and FCC alloys.

Figure 8 shows the load–displacement (P–V) curves and fracture surfaces of each hydrogen-charged 9% Ni specimen. At −80 °C, predominantly ductile fracture features were observed, although some regions showed hydrogen-induced brittle morphology. The P–V curve displayed a typical post-peak softening behavior. At −100 °C and −130 °C, although differences in maximum load were observed, the CTOD values were relatively consistent, and fracture surfaces showed mixed features, indicating the combined effects of temperature and hydrogen. At −160 °C, the fracture surfaces were mostly ductile with partial brittle areas, and the P–V curves showed softening after peak load, similar to the hydrogen-free condition. This suggests that at −160 °C, the influence of temperature degradation was dominant, whereas the hydrogen effect was significantly diminished.

Figure 9 shows the load–displacement (P–V) curves and fracture surfaces of each hydrogen-charged STS 316L specimen. Due to the influence of hydrogen, a reduction in displacement was observed, with the most pronounced effects occurring in the temperature range of −10 °C to −80 °C. At −140 °C; however, the influence of hydrogen was minimal and scarcely evident. For materials with a BCC crystal structure, a distinct change in mechanical response was observed across the −70 °C to −160 °C range, indicating strong sensitivity to HE within this interval. In contrast, FCC-structured materials demonstrated a delayed response to hydrogen effects, suggesting that hydrogen-related degradation becomes prominent only at relatively higher subzero temperatures. This difference was attributed to the lower hydrogen diffusivity and uptake typically associated with FCC materials.

Across all the test temperatures, similar P–V curve patterns and fracture modes were recorded. Interestingly, the overall displacement increased as the temperature decreased. However, based on conventional tensile testing results [29], it is generally observed that hydrogen charging leads to a decrease in displacement, highlighting the complex interaction between hydrogen and temperature effects on the fracture behavior.

3.4. Limit Temperature for Hydrogen Embrittlement Free Behavior

To assess the influence of HE on the fracture toughness, CTOD tests were conducted under both hydrogen-free (WO-H) and hydrogen-charged (W-H) conditions. Figure 10 shows the load–displacement (P–V) curves at various test temperatures. Figure 10a shows the P–V curves of the 9% Ni steel. At −80 °C, the WO-H specimen exhibited a maximum load of 19.6 kN and a displacement of 3.2 mm. In contrast, the W-H specimen recorded a maximum load of 17.2 kN (approximately 88% of the WO-H value) and a displacement of 1.4 mm (44% of the WO-H value). At −100 °C, the WO-H-#1 specimen showed a maximum load of 25.9 kN and 2.8 mm displacement, whereas the corresponding W-H-#1 specimen exhibited 20.8 kN and 1.2 mm, respectively. The WO-H-#2 specimen reached 22 kN and 2.6 mm, compared to 19 kN and 1.3 mm in W-H-#2, representing 86% and 50% of the WO-H values, respectively. At −130 °C, the WO-H specimen exhibited a maximum load of 21.6 kN and a displacement of 2.6 mm. The W-H-#1 and W-H-#2 specimens recorded 22.6 kN and 1.4 mm, and 21 kN and 1.6 mm, respectively—104% and 97% of the load, and 54% and 62% of the displacement compared to WO-H. At −160 °C, the WO-H specimen reached 22.6 kN and 2.1 mm, whereas the W-H specimen showed 19.9 kN and 1.9 mm, corresponding to approximately 88% of the load and 90% of the displacement of the WO-H condition. Figure 10b shows the P–V curves of STS 316L steel. The sensitivity of the FCC structure to hydrogen was lower than that of the BCC structure. As a result, hydrogen-induced effects were less evident in the load–displacement (P–V) curves. Furthermore, hydrogen charging is challenging because of the inherently low hydrogen diffusivity of FCC materials. To facilitate hydrogen uptake, the specimens were fabricated with a reduced thickness of 2 mm. However, this thin geometry makes it difficult to obtain stable and clearly defined P–V curves during fracture testing.

Hydrogen charging resulted in a reduction in both load and displacement. In certain cases, the localized stress concentration due to hydrogen led to slightly increased loads, although the general trend indicated reduced mechanical performance. These findings are consistent with those of conventional tensile testing, in which hydrogen charging has been shown to reduce the ductility. In 9% Ni steel, at −160 °C, the W-H specimen unexpectedly exhibited a higher load than the WO-H specimen, which is attributed to differences in the initial crack size (a₀/W), rather than an intrinsic improvement in fracture resistance. At such ultralow temperatures, hydrogen mobility is significantly suppressed, and the fracture behavior is predominantly governed by the temperature effect alone.

Figure 11 shows the variation in the CTOD values with temperature under both hydrogen-free and hydrogen-charged conditions in the 9% Ni and STS 316L steels. Figure 11a shows the effect of HE on 9% Ni steel. Under the WO-H condition, the CTOD values decreased with decreasing temperature for both steels, which is consistent with the results of previous studies. For the W-H condition, a more pronounced reduction in CTOD was observed at higher subzero temperatures (e.g., −80 °C), where both HE and thermal effects are active. The overlap of HE and temperature-induced brittleness led to a combined degradation in fracture toughness. However, at −160 °C, the CTOD values for WO-H and W-H were nearly identical, indicating that hydrogen had minimal influence under these conditions because of the suppression of diffusivity and activity at cryogenic temperatures. Figure 11b shows the effect of HE on STS 316L steel. Experiments were conducted on hydrogen-free (WO-H) and hydrogen-charged (W-H) specimens at temperatures ranging from −10 to −140 °C. Overall, the influence of hydrogen was found to be minimal. In particular, at temperatures below −80 °C, the effect of hydrogen was negligible, with no discernible difference in fracture toughness values. At relatively higher subzero temperatures, such as −50 °C, −30 °C, and −10 °C, slight differences were observed between the two conditions; however, these differences were small and did not indicate a significant effect, unlike the clear trend observed in 9% Ni steel. This behavior is attributed to the intrinsic characteristics of FCC structures, in which the hydrogen diffusion rate is significantly lower than that of BCC structures. Consequently, the hydrogen-induced degradation of FCC materials tends to occur at comparatively higher temperatures. At −150 °C, the effect of HE was minimal, and the observed reduction in fracture toughness was primarily attributed to the intrinsic degradation of mechanical properties due to the low temperature. Therefore, for STS316L, the effect of hydrogen embrittlement under the present experimental conditions was considered minimal, making it difficult to establish a distinct trend.

Figure 12 presents the SEM analysis of the fracture surface in 9%Ni of WO-H and W-H specimens at −80 °C. The WO-H specimen displayed a fully ductile fracture surface characterized by dimples and micro-voids. The W-H specimen, however, exhibited a partially brittle fracture surface, indicating HE and explaining the observed reduction in toughness. For the STS316L specimens, fracture surfaces at −10 °C are shown for both the WO-H and W-H conditions. While both specimens predominantly exhibited ductile fracture morphology, a slight brittle feature was observed at the notch tip of the W-H specimen. This localized brittle region is considered to result from shallow hydrogen penetration. Although it contributed to a minor reduction in fracture toughness at −10 °C, the penetration depth was insufficient to significantly affect the overall fracture toughness. Figure 13 shows the SEM analysis at −160 °C; the W-H specimen primarily exhibited ductile fracture features, with only limited brittle areas. As hydrogen activity declined with decreasing temperature, the dominant factor influencing fracture behavior became the inherent loss of plastic deformability and atomic cohesion due to the cryogenic environment, rather than HE. In the WO-H specimens, dimples, voids, and partial brittle fracture features were observed. At −160 °C, both W-H and WO-H specimens exhibited similar CTOD values. This suggests that even in the presence of internal hydrogen, the activity of hydrogen is significantly suppressed at such low temperatures, and the observed reduction in fracture toughness is primarily due to the effects of temperature alone. As the temperature decreases, the effects of HE are progressively diminished due to the reduced mobility of hydrogen. Consequently, the occurrence of hydrogen-induced brittle fracture is suppressed, whereas the increase in brittleness is primarily attributed to the intrinsic effects of low temperature.

Figure 14 presents the SEM analysis of specimens tested at −253 °C, under conditions simulating exposure to liquefied hydrogen. At this ultra-low temperature, both WO-H and W-H specimens exhibited fully brittle fracture surfaces. These results experimentally confirm that at cryogenic temperatures near −253 °C, hydrogen activity is negligible, and fracture behavior is governed solely by thermal effects rather than HE. These findings confirm that hydrogen has a pronounced embrittling effect at moderately low temperatures (e.g., −80 °C to −130 °C). In contrast, at ultra-low temperatures (e.g., −160 °C), the embrittling influence of hydrogen is substantially reduced due to limited diffusion, and fracture behavior is governed primarily by temperature-induced effects.

4. Conclusions

This study systematically investigated the effects of hydrogen embrittlement on the fracture toughness of 9% Ni steel and STS 316L stainless steel across a range of cryogenic temperatures, including −253 °C. Quantitative analysis of the hydrogen content using TDS confirmed sufficient hydrogen uptake in the BCC-structured 9% Ni steel to induce embrittlement. In contrast, hydrogen charging was more challenging in the FCC-structured STS 316L. CTOD tests under hydrogen-free (WO-H) and hydrogen-charged (W-H) conditions revealed the following:

(1)

Temperature dependence: Both 9% Ni and STS 316L steels showed a gradual decrease in CTOD values with decreasing temperature under hydrogen-free conditions, consistent with typical reductions in toughness owing to increased atomic binding and reduced ductility at cryogenic temperatures. Nevertheless, ductile fracture behavior was maintained at all temperatures.

(2)

Hydrogen embrittlement behavior: For 9% Ni steel under hydrogen-charged conditions, a significant reduction in fracture toughness was observed at subzero temperatures (−80 °C to −130 °C), where hydrogen mobility is sufficient to promote embrittlement. In contrast, the effect of hydrogen on STS316L specimens was minimal. Owing to the very shallow penetration of hydrogen, only a slight influence was observed, and its effect on fracture toughness was minimal. Therefore, within the experimental methodology and scope of this study, the influence of hydrogen on the fracture toughness of STS316L was determined to be minimal. This reduction was particularly pronounced in the 9% Ni steel because of its BCC crystal structure, which allowed for more efficient hydrogen diffusion.

(3)

Transition temperature for HE-free behavior: For 9% Ni steel, hydrogen influenced fracture toughness at temperatures above −160 °C, but its effect was negligible below this temperature. Thus, −160 °C and lower can be regarded as a practical temperature range in which the influence of hydrogen embrittlement can be disregarded. In the case of STS316L, hydrogen penetration was minimal, resulting in only slight differences in fracture toughness up to −80 °C; below this temperature, the effect of hydrogen was essentially absent.

(4)

Fracture morphology and SEM observations: SEM analysis corroborated the mechanical findings, showing predominantly ductile features at higher temperatures and increasing brittleness with decreasing temperature. The 9% Ni steel without hydrogen charging exhibited brittle morphology over nearly the entire fracture surface. In contrast, STS316L without hydrogen charging displayed limited brittle features localized at the notch tip, while the overall fracture surface remained ductile.

Overall, this study demonstrated that while hydrogen embrittlement poses a significant threat to fracture toughness at moderately low temperatures, its effect diminishes as the temperature decreases because of reduced hydrogen diffusivity. These results provide critical insights into the safe operational limits of hydrogen-exposed structural materials in cryogenic environments, supporting the design of hydrogen infrastructures with enhanced resistance to embrittlement.