Effect of heat treatment on hydrogen permeation behaviour of AISI 4135 steel under splash zone conditions

Materials and specimen

The material used in this paper was an AISI 4135 high strength steel. The chemical compositions are shown in Table 1, and the heat treatment processes are shown in Table 2. The A, B, C and D shown in Table 2 are then used to indicate the heat treatment conditions throughout the manuscript. For the hydrogen permeation tests, circular plate specimens 40 mm in diameter and 0.5 mm in thickness were used. Both surfaces of the specimens were polished with emery paper up to # 600. One side of the specimens was plated with Ni in Watt's bath (NiSO₄·6H₂O 250 g L^{− 1}, NiCl₂·6H₂O 45 g L^{− 1}, H₃BO₃ 40 g L^{− 1}) at room temperature with a current of 3 mA cm^{− 2} for 3 min. ²⁰ The estimated thickness of the Ni plating was ∼180 nm. For the measurement of pH beneath rust layer, specimens of 10 mm in diameter and 20 mm in height were prepared.

OPEN IN VIEWER

Table 1

Chemical compositions of AISI 4135 steel

Element	C	Si	Mn	P	S	Cr	Mo	Ni	Fe
wt-%	0.399	0.293	0.509	0.015	0.014	0.903	0.204	0.08	Balance

OPEN IN VIEWER

Table 2

Heat treatment conditions of AISI 4135 steel specimens

Specimen	Heat treatment condition
A	860°C thermal insulation 50 min, 370°C isothermal quenching 30 min, air cooling
B	860°C thermal insulation 50 min, oil quenching, 200°C tempering 1 h, air cooling
C	860°C thermal insulation 50 min, oil quenching, 550°C tempering 1 h, air cooling
D	1120°C thermal insulation 20 min, 860°C thermal insulation 10 min, oil quenching, 200°C tempering 1 h, air cooling

Simulation of splash zone conditions

The schematic diagram of the indoor apparatus for the simulation of splash zone conditions is shown in Fig. 1. The electrolytic cell and spray chamber were made by organic glass. The wetting conditions of specimens were simulated by spraying sea water on the surface. The tidal changes and the severity of splash were simulated by adjusting the period of spraying and spraying frequency.

OPEN IN VIEWER

1

Schematic of indoor apparatus for simulated splash zone conditions (C: counter electrode, R: reference electrode, W: working electrode)

Hydrogen permeation tests

Hydrogen permeation tests were performed under simulated splash zone conditions as shown in Fig. 1 at room temperature. The double cell that is similar to the Devanathan–Stachurski's cell was used. ²¹ Two spraying intervals of 1 and 30 min were employed to simulate two different severities of sea water splash in this study. As illustrated in Fig. 1, an Hg/HgO electrode and a Pt wire were used as reference and counter electrodes for the hydrogen detection cell. The cell was filled with 0.2M NaOH solution, and the Ni plated specimen was polarised using a potentiostat at 0.0 mV versus Hg/HgO electrode for >24 h until the residual current < 100 nA cm^{− 2} was reached. Then, the hydrogen permeation current was recorded as the specimen corroded under simulated splash zone conditions.

pH and corrosion potential measurement

The pH beneath the rust layer of AISI 4135 steel specimens was monitored using a W/WO₃ electrode ²² prepared by electrooxidation in 2.0M H₂SO₄ solution between 0.76 and 2.0 V. Saturated calomel electrode was used as reference electrode. The W/WO₃ electrode was calibrated using standard pH buffer solutions. Sea water droplet was put on the specimen surface at room temperature simultaneously whenever to start the pH and the corrosion potential measurements.

Electrochemical impedance spectroscopy

The experiment was performed in a typical three-electrode system. A saturated calomel electrode and a Pt electrode were used as reference and counter electrodes. Electrochemical impedance spectroscopy tests were measured with specimens under sea water film at room temperature. The electrochemical impedance spectroscopy was performed in a frequency range from 10⁵ to 10^{− 2} Hz. Parallel experiments were measured to verify the reproducibility, and the most typical results were reported. The number of tests was denoted as specimen size in the section on ‘Results and discussion’.

Steel microstructure

Figure 2 illustrates the microstructure for steel subjected to different heat treatments. It can be seen that microstructure of specimen A is mainly bainite and retained austentite (Fig. 2a); the main structure of specimen B is short lath martensite (Fig. 2b); the microstructure of specimen C is tempered martensite, some ferrite and pearlite (Fig. 2c); the specimen D is coarse lath martensite with prior austenite grains being visualised (Fig. 2d).

OPEN IN VIEWER

2

Metallographic observation on heat treated specimens

Hydrogen permeation currents under simulated splash zone conditions

The hydrogen permeation currents recorded under spraying interval of 1 and 30 min are shown in Figs. 3 and 4. The hydrogen currents have the same variation trend under different spraying interval. Some minor decrease at the beginning of the spray was also observed. As the tide lowered, which was during the dry process, the hydrogen permeation current decreased with time after a peak value was reached. The decrease in hydrogen permeation current at the beginning of the spray is mainly due to the pH increase after the spray, since the pH of fresh sea water is ∼8. When the specimen surface was wetted by sea water, the corrosion reaction occurred. The main cathodic reaction corresponds to the anodic dissolution of Fe (equation (1)) is oxygen reduction reaction (equation (2)). At the same time, the hydrogen reduction reaction also takes place partially (equations (3) and (4)), and the adsorbed hydrogen on the surface permeated into the metal (equation (5)). Therefore, the hydrogen permeation current can be recorded with the progress of corrosion process. During the dry stage after sea water spray, the rust formed with the continuing corrosion and the evaporation of sea water, concentrations of H⁺ and chloride ion (equations (6) and (7)) would increase and then enhance hydrogen entry before the rust layer is totally dried up.^11,23,24 The reactions taking place on the rusted metal surface can be described as follows (1) (2) (3) (4) (5) (6) (7) (8) (9) If we integrate the area below each curve shown in Figs. 3 and 4, the amount of hydrogen permeated through specimens can be obtained. The calculated values of different cycles are shown in Tables 3 and 4. More clearly, the sequence of total amount of hydrogen permeated is A < D < C < B.

OPEN IN VIEWER

3

Hydrogen permeation currents under spraying interval of 1 min

OPEN IN VIEWER

4

Hydrogen permeation currents under spraying interval of 30 min

OPEN IN VIEWER

Table 3

Amount of hydrogen permeated under spraying interval of 1 min

Number of cycles	Amount of hydrogen permeated (H+/mol)
	A	B	C	D
1	1.706 × 10− 7	4.015 × 10− 7	4.181 × 10− 7	2.117 × 10− 7
2	2.003 × 10− 7	4.164 × 10− 7	4.131 × 10− 7	2.495 × 10− 7
3	2.636 × 10− 7	4.255 × 10− 7	4.167 × 10− 7	3.293 × 10− 7
4	3.013 × 10− 7	4.786 × 10− 7	4.976 × 10− 7	3.750 × 10− 7
5	3.678 × 10− 7	7.042 × 10− 7	5.446 × 10− 7	4.044 × 10− 7
6	4.472 × 10− 7	7.792 × 10− 7	7.716 × 10− 7	4.187 × 10− 7
7	4.012 × 10− 7	6.277 × 10− 7	6.773 × 10− 7	3.871 × 10− 7
8	3.584 × 10− 7	5.666 × 10− 7	5.755 × 10− 7	3.724 × 10− 7
Total	2.511 × 10− 6	4.400 × 10− 6	4.315 × 10− 6	2.748 × 10− 6

OPEN IN VIEWER

Table 4

Amount of hydrogen permeated under spraying interval of 30 min

Number of cycles	Amount of hydrogen permeated (H+/mol)
	A	B	C	D
1	1.664 × 10− 7	3.216 × 10− 7	2.015 × 10− 7	2.091 × 10− 7
2	1.845 × 10− 7	4.010 × 10− 7	3.161 × 10− 7	2.353 × 10− 7
3	2.315 × 10− 7	4.085 × 10− 7	3.647 × 10− 7	2.512 × 10− 7
4	2.810 × 10− 7	4.200 × 10− 7	4.653 × 10− 7	3.004 × 10− 7
5	3.265 × 10− 7	4.506 × 10− 7	4.396 × 10− 7	3.246 × 10− 7
6	3.738 × 10− 7	5.408 × 10− 7	5.147 × 10− 7	3.947 × 10− 7
7	3.434 × 10− 7	5.357 × 10− 7	4.108 × 10− 7	2.529 × 10− 7
8	2.939 × 10− 7	5.302 × 10− 7	3.155 × 10− 7	2.604 × 10− 7
Total	2.201 × 10− 6	3.609 × 10− 6	3.028 × 10− 6	2.229 × 10− 6

As shown in Tables 3 and 4, the amount of hydrogen permeated increased with the cycle, and then gradually decreased after showing a maximum in sixth cycle under simulation of splash zone conditions. It is attributed to the rust formation on the steel surface. In early cycles, the corrosion products can provide transmission channels for moisture and oxygen, which is the main reason for the decrease in pH (equations (8) and (9)). The H⁺ reduction reaction was accelerated by a decrease in pH at the metal/corrosion products interface. ⁸ After six cycles, the accumulation of corrosion products prevented hydrogen entry due to the low permeability of hydrogen through the rust layer. The lower diffusivity of hydrogen in rust layer and the lower coverage of hydrogen on the metal surface after massive corrosion product formation can be the reason of lower hydrogen entry after prolonged spray cycles.

In addition, the amount of hydrogen permeated under spraying interval of 1 min is higher than specimen under spraying interval of 30 min. This is presumably because of the increase in water spray time when shortening the time of spray interval. The concentration of dissolved oxygen and chloride ion can be high. The Fe²⁺ can be further oxidised (equation (8)). Moreover, chloride ions can migrate through the rust layer to the metal/rust interface and increase the water content in the rust. ¹¹ These factors can further lower the pH value and increase the hydrogen entry.

Moreover, the repeated experiments with different heat treated specimens were performed under same conditions to confirm the reproducibility. As shown in Figs. 5 and 6, the amount of hydrogen permeated is A < D < C < B.

OPEN IN VIEWER

5

Amount of hydrogen permeated under spraying interval of 1 min

OPEN IN VIEWER

6

Amount of hydrogen permeated under spraying interval of 30 min

pH and corrosion potential measurements

In order to further understand the hydrogen origin during splash corrosion process and the factors that affect the hydrogen permeation, the pH beneath rust layer and the corrosion potential of the specimen surface were measured. Figure 7 shows the change of measured pH value and corrosion potential when the specimen surface was wetted by dropping 0.5 mL of sea water. The pH value and corrosion potential decreased with time. The difference in initial corrosion potentials is small among the four specimens with different heat treatments. Their corrosion potential shifted toward negative with time and reached a steady value after 1 h. The hydrogen evolution action occurred easily under these conditions refer the potential pH diagram of Fe–H₂O. ²⁵ The pH decrease accelerates H⁺ reduction and results in hydrogen permeation current increase as shown in Figs. 2 and 3. In addition, the rust layer is moistened by sea water. The hydrolysis reaction of Fe²⁺ leads to lowering pH value ⁸ and increasing the efficiency of hydrogen permeation (equations (8) and (9)).

OPEN IN VIEWER

7

pH value and corrosion potential variation after wetting specimen surface with 0.5 mL of sea water at room temperature

Rust layer analysis

The rust layer of different heat treated specimens was observed by SEM, and the micrographs are shown in Fig. 8. The whole surface of each specimen was covered with rust layer after eight cycles of experiment. These corrosion products are mainly in the form of irregular particles. The diameter of the largest particle is ∼10 μm. Pores and cracks exist among these corrosion products.

OPEN IN VIEWER

8

Micrographs of rust layer observed by SEM

Corrosion products of different heat treated specimens that had been corroded for 8 cycles were analysed by XRD. Figure 9 shows that the rust layers are composed of lepidocrocite (γ-FeOOH), goethite (α-FeOOH) and hematite (α-Fe₂O₃). From the XRD pattern and SEM, the corrosion products have no apparent differences. That is to say, the selected heat treatments have no prominent impact on the form of corrosion products.

OPEN IN VIEWER

9

X-ray diffraction pattern of rust products (G: goethite, L: lepidocrocite, H: hematite)

Electrochemical impedance spectroscopy

The impedance measurements obtained after 15 min of exposure under natural sea water electrolyte film were shown in Fig. 10. The four specimens with different microstructures exhibit similar behaviour. A fitting process was performed using equivalent circuit analogies. The equivalent circuit that was the typical Randles one as illustrated in Fig. 11, where R _s is solution resistance, R _ct is the charge transfer resistance and CPE is a constant phase element. The fitted results of R _ct are shown in Table 5.

OPEN IN VIEWER

10

Electrochemical impedance spectroscopy plots with specimens under sea water film at room temperature

OPEN IN VIEWER

11

Equivalent circuits of fitting the impedance results

OPEN IN VIEWER

Table 5

Charge transfer resistance obtained by equivalent circuit fitting

Specimen	R ct
	A	B	C	D
1	1973	1560	1492	2022
2	2131	…	2183	1952
3	…	2320	1868	1904
4	2309	1604	1892	2481
5	1598	2354	…	2442
6	2560	…	1782	2247
7	1875	1409	1893	1779
8	…	1639	1887	1822
9	1910	1491	2458	2452
10	…	1567	…	1553
11	2017	1531	2436	2457
12	2416	1725	2035	…
13	2042	…	1688	2025
14	2226	1912	2100	2329
15	2104	1536	2293	…

As shown in Table 5, due to the scattering of impedance data, mathematical statistics is used to compare the differences in corrosion behaviour among the specimens. The analysis shows that the average value of charge transfer resistance for heat treatments A, B, C and D have no significant effects on corrosion behaviour under nature sea water film. In addition, Yu et al. ²⁶ had investigated the corrosion behaviour of heat treated AISI 4135 steel by means of potentiodynamic polarisation experiments and found that the heat treatment processes have no profound effect on the average open circuit potential and corrosion current density at same temperature.

From above analysis, it can be deduced that the corrosion behaviour of different heat treated specimens has no significant contribution to the differences in hydrogen permeation behaviour. Moreover, different heat treatments also have no prominent impact on corrosion products. Therefore, the differences in hydrogen permeation behaviours were mainly attributed to the differences in microstructures of the specimens. It is well known that the specimen microstructure has a significant effect on the resistance to hydrogen induced cracking. ²⁷ A study ²⁸ has shown that the hydrogen diffusion coefficient of quenched specimen is higher than that of quenched and tempered specimen. A previous study ²⁹ also has shown that the corrosion rate of the steel with a martensitic microstructure was higher than the steel with bainitic microstructure. Moreover, the steel with bainite shows the lowest values of the hydrogen permeation rate and effective diffusivity. ³⁰ Consequently, the hydrogen permeated through specimen A with bainitic microstructure should be lower than other specimens. The fine microstructures with tempered martensite steels and residual austenite grains are able to distribute hydrogen over spread irreversible traps and reduce the hydrogen diffusion coefficient. ³¹ So, the specimen B with martensitic microstructure shows the highest amount of hydrogen permeated through specimen. In addition, the separation of carbide can decrease the content of alloy element and reduce the corrosion resistance. Therefore, the amount of hydrogen permeated through specimen C is higher than that of specimen D. This was consistent with the experiment results, which were obtained by hydrogen permeation tests.