Abstract
The precipitates and hydrogen permeation behaviour were investigated in high strength low alloy steel produced by thermomechanical controlled processing with air/water cooling after finishing rolling, and the water cooled specimens were further tempered at various temperatures. Two types of precipitates have been found in the specimens. One is TiN with the size ranging from 50 to 300 nm, and the other one is fine NbC. The cooling and tempering treatment conditions affect the precipitation behaviour of NbC particles in α-Fe, leading to the difference in hydrogen permeation. The apparent hydrogen diffusivity in the air cooled specimen is lower than that in the specimen quenched and subsequently tempered at 300°C when the charging current density is 10 mA cm−2. Increasing the tempering temperature to 500°C leads to the decrease of apparent hydrogen diffusivity; but the value is still higher than that in the air cooled specimen. However, the apparent hydrogen diffusivity slightly increases with further increasing tempering temperature from 500 to 650°C.
Introduction
High strength low alloy (HSLA) steels produced by thermomechanical controlled processing (TMCP) have attracted more attention for their high strength, high toughness and good weldability.1,2 Multi-microalloying of Nb and Ti, leading to the formation of small precipitates such as TiN and NbC, influences the mechanical properties of HSLA steels.3 The TMCP with an accelerated cooling process developed as a plate production technology can improve strength and toughness by controlling the final microstructure.4 The precipitation of small particles during the controlled rolling and controlled cooling process also plays an important role in the improvement of mechanical properties of HSLA plates.
It is recognised that hydrogen dissolved in HSLA plates can cause a detrimental effect on mechanical properties.5 This effect is often referred to as hydrogen embrittlement, which is an extremely complex phenomenon. Although several different mechanisms were proposed for hydrogen embrittlement,6,7 it is difficult to establish an effective correlation between hydrogen embrittlement susceptibility and material characteristics. It is believed that the hydrogen embrittlement susceptibility is related to a critical concentration of hydrogen causing subsequent embrittlement. Traps in steels are very important for resistance to hydrogen embrittlement by means of decreasing hydrogen diffusivity and increasing critical hydrogen concentration.8 Dislocations, interfaces and precipitates are generally recognised as trapping sites. These traps can be classified as reversible and irreversible ones according to their binding energy for hydrogen atoms.9,10 Precipitates such as TiC and NbC dispersed in steels are usually considered as effective traps to decrease hydrogen diffusivity and increase critical hydrogen concentration.11 Thus these types of traps improve the hydrogen embrittlement resistance of HSLA steels.12
For HSLA plates produced by TMCP, a controlled cooling process after hot rolling inevitably affects the precipitation behaviour of carbides. However, the hydrogen permeation behaviour of the HSLA plates produced by TMCP with an accelerated cooling process is not clear. So far, an electrochemical permeation technique has been successfully applied to measurement of hydrogen diffusivity.9,13 Therefore, the purpose of this paper is to investigate the hydrogen permeation behaviour by the electrochemical method as well as the precipitates by transmission electron microscopy in the HSLA steel prepared by controlled rolling and controlled cooling.
Experimental
The steel investigated was prepared by vacuum melting and its chemical compositions are: 0·035C, 0·20Si, 1·50Mn, 0·02Ti, 0·05Nb, 0·004N, ⩽0·003S, ⩽0·01P (wt-%). The ingot of 150 kg was heated >1200°C and then forged into billets >950°C. The billets were reheated to ∼1200°C and subsequently controlled rolled into plates. The finish rolling temperature is >850°C. After finish rolling, the plates were air cooled or quenched to room temperature. The quenched plates were tempered at 300, 500 or 650°C for 2 h respectively.
Both carbon extraction replicas and thin foils were prepared for examinations of precipitates. The precipitates were observed using an H-800 transmission electron microscope (TEM) operating at 200 kV and a JEM-2010 high resolution transmission electron microscope (HRTEM) operating at 200 kV.
Hydrogen permeation experiments were carried out by using the electrochemical method developed by Devanathan and Stachurski.9 Disc shaped specimens were cut from the plates and mechanically ground to 1 mm in thickness. The specimens were then electrolytically polished in a solution of perchloric acid, washed several times in distillated water and degreased with acetone. Both sides of the specimens were electroplated with Ni. The thickness of the plated Ni layer was ∼0·01 mm. Each specimen was used as a membrane separating an electrochemical cell into two halves. In the left half, hydrogen was introduced by cathodic charging on the left side of the membrane. The charging process was performed by using a galvanostatic method. Hydrogen coming through the right side of the membrane was ionised at a potential of 200 mV(Hg/HgCl); and the output flux of hydrogen was measured by monitoring the current in the right half of the electrochemical cell. Both cathodic and anodic solutions were 0·2 N NaOH, which were deaerated with N2 to reduce the background current.11
All hydrogen permeation curves were measured at 20°C for the disc shaped specimens. Half-time method may be used to calculate diffusion coefficient properly, but the test takes a very long time and it is difficult to reach a permeation balance state. In the present work, therefore, the apparent hydrogen diffusivity D app was calculated from each hydrogen permeation curve by the time lag method9,14
Results and discussion
With TEM observations on the carbon extraction replicas, all specimens were found mainly containing two types of precipitates. One is the cubic shaped precipitates with the size ranging from 50 to 300 nm, showing the typical morphology character of TiN. Figure 1a shows the TEM image obtained from the carbon extraction replica of the air cooled specimen. Figure 1b is the selected area electron diffraction pattern from the TiN particle marked by an arrow in Fig. 1a. It can be seen that the TiN particle is much larger than the other type of precipitates in an irregular spherical shape. It is well known that most particles dissolve to the austenite solid solution but TiN particles still remain when the reheating temperature is elevated to 1200°C. Thus the large TiN particles probably formed in the reheating process.
a TEM image obtained from carbon extraction replica showing precipitates in air cooled sample and b SAD pattern from TiN particle marked by arrow in a
Figure 2a exhibits the HRTEM image obtained from a thin foil of the air cooled specimen, showing a particle in the irregular spherical shape. Figure 2b presents the inverse fast Fourier transform image of the selected area in Fig. 2a. It was found that the particle has a face centred cubic structure and the lattice parameter a = 0·4434 nm, which is close to the reported value a = 0·4468 nm for NbC.15 Thus this type of precipitates is believed to be NbC. The slight decrease in lattice parameter may be caused by complex precipitation of (Nb1−xTix)(C1−yNy) or formation of nonstoichiometric NbCx.16
a HRTEM image obtained from thin foil showing NbC particle precipitated in matrix of air cooled sample and b inverse fast Fourier transform image of selected area in a
NbC particles usually precipitate during the controlled rolling process. However, it is noteworthy that the solubility of Nb in γ-Fe is larger than that in α-Fe.17 This means that NbC particles could further precipitate in α-Fe during the cooling process. Thus the precipitation of NbC in α-Fe can be suppressed by quenching after finish rolling. During the tempering of the quenched specimens at a moderate temperature, NbC particles would reprecipitate in α-Fe. Therefore, the cooling condition should influence the precipitation behaviour of NbC in α-Fe.
Figure 3 compares the TEM images obtained from the thin foils of the air cooled specimen and the specimens quenched and subsequently tempered at 300, 500 and 650°C. It can be seen that a large amount of fine NbC particles precipitate in the ferrite matrix, grain boundaries and on dislocations in the air cooled specimen; and the large majority have a size <20 nm, as shown in Fig. 3a. However, the specimen quenched and subsequently tempered at 300°C has a smaller quantity of NbC particles than the air cooled specimen (Fig. 3b). This confirms that the precipitation of NbC particles for the quenched steel could not occur in the tempering process at 300°C. It is interesting that NbC particles could further precipitate during the tempering processes at 500 and 650°C. Compared with the particle size of NbC in the air cooled specimen, those in the specimens quenched and subsequently tempered at 500 and 650°C are slightly larger (∼20 nm), as shown in Fig. 3c and d.
Images (TEM) obtained from thin foils showing NbC particles precipitated in samples a air cooled, b quenched and tempered at 300°C, c quenched and tempered at 500°C and d quenched and tempered at 650°C
The hydrogen permeation curves at different charging current densities are present in Fig. 4. It can be seen that hydrogen diffusion is accelerated with increasing charging current density from 1 to 10 mA cm−2. Furthermore, the hydrogen diffusion behaviour in distinct specimens is also found to be different. From these curves, the lag times of hydrogen permeations were obtained and the apparent hydrogen diffusivities D app were calculated by using equation (1).
Hydrogen permeation curves measured at 20°C for samples a air cooled, b quenched and tempered at 300°C, c quenched and tempered at 500°C and d quenched and tempered at 650°C
Figure 5 shows the change in apparent hydrogen diffusivity with increasing charging current density for each specimen. It can be seen that the apparent hydrogen diffusivity gradually rises to a steady value with increasing the charging current density up to 10 mA cm−2. This result is related to traps in steels because hydrogen diffusion is strongly impeded by traps. Ideal irreversible traps are often saturable at a low lattice hydrogen concentration so that the apparent hydrogen diffusivity increases with hydrogen concentration once the irreversible traps are filled.18 For a lattice only containing irreversible traps, the apparent hydrogen diffusivity D app increases to ideal lattice diffusivity D L with increasing charging current density until the irreversible traps are saturated. However, reversibly trapped hydrogen is in dynamic equilibrium with dissolved hydrogen in lattice so that the apparent hydrogen diffusivity D app is always less than D L.18 In the present case, a large amount of fine NbC particles work as strong irreversible traps. Thus the apparent hydrogen diffusivity D app increases to a steady value, indicating the irreversible traps reach saturation quickly when the charging current density increases up to 10 mA cm−2. Even so, the steady value should be less than ideal lattice diffusivity D L because each specimen also contains reversible traps besides the irreversible ones.
Apparent hydrogen diffusivity versus charging current density measured at 20°C for samples a air cooled, b quenched and tempered at 300°C, c quenched and tempered at 500°C and d quenched and tempered at 650°C
Moreover, the apparent hydrogen diffusivities measured at a charging current density of 10 mA cm−2 for distinct specimens are quite different, as shown in Fig. 5. The value for the specimen quenched and tempered at 300°C is larger than those for others. These results seem to be associated with NbC precipitates. It is generally accepted that the incoherent NbC precipitates with a high bonding energy act as irreversible traps which can increase total hydrogen concentration in steel.11 If the irreversible traps are saturated by hydrogen, they no longer interact with the hydrogen atoms dissolved in lattice and affect the diffusion of hydrogen.18 Thus the difference in the apparent hydrogen diffusivity measured at 10 mA cm−2 for the distinct specimens should be caused by reversible traps. Usually the reversible traps with low bonding energies include dislocations, boundaries, substitutional elements, etc. However, the above results indicate that the difference in apparent hydrogen diffusivity is also related to NbC precipitates. It was reported that both reversible and irreversible traps can be produced by the precipitation of the coherent TiC, VC and NbC.19 Although the exact fractions of the coherent and incoherent NbC particles could not be determined, the quantity of the coherent NbC particles might be responsible for the different apparent hydrogen diffusivities mentioned above. Probably, the coherent NbC particles can precipitate to some extent in α-Fe during the air cooling process. On the contrary, the incoherent NbC particles easily precipitate in the quenched specimens during the subsequently tempering process if the tempering temperature is enough high. When the quenched steel was tempered at 300°C, both the coherent and the incoherent NbC particles were much less. Although NbC particles could sufficiently precipitate in α-Fe when the quenched steel was tempered at 500 or 650°C, the fraction of the coherent NbC particles might be less than that in air cooled specimen. Therefore, the air cooled specimen has the lowest apparent hydrogen diffusivity while the specimen quenched and tempered at 300°C shows the highest value. Similarly, increasing tempering temperature from 500 to 650°C might lead to the further decrease of the fraction of coherent NbC particles, causing the slight increase of apparent hydrogen diffusivity.
Conclusions
In the investigated HSLA steel produced by TMCP there are two types of precipitates. One is TiN with the size ranging from 50 to 300 nm, and the other one is fine NbC. The quantity of NbC precipitates in the air cooled specimen is more than that in the specimen quenched and subsequently tempered at 300°C. Although NbC particles can sufficiently precipitate when the quenched steel was tempered at 500 and 650°C, the size of the NbC particles is slightly larger than that in the air cooled specimen. Such a precipitation behaviour influences the hydrogen permeation. The apparent hydrogen diffusivity for each specimen gradually rises to a steady value with increasing the charging current density up to 10 mA cm−2. The apparent hydrogen diffusivity at 10 mA cm−2 for the air cooled specimen is lower than that for the specimen quenched and tempered at 300°C. Increasing the tempering temperature to 500°C leads to the decrease of apparent hydrogen diffusivity, but the value is still higher than that in the air cooled specimen. However, the apparent hydrogen diffusivity slightly increases with further increasing tempering temperature from 500 to 650°C. The experimental results imply that the tempering operation is favourable for improving the hydrogen embrittlement resistance of the direct quenched steel; but for an optimised effect, the tempering process parameters should be carefully determined.