Search for articles:
Anwen Zhang
Corresponding authors:
Received: 2018-09-27
Revised: 2018-10-23
Accepted: 2019-02-7
Online: 2019-07-20
Copyright: 2019 Editorial board of Journal of Materials Science & Technology Copyright reserved, Editorial board of Journal of Materials Science & Technology
More
Abstract
Effect of segregation behaviors of P at different cooling rates on the precipitation of γ″ and γ′ phases and the corresponding strength are investigated. The precipitation of γ″ and γ′ phases during cooling is sensitive to P concentration. With increasing the concentration of P, the amount of γ″ and γ′ particles increases after air cooling. With decreasing the cooling rate, the accelerating effect of P on the precipitation of γ″ and γ′ phases decreased first and then increased, which demonstrates the concentration of P dissolved in the grain interior decreases first and then increases. The different effects of P on γ″ and γ′ phases with different cooling rates were analyzed by the kinetic characteristic of nonequilibrium grain-boundary segregation. The characteristic of nonequilibrium grain-boundary segregation of P in superalloy is further confirmed, and the phenomenon caused by critical cooling rate is captured.
Keywords：
Although phosphorus had been taken as a common deleterious element in steels and superalloys for a long time period [1], it has been documented that appropriate quantities of P can significantly prolong the creep and stress rupture lives of some wrought superalloys including IN718 alloy [[2], [3], [4], [5]]. Essentially, P is a trace element with strong segregation tendency and tend to be segregated at grain boundaries [6,7]. It has been deemed that P has no effect on the grain interior [5,8,9], and the beneficial effect of P on superalloys has been attributed to its effect on grain boundaries [2,10]. However, the most recent studies revealed that P accelerated the precipitation of γ″ and γ′ phases in superalloys during air cooling [11,[11], [12], [13]]. Therefore, P has significant effects on both grain boundaries and grain interiors, and studying the distribution and segregation behavior of P in superalloys is important for the proper application of P on improving superalloys.
The segregation of impurities to grain boundaries occurs by two mechanisms, viz., equilibrium grain-boundary segregation (EGS) and nonequilibrium grain-boundary segregation (NGS) [14,15]. The NGS has two typical characteristics, viz., the phenomenon of critical time during annealing and the phenomenon of critical cooling rate during cooling [16]. The P content on the grain boundary can reach peak value in the two critical states. Many studies captured the phenomenon of critical time of P in steels during annealing, which confirmed the NGS behavior of P in steels [17,[16], [17], [18], [19]]. However, the investigation focused on the segregation behavior of P in superalloys is far fewer. Zheng et al. revealed that the segregation behavior of P in IN718 alloy did not match the laws of EGS during annealing, which indicated that the NGS behavior of P occurred in IN718 alloy [6]. Unfortunately, the segregation behaviors of P during cooling in both steels and superalloys are difficult to be captured, and the segregation behaviors of P during cooling have not been revealed by experiments up to now.
The effect of cooling rate on the segregation behavior of P in IN718 alloy was investigated in this study, and the effects of the segregation behavior of P on the precipitation of γ″ and γ′ phases and the strength of the alloy were revealed.
An IN718 master ingot was prepared via vacuum induction melting (VIM), the chemical composition (wt%) of the master ingot was: 0.043 C, 19.15 Cr, 52.3 Ni, 3.08 Mo, 0.48 Al, 0.90 Ti, 5.30 Nb, 0.005 B and balance Fe. The master alloy was cut and remelted. No P was intentionally doped in the normal IN718 alloy, but the previous study showed that the optimal content of P was about 0.022 (wt%) in the IN718 alloy [20]. Therefore, the alloys with low P content (no P was intentionally doped), optical P content and high P content were designed. The P contents of the three IN718 alloys in this study are 0.001, 0.023 and 0.034 wt% for Alloys 1, 2 and 3, respectively. The three ingots were homogenized by 1120 ℃ × 20 h + 1160 ℃ × 20 h + 1190 ℃ × 50 h AC and then forged at 1100 ℃.
To investigate the effects of P on the precipitation of γ″ and γ′ phases during air cooling in IN718 alloy, the alloys with various P contents were treated by 1120 ℃ × 3 h AC. To make the cooling rates of the samples same, the samples with various P contents were taken out of the heat treatment furnace together. To exclude the effect of grain growth on the segregation behavior of P in IN718 alloy, some IN718 specimens were treated by 1120 ℃ × 1 h + 1190 ℃ × 3 h WQ firstly. Annealing at 1120 ℃ aims to dissolve the phases with low melting points, and annealing at 1190 ℃ aims to make the grain fully grow. The specimens were treated by 1120 ℃ × 3 h AC subsequently.
Alloys with different P contents were machined into three geometries (Fig. 1). The cubic specimens were machined into three different sizes, 20 mm × 20 mm × 20 mm, 30 mm × 30 mm × 30 mm, 40 mm × 40 mm × 40 mm. To investigate the effect of cooling rate on the strengthening effect of P, the micro-hardness was measured at various positions of the different specimens. The micro-hardness testing performed on the specimens with the geometry A (Fig. 1) was carried out on the upper surfaces; the micro-hardness testing performed on the round bars with the geometry B was carried out along the central axis (Fig. 1, indicated by dotted line); the micro-hardness testing performed on the cubic specimens with the geometry C was carried out along the central axis of the cubic specimens (Fig. 1, indicated by dotted line). The grain-interior Vickers micro-hardness testing was carried out at room temperature under the loading condition of 200 g for 15 s. The micro-hardness values were an average of at least 5 measurements per sample.
Fig. 1. Specifications of the alloys for micro-hardness test.
The standard tensile specimens were machined and tested at room temperature. The gage diameter is 5 mm, and gage length is 25 mm. The microstructures were examined by optical microscopy (OM), scanning electron microscopy (SEM), and transmission electron microscopy (TEM). The grain sizes were quantified by OM and the IAS8 mage analysis software.
A PHI-700 NanoAES was utilized to evaluate the concentration of P on the grain boundary. The parameters for the Auger measurement were as follows: Gun Voltage = 10 kV, Energy Resolution = 1‰, Vacuum < 3.9 × 10-9 Torr, Beam Diameter = 20 nm. The specimens were sputtered by Ar + for 1 min before the quantitative analysis.
As shown in Fig. 2, the grains of IN718 alloys grew rapidly during annealing at 1120 ℃. After annealing at 1120 ℃ for 20 min, the grains stopped growing, and the average grain size was about 120 μm. Both grain growth rates and final grain sizes of alloys 1, 2 and 3 are the same.
Fig. 2. Grain sizes of the alloys with different P contents annealed at 1120 ℃ for various time.
The specimens with geometry A in Fig. 1 were treated by 1120 ℃ × 3 h AC. As shown in Fig. 3, the micro-hardness of the grain interior increases with the increasing P content. The TEM results reveal that no γ″ and γ′ phases precipitate in the matrix of alloy 1. However, γ″ and γ′ phases precipitate in the matrix of alloys 2 and 3, and the amount of the particles in alloy 3 is the most. It indicates that P can accelerate the precipitation of γ″ and γ′ phases and strengthen the matrix during air cooling, and the accelerating effect enhances with the increasing P content. In other words, the micro-hardness and the corresponding γ″ and γ′ phases in the matrix after air cooling are sensitive to the P content.
Fig. 3. Micro-hardness, γ″ and γ′ phases and corresponding [
The round bars with geometry B in Fig. 1 were treated by 1120 ℃ × 3 h AC, and the tensile testing was performed subsequently. Alloy 2 is obviously harder than alloy 1 (Fig. 3), but the yield strengths of the two alloys are almost the same (Fig. 4). However, the yield-strength value of the alloy 3 is double that of alloys 1 and 2. The ultimate-strength change with the increasing P content is the same with the yield-strength change. The results in Fig. 2 indicate that the grain sizes of the three alloys treated by 1120 ℃ × 3 h AC are the same, so the different tensile-strength values of the alloys are irrelevant to the grain size.
Fig. 4. Tensile strengths of the IN718 alloys treated by 1120 ℃ × 3 h AC.
As a precipitation strengthening alloy strengthened by γ″ and γ′ phases [21,22], both the matrix micro-hardness and the tensile strength of IN718 alloy are predominated by γ″ and γ′ phases. Therefore, the tensile strength change and the micro-hardness change should be consistent. However, the results in this work reveal an abnormal phenomenon: the tensile strength change (Fig. 4) is inconsistent with the micro-hardness change (Fig. 3) with the increasing content of P.
The previous study [12] indicated that P decreased the formation enthalpies of γ′ and γ′′ phases when it occupied the Ni lattice sites in the two phases and accelerated the precipitation of the two phases. The quantities of the precipitates increase with the increasing P content (Fig. 3) because more P atoms in the matrix prove more sites for nucleation.
The micro-hardness testing was performed on the surfaces of the small samples with the geometry A in Fig. 1, and the tensile testing characterized the strengths of the gage sections of the standard tensile specimens. Although all the specimens were treated by 1120 ℃/3 h AC, the cooling rates varied with the sizes. The different effects of P on micro-hardness and yield strength may result from different cooling rates. To confirm this speculation, the micro-hardness testing was performed along the central axis of the round bars with the geometry B (indicated by dotted line, Fig. 1). With the increasing distance from the end of the bar along the axis, the cooling rate decreases.
As shown in Fig. 5, the micro-hardness at various positions in alloy 1 is almost the same, and no γ″ and γ′ phases precipitate in the matrix. It indicates that the effect of cooling rates on the precipitation of γ″ and γ′ phases in alloy 1 can be ignored. The micro-hardness at the end of the bar of alloy 2 is higher than that of alloy 1, and γ″ and γ′ phases precipitate in the matrix at this position due to the accelerating effect of P on the two phases during air cooling. However, increasing the distance from the end of the bar, the micro-hardness of alloy 2 decreases rapidly and becomes the same with that of alloy 1, and the TEM results show that no γ″ and γ′ phases precipitate in the matrix of alloy 2. It indicates that the accelerating effect of P on the precipitation of γ″ and γ′ phases in alloy 2 disappears when the cooling rate is relatively low. The micro-hardness at the end of the bar of alloy 3 is higher than that of alloy 2, and much more γ″ and γ′ phases precipitate in the matrix, which consists with the results in Fig. 3. Increasing the distance from the end of the bar, the micro-hardness of alloy 3 decreases rapidly and then stays stable, but it is always higher than that of alloys 2 and 1, and the γ″ and γ′ phases are always observed in the matrix. The tensile test characterizes the strength of the gage section of the specimen. Therefore, the strength of alloy 2 is the same with that of alloy 1, and the strength of alloy 3 is much higher than that of alloy 1. As a consequence, the tensile-strength change (Fig. 4) is inconsistent with the micro-hardness change (Fig. 3), which results from the different cooling rates of the specimens with different sizes.
Fig. 5. Micro-hardness and corresponding γ″ and γ′ phases at the positions with various distances from the ends of bars along the axial.
The NGS of P in IN718 alloy has been confirmed by the reductio ad absurdum [6]. Based on the theory of NGS, the solute-vacancy complexes can diffuse to the grain boundaries during cooling from high temperature [[23], [24], [25]]. Therefore, the effects of P on the precipitation of γ″ and γ′ phases depend on the P content dissolved in the matrix when the specimens are cooled to the precipitation temperatures of γ″ and γ′ phases. The cooling rate at the end of bar is relatively fast during cooling, and the cooling time is insufficient for all P-vacancy being segregated to the grain boundaries. The remanent P in the grain interior accelerate the precipitation of γ″ and γ′ phases and strengthen the matrix. With slower cooling rate, the time is sufficient for much P-vacancy complexes being segregated to grain boundaries, so the micro-hardness of alloys 2 and 3 at the positions more than 1 mm from the end of the bar decreases markedly (Fig. 5). However, the P content in alloy 3 is high, and the concentration of P on the grain boundary is high or reaches saturation, which makes the driving force of NGS low. Some P atoms are always left in the grain interior of alloy 3. As a result, the precipitation of γ″ and γ′ phases is accelerated, and the matrix is strengthened.
To further confirm the above discussions, the P concentrations on grain boundaries at different positions should be characterized and compared. The P content of alloy 3 is the highest, and it is most likely to be measured effectively by auger electron spectroscopy (AES). Therefore, the grain-boundary P contents of alloy 3 at various positions of the round bar were measured by AES. As shown in Fig. 6, the grain-boundary P content at 30 mm distance from the end of bar is higher than that at the end of the bar, which strongly verifies the above discussions. In addition, the beam diameter is larger than the width of grain boundary, so the accurate concentration of P on the grain boundary should be much higher than the measured value.
Fig. 6. AES spectrum of the positions with various distances from the end of the W3 bar along the axial: (a) the P peak of AES spectrums of the positions with various distances from the end of the bar; (b) the average P contents of the positions with various distances from the end of the bar.
According to the NGS theory [26], the critical cooling rate can lead to the peak of P content on the grain boundary during cooling, and the P content on grain boundary should increase first and then decrease with decreasing cooling rate. Unfortunately, the segregation behavior of P during cooling in both steels and superalloys has not been revealed by experiments, and the phenomenon of critical cooling rate during cooling has not been captured up to now. According to the NGS theory [26], the critical cooling rate can lead to the peak of P content on grain boundary, and the grain-boundary P content increases first and then decreases with decreasing cooling rate. Therefore, the matrix micro-hardness should decrease first and then increase. To capture the critical cooling rate of P and confirm the discussion in Section 4.2, the specimens were cooled with slower rate. Cubic specimens of alloy 2 with various sizes were treated by 1120 ℃ × 1 h + 1190 ℃ × 3 h WQ firstly to exclude the effect of grain growth on the segregation behavior of P. Then the specimens were treated by 1120 ℃ × 3 h AC. The grain-interior micro-hardness testing was performed along the dotted line in Fig. 7(a). As shown in Fig. 7, the micro-hardness of side surfaces is high, while the micro-hardness decreases markedly and then stays invariable with the decreasing cooling rate. The TEM results indicate that a great amount of γ″ and γ′ phases precipitate in the matrix on the side surface of the specimen, while no γ″ and γ′ phases precipitate in the matrix of the center. The micro-hardness curves consist with that of alloy 2 in Fig. 5, but still the critical cooling rate is not captured.
Fig. 7. Micro-hardness and γ″ and γ′ phases at the positions with various distances from the sides of the cubes along the axial, alloy 2: (a) 20 mm × 20 mm × 20 mm; (b) 30 mm × 30 mm × 30 mm; (c) 40 mm × 40 mm × 40 mm.
The vacancy concentration is high when the sample is soaked at high temperature, which makes the high driving force of NGS at the same cooling rate [27,28]. In this case, the de-segregation of P from the grain boundary is difficult. Therefore, the soaking temperature decreased subsequently to capture the critical cooling rate and confirm the discussion in Section 4.2. Other cubic specimens of alloy 2 were treated by 1120 ℃ × 1 h + 1190 ℃ × 3 h WQ firstly and then treated by 1050 ℃ × 3 h AC. The micro-hardness test was carried out along the dotted line in Fig. 8(a). The micro-hardness curve of the 20 mm × 20 mm × 20 mm cube increases slightly and then decreases to an invariable value. However, the micro-hardness curves of the 30 mm × 30 mm × 30 mm and 40 mm × 40 mm × 40 mm cubes present increasing-decreasing-increasing trend (Fig. 8(b) and (c)). The TEM results indicate that the micro-hardness change results from the various precipitation of γ″ and γ′ phases. The cooling rates of the positions at a distance of 7.5 mm from the side surface of 30 mm × 30 mm × 30 mm cube and 5 mm from the side surface of 40 mm × 40 mm × 40 mm cube are the critical cooling rates. The cooling rate slows as the size enlarge, so the critical place of the 40 mm × 40 mm × 40 mm cube is closer to the side surface.
Fig. 8. Micro-hardness and γ″ and γ′ phases at the positions with various distances from the sides of the cubes along the axial, alloy 2: (a) 20 mm × 20 mm × 20 mm; (b) 30 mm × 30 mm × 30 mm; (c) 40 mm × 40 mm × 40 mm.
What’s more, the results in Fig. 8 indicate that the micro-hardness at the positions near the side surfaces of the cubes increases slightly with the decreasing cooling rate. Although decreasing the cooling rate can reduce the P content in the grain interior, it can also provide more time for the precipitation of γ″ and γ′ phases. Therefore, the micro-hardness increases slightly at the positions near the side surfaces of the cubes. The results in Fig. 7 reveal that the micro-hardness on the side surface of the cube increases with the increasing of the size, which further proves the above discussion.
The γ″ and γ′ phases in the IN718 alloy without P addition can also precipitate when the stay time at the precipitation temperature of γ″ and γ′ phases is long enough. It indicates that the γ″ and γ′ phases can precipitate when the cooling rate is slow enough, which makes us wonder the high micro-hardness near the center of the cube in Fig. 8 attributes to the slow cooling rate or the effect of P. As shown in Fig. 9, the micro-hardness of alloy 1 is always low in every state, which indicates that the γ″ and γ′ phases cannot precipitate in the matrix in every state. So the high micro-hardness near the center of the cube in Fig. 8 is attributed to the effect of P.
Fig. 9. Micro-hardness of the positions with various distances from the sides of the cube along the axial, alloy 1 pre-treated by 1120 ℃/1 h + 1190 ℃/3 h WQ： (a) 30 mm × 30 mm × 30 mm, 1120 ℃/3 h AC; (b) 40 mm × 40 mm × 40 mm, 1120 ℃/3 h AC; (c) 30 mm × 30 mm × 30 mm, 1050 ℃/3 h AC; (d) 40 mm × 40 mm × 40 mm, 1050 ℃/3 h AC.
(1)The precipitation of γ″ and γ′ phases during air cooling is sensitive to the P content in the grain interior of IN718 alloy. The amount of γ″ and γ′ phases increases with the increasing P content, so the matrix is enhanced gradually.
(2)With the decreasing cooling rate, the accelerating effect of P on the precipitation of γ″ and γ′ phases diminishes first and then enhances, which demonstrates the concentration of P dissolved in the grain interior decreases first and then increases.
(3)The characteristic of nonequilibrium grain-boundary segregation of P in superalloy is further confirmed, and the phenomenon caused by critical cooling rate is captured.
The authors have declared that no competing interests exist.
WeChat
/
| 〈 |
|
〉 |
