Effect of alloying elements on delta ferrite formation temperature

The delta ferrite formation temperature, which is defined as the lowest temperature at which delta ferrite is stable, is an important parameter in 9 wt.% Cr steels. During the manufacturing process, the 9 wt.% Cr steels are typically required to be hot rolled at 1250°C. If the delta ferrite formation temperature of the steel is below the hot rolling temperature, delta ferrite may form during hot rolling and is very difficult to remove by subsequent heat treatment, including normalising and tempering. The presence of a large amount of delta ferrite in the microstructure of the steel is detrimental as delta ferrite grains always adopt an elongated morphology along the rolling direction and can increase the brittleness of the steel.

In this study, the effect of various alloying elements including C, Mn, Al, N and W on the predicted delta ferrite formation temperature has been studied using thermodynamic calculations. Each of the elements was varied independently by ±30% of its original composition in the Tenaris P92 steel. The delta ferrite formation temperatures of each composition variations were then calculated and plotted in Figure 4.9 against the amount of element variation.

77

Figure 4.9: Effect of alloying elements on the predicted delta ferrite formation temperature of Tenaris P92

It can be observed from Figure 4.9 that carbon and tungsten are the two elements which have the most significant impact on the predicted delta ferrite formation temperature. Increasing the carbon content in the alloy can significantly raise the delta ferrite formation temperature, whilst increasing the tungsten can significantly decrease the delta ferrite formation temperature. The effects of these two elements on the delta ferrite formation temperature can be explained by their roles in the phase transformation between delta ferrite and austenite.

Carbon is a strong austenite stabiliser which promotes the formation of austenite and therefore increases the delta ferrite formation temperature whilst tungsten is a strong ferrite stabiliser which stabilises the ferrite phase at low temperature and therefore decreases the delta ferrite formation temperature [9]. Manganese, nitrogen and aluminium can also affect the predicted delta ferrite formation temperature. However, as shown in Figure 4.9, the effects of these elements are not as significant as W and C.

1160 1180 1200 1220 1240 1260 1280

-40% -30% -20% -10% 0% 10% 20% 30% 40%

Delta Ferrite formation T (°C)

Percentage of variance

C Mn Al N W

78 4.2.2 Sensitivity study on MarBN

The effects of some key alloying additions including C, B, N, and W on the phase stabilities in the MarBN steel are studied in this section. In addition, the effect of various alloying elements on the delta ferrite formation temperature of the MarBN steel is investigated at the end of this section.

4.2.2.1 Effect of C

The effect of carbon on the predicted amount of Laves phase and M23C6 at 600°C in MarBN steel is shown in Figure 4.10 (a), whilst the effect of carbon on the predicted M₆C formation is shown in Figure 4.10 (b).

Figure 4.10: (a) Effect of carbon content on the predicted amount of Laves phase and M₂₃C₆ in MarBN steel; (b) The effect of carbon content on the predicted formation of M₆C phase in MarBN steel

The effect of carbon on the predicted amount of the M₂₃C₆ and Laves phase is identical to that in the Tenaris P92 steel. As shown in Figure 4.10 (a), the predicted amount of M₂₃C₆ increases as the carbon content in the alloy increases, whilst the predicted amount of Laves phase decreases as the carbon content increases.

The carbon concentration in the MarBN steel can also affect the formation of M₆C, which is predicted to be a tungsten rich carbide. As shown in Figure 4.10 (b), the temperature range in which M6C phase is predicted to be stable broadens as the carbon content in the steel increases. The M6C phase is predicted only in MarBN steel rather than in the P92 type steels probably due to the high tungsten content in the alloy. The contribution of the M6C phase to the creep resistance of the steel is unknown as no relevant literature has been found to date.

0.0%

Laves at 600C M23C6 at 600C

Mass Percentage -30%C

79 4.2.2.2 Effect of B

Boron is added deliberately into MarBN steel to improve the steel’s creep resistance [27, 28].

According to Abe et al. [74, 77 and 78], boron is able to stabilise M₂₃C₆ from coarsening during creep exposure and therefore extends the creep life of the steel. Thermodynamic calculations carried out in this study predict a small amount of boron in the M₂₃C₆ phase whilst the majority of boron is predicted to present in the M₂B phase, which is a Cr and W rich boride. The effect of boron concentration on the predicted amount of M₂B is shown in Figure 4.11.

Figure 4.11: The effect of boron content on the predicted amount of M₂B phase in MarBN steel.

As shown in Figure 4.11, the predicted amounts of the M2B phase at 600°C, 780°C and 1070°C all increase as the boron content in the steel increases. The experimental observation of a M2B phase in 9 wt.% Cr steel has not been found in the literature, although Horiuchi et.

al. [28] have found some un-identified W rich borides in their research. Therefore, it is considered that the prediction of M2B phase needs to be further verified by experimental results the details of which are discussed further in Chapter 5.

4.2.2.3 Effect of N

The effect of nitrogen on the predicted amount of VN type MX at 780°C is shown in Figure 4.12 (a) and the predicted effect of nitrogen on the Z phase formation is shown in Figure 4.12 (b).

M2B at 600C M2B at 780C M2B at 1070C

Mass Percentage -30% B

-15% B Base 15%B 30% B

80

Figure 4.12: (a) The effect of nitrogen content on the predicted amount of VN type MX in MarBN; (b) The effect of nitrogen content on the predicted amount of Z phase in MarBN

As shown in Figure 4.12 (a) and (b), the increase in nitrogen content in MarBN promotes the formation of VN type MX and Z phase simultaneously, which is identical to the effect of nitrogen in P92 steel. However, due to the boron addition in MarBN, the nitrogen content in MarBN steel should be carefully adjusted together with boron to avoid the formation of boron nitride. The effect of synergistic boron and nitrogen addition on the formation of BN has been studied in detail using thermodynamic calculations and the results are discussed in section 4.3 4.2.2.4 Effect of W

Identical to the findings in P92 steel, tungsten promotes the formation of Laves phase in MarBN steel. As shown in Figure 4.13, the amount of predicted Laves phase at 600°C increases with the tungsten concentration in the steel.

Figure 4.13: The effect of tungsten on the predicted amount of Laves phase at 600°C in MarBN steel 0.00%

81 In addition to its effect on Laves phase formation, the concentration of tungsten in MarBN steel can also affect the M6C formation. As shown in Figure 4.14 (a), both the predicted temperature range and the amount of M₆C phase increase as the tungsten content increases in the alloy. In the composition variation which contains 70% of the original tungsten addition, no M₆C phase was predicted. The promotion of the M₆C phase also has impact on the formation of other phases. The comparison between Figure 4.14 (a) and (b) shows that the amount of predicted NbC type MX phase decreases as the amount of M₆C increases in the temperature range in which M₆C is stable. The simultaneous promotion of the M₆C phase and the suppression of the NbC type MX phase suggest that tungsten makes M₆C more thermodynamically stable and the formation of M₆C consumes the available carbon content in the steel which is needed for NbC formation. Although the data concerning the effect of M6C phase on the creep resistance of the steel are not available, the thermodynamic calculation results indicate that M6C may be a detrimental phase as it suppresses the formation of NbC type MX, which is one of the important strengthening phases in 9 wt.% Cr steel.

Figure 4.14: (a) The effect of tungsten on the formation of M6C phase in MarBN steel, it should be noted that at -30% W, no M₆C phase was predicted; (b) the effect of tungsten on the formation of NbC type MX

in MarBN steel

Mass Percentage of M6C phase

Temperature (°C)

Mass Percentage of NbC rich MX

Temperature (C)

82 4.2.2.5 Effect of alloying elements on the delta ferrite formation temperature

The effect of various alloying elements including C, Mn, Al, B, Co, N and W on the predicted delta ferrite formation temperature of MarBN steel has been studied in a similar way to that in P92 steel. The results are shown in Figure 4.15, which plots the calculated delta ferrite formation temperatures against the percentage of each composition variation.

Figure 4.15: Effect of alloying elements on the predicted delta ferrite formation temperature in MarBN steel

The results shown in Figure 4.15 suggests that Co and C are the two strong austenite stabilisers which are able to raise the delta ferrite formation temperature whilst W is a strong ferrite stabiliser which reduces the delta ferrite formation temperature significantly. In addition, Mn, Al and B have little impact on the predicted delta ferrite formation temperature.

The calculation results highlights one of the alloy design concepts of the MarBN steel, because one of the big differences between a MarBN (3W3Co) and a conventional P92 (2W0Co) is the increased tungsten and cobalt contents in the MarBN steel. Tungsten can provide solid solution strengthening to the steel, however, the increased amount of tungsten is needed to be balanced by the addition of cobalt to avoid delta ferrite formation during manufacturing.

83 4.3 The Effect of Boron and Nitrogen Concentrations on BN Formation

Recent research has indicated that the addition of boron into 9 wt.% Cr steels can significantly increase the creep resistance of the steel [28, 74, 77 and 78]. It is considered that boron can stabilise the M23C6 from coarsening during creep exposure [77, 78]. In the ideal case, the boron stabilising effect can be combined with the nitrogen strengthening effect and result in a much improved creep resistance of 9 wt.% Cr steel. However, the simultaneous addition of boron and nitrogen may result in the formation of BN, which potentially cancels the strengthening effect of both elements. Therefore, the avoidance of BN formation in the 9 wt.% Cr steel by controlled addition of boron and nitrogen is one of the key issues in new steel design.

Sakuraya et al. [81] have studied the formation conditions of BN in a variety of 9-12 wt.% Cr steels and developed a boron-nitrogen solubility product line from experimental data as shown in Figure 4.16.

Figure 4.16: Composition diagram for boron and nitrogen showing the formation conditions for BN in the temperature range of 1050-1150°C in 9-12 wt.% Cr steels [81]

In the current study, the formation conditions for BN have been studied using thermodynamic calculations. In the present work, a large number of calculations were performed using the Tenaris P92 composition as the base composition with the boron and nitrogen concentrations systematically varied at the expense of iron. The calculation results are summarised and plotted in Figure 4.17 in a similar manner to Figure 4.16. In Figure 4.17, each boron and nitrogen variation is plotted in a boron-nitrogen composition diagram; if BN is predicted to

84 be stable in the temperature range 1050-1150°C, a blue diamond was used, if BN is not predicted to present at that temperature range, a red square was used. It can be observed from Figure 4.17 that the numerous data points have revealed a boundary which denotes a composition range in which BN is predicted to be present and also a composition range in which BN is not predicted.

Figure 4.17: Thermodynamically calculated composition diagram for boron and nitrogen showing the formation conditions for BN in the temperature range of 1050-1150°C in the standard Tenaris P92 steel

To simplify the predictions on the formation rules of BN, three arbitrary zones were defined based on the calculation results and are illustrated in Figure 4.18. An arbitrary marginal zone is defined around the boundary between BN predictions and no BN predictions. A conservative BN formation condition is then defined using the boundary between the ‘No BN Zone’ and the ‘Marginal Zone’. It can be observed from Figure 4.18 that if the boron addition is between 0.0001 wt.% and 0.04 wt.%, the nitrogen addition should be kept below 0.016 wt.% to avoid BN formation; if the boron addition is lower than 0.0001 wt.%, the nitrogen addition should be kept below 0.1 wt.% to avoid BN formation. For comparison purposes, the BN solubility product line developed by Sakuraya et al. [81] has been superimposed on Figure 4.18. It can be observed that the BN formation condition defined in this study is generally in agreement with the experimentally developed solubility line. Compared to the condition given by Sakuraya et al. [81], the conditions developed in the present study

1E-05 0.0001 0.001 0.01

0.001 0.01 0.1

Boron (wt.%)

Nitrogen (wt.%)

YES NO

85 indicates that more boron can be added when the N content is about 0.01 wt.% However, it should be noted that the validity of thermodynamic predictions needs to be confirmed by experimental data.

Figure 4.18: Thermodynamically calculated composition diagram for boron and nitrogen showing the formation condition for BN in the temperature range of 1050-1150°C in the standard P92 steel with the

simplified rules of formation illustrated and the BN solubility product [81] superimposed