Materials Physics
Effect of Nb on Thermal-Shock Resistance of Austenitic Heat Resistant Cast Steel
2020 Volume 61 Issue 9 Pages 1711-1716
Details
2020 Volume 61 Issue 9 Pages 1711-1716
The effect of Nb on microstructure and thermal-shock resistance was investigated for as cast and annealed JIS SCH21 steels. Primary NbC carbide were crystallized together with M23C6 at grain boundary in as cast steel. Decomposition of primary M23C6 was detected, while NbC was also hardly decomposed when annealed at elevated temperature. It was found that chromium carbides had low thermal stability but primary NbC carbides played a role in preventing propagation of micro-cracks and restraining shape deformation under condition both heating and quenching. As a result, thermal-shock resistance was improved in the Nb-added steel.
This Paper was Originally Published in Japanese in J. Japan Inst. Met. Mater. 83 (2019) 474–478.
Fig. 12 Partitioning of alloying elements analyzed by STEM-EDS for micro-crack in Nb steel occurred when thermal shocked at 1223 K for 300 cycles.
Conventional austenitic heat resistant cast steels such as JIS SCH13, SCH21 and SCH13 are used to make heat treatment trays, jigs for loading heat-treated automotive parts and furnace components.1,2) When heating and rapid cooling (water cooling or oil cooling) are repeated for long time under the severe heat-treatment environment with air atmosphere or carburizing, the product life is remarkably reduced by the initiation of the cracks and outward deformation.3)
It is generally known that the addition of nickel (Ni) is effective to improve heat resistance.4) However, Ni is relatively expensive alloying element and its price fluctuation is so large. Addition of a large amount of Ni could not be the better choice for improving high-temperature mechanical properties from a viewpoint of production cost. Therefore, most of studies concerning about improvement of heat resistance such as creep property by adding a few alloying elements of Nb, Ti, Mo and so on instead of Ni have been actively conducted.5–7)
On the other hand, with regard to heat treatment jigs, it is necessary to optimize the chemical compositions of cast steels to attempt a longer product life or maintain a fixed lifetime in even thinner-walled jigs with lighter weight.
Apart from such chemical compositions and structures, there is a problem that cracks and fractures occur on the surface in even defect-free products under the thermal shock environment where heating and cooling cycles are repeated. However, there is little report about cracking factors, which is regarded as the problem on practical use.
As described above, several studies about effects of alloying elements on creep property, oxidation resistance and carburization resistance of austenitic cast steels have been reported in recent years.8,9) However, there hardly is the investigation about the thermal shock resistance of the steels which becomes factors of crack and fracture occurrence. Especially mechanism of crack generation and deformation and suppression effect under thermal shock environment have not been clarified.
Therefore, in this study, focusing on Nb, one of effective alloying elements for improving the high-temperature strength, we investigated the effect of Nb addition on thermal shock resistance of austenitic heat resistant cast steel by performing thermal shock test assuming the actual use environment.
High-temperature mechanical properties of heat-resistant cast steel are strongly influenced by ferrite (α phase) and sigma phase (σ phase), which cause high-temperature embrittlement.7) Presence of these phases mainly depends on the chemical composition of steel. Therefore, investigation of high-temperature properties requires alloy design which is sufficiently taken into account high-temperature embrittlement. In this study, alloy design was performed using conventional thermodynamic calculation software Thermo-Calc (database: TCFE5), and the carbon content was fixed at about 0.32 mass% to prevent high-temperature embrittlement. As a result, based on chemical compositions of JIS SCH21, two JIS G0307a experimental steels were prepared using a high frequency induction furnace: 1) Fe–20Ni–23Cr–0.33C–0.02Nb (hereafter referred to as base steel) and 2) the base steel with adding 1.25%Nb. With regard to the casting conditions of the Y block, the molten steel was heated to 1913 ± 20 K. After holding for 0.6 ks, the molten steels were tapped lightly at 1873 K into the ladle and then poured into sand molds of Y block. Thermal shock test specimens were mechanically machined from the round bars which were cut out from the bottom of Y block of both steels in as-cast state. Table 1 shows the main chemical compositions of experimental steels.
Figure 1 shows a schematic diagram of the thermal shock machine and test specimen. The outer diameter D, inner diameter and thickness of the test specimen are 25 mmφ, 14 mmφ and 7 mm, respectively. The test specimens were subjected to heating at 1223 K for 0.6 ks and water cooling for 0.12 ks as one cycle. Total cycles were 300 times. The outer diameter deformation rate of the test specimen every 20 cycles was defined and calculated by the following equation.
| \begin{align} &\text{Outer diameter deformation rate (%)}\\ &\quad = |\text{D}_{\text{w300}} - \text{D}_{\text{h300}}|/\text{D}_{\text{w0}} \times 100 \end{align} | (1) |
Experimental apparatus of thermal shock equipment (a) and dimensions of samples (b).
Observation of micro-cracks generated by the thermal shock test in each sample was performed using an optical microscope, scanning electron microscope (ELIONIX, ERA8900FE) and scanning transmission electron microscope (STEM, JEOL, JEM-ARM200CF). The STEM sample was prepared using a dual beam FIB (FEI, Quanta 3D 200i), which can perform sampling by limiting the location of micro-cracks generation. In addition, in order to quantitatively evaluate micro-cracks, the average area (measured at 41 places) and average length (measured at 18 places) of the cracks were obtained using image process method. The distribution of each alloying element was analyzed using energy dispersive X-ray spectroscopy (EDS, EDAX, AMETEK). Furthermore, in order to reproduce the microstructural change during thermal shock test, microstructural analysis of the samples which were subjected to annealing treatment at 1223 K for 180, 360 and 720 ks in air atmosphere was similarly performed. In addition, X-ray diffractometry (XRD, PANalytical, Empyrean, target: Co) was used to identify carbides.
On the other hand, regarding with evaluation of physical properties, the measurement of thermal expansion coefficient was performed using a thermomechanical analyzer (SHIMADZU, TMA-60H). 3 mmφ × 10 mm test specimen was heat-treated under a load of 0.098 N in air atmosphere at a temperature range of 373 to 1273 K with a heating rate of 0.167 K/sec. The average thermal expansion coefficient was calculated every 100 K based on the obtained displacement curve. Measurement of thermal conductivity was performed from room temperature to 1073 K by the laser flash method.
Figure 2 shows the XRD results of the base steel and the Nb steel in as-cast state and after annealing at 1223 K for 180, 360 and 720 ks. Remarkable diffraction peaks attributed to austenite (γ phase) are observed, whereas the diffraction peak intensity considered to be carbide is extremely weak in all samples. Two kinds of diffraction peaks of carbide indicated by ○ and △ marks were revealed to be M23C6 and NbC, respectively. It was found that only M23C6 crystallized in the base steel, whereas M23C6 and NbC coexisted in the Nb steel. It is possible to think that the strength of NbC peak intensity is high compared with M23C6 and NbC preferentially crystallizes by Nb addition when aiming at the peak intensity ratio of M23C6 and NbC carbide in the Nb steel here. No diffraction peak of the α phase and the σ phase adversely influencing the high-temperature properties were detected except confirming carbides which existed in the austenite matrix of both steels.
X-ray diffraction spectra obtained from base and Nb steels in as-cast and annealed at 1223 K for various times.
Microstructures of the base steel and the Nb steel in as-cast state (hereafter, as cast) are shown in Fig. 3. Microstructures of both steels consisted of austenitic matrix and massive M23C6, and net-like NbC was detected in the Nb steel. These are typical microstructures of heat-resistant cast steels.
Optical micrographs of as cast steels.
Figure 4 shows the results of thermal shock test. The second horizontal axis indicates the total high-temperature holding time at each cycle number. Under thermal shock environment where heating and cooling are repeated, there is little change in thickness of the specimens, but deformation of outer diameter occurred in both steels. Here, the total high-temperature holding time of a specimen by 300 cycles of thermal shock test is equal to 180 ks of annealing time. The outer diameter deformation rate after 300 cycles of thermal shock test was calculated from eq. (1). As a result, the deformation rates in base steel and Nb steel were 2.5% and 1.5%, respectively. After 300 cycles of thermal shock tests, micro-cracks occurred on surface of specimens in both steels. The micro-cracks observed on the surface of specimen after the thermal shock test are illustrated in Fig. 5. Many of the micro-cracks that occurred in the base steel were larger than the Nb steel. Table 2 shows the average length and average area of micro-cracks observed in both steels. The micro-cracks that occur due to thermal shock test decrease not only average length but also average area of micro-cracks by adding Nb. This result exhibits a good correlation with the measured deformation rate of the outer diameter through the thermal shock test, and it is considered that the deformation rate caused by thermal shock depends on the degree of development of micro-crack.
Deformation rate measured through thermal shock test from 293 K to 1223 K.
SEM micrographs showing micro-cracks in base and Nb steels occurred during thermal shock test at 1223 K for 300 cycles.
Generally, under the environment where heating and cooling are repeated, the steels suffer deformation by expansion and shrinking or thermal stress due to a temperature gradient. Therefore, thermal expansion coefficient and thermal conductivity are considered to be the important factors.4) For details, the smaller coefficient of thermal expansion lowers the deformation by expansion and shrinking. The larger thermal conductivity or the smaller temperature gradient decrease the generated thermal stress. Figure 6 and Fig. 7 are the measurement results of thermal expansion coefficient and thermal conductivity, respectively. The thermal expansion coefficients of both steels increase with increasing temperature. It is clear from Fig. 6 that the addition of Nb is effective to the decrease of the thermal expansion coefficient, and this result is consistent with the report of Ueda et al.2) Furthermore, it is considered from Fig. 7 that addition of Nb contributes to increase of the thermal conductivity. As described above, addition of Nb is also effective in lowering the coefficient of thermal expansion. Therefore, it expects to be effective in the improvement of thermal shock resistance.
Temperature dependency of thermal expansion coefficient.
Temperature dependency of thermal conductivity.
Figure 8 shows the microstructure of micro-cracks on the surface of specimen after thermal shock test. Cr carbides were observed in the vicinity of micro-cracks that occurred in both steels by the thermal shock test. This carbide is identified as primary M23C6 based on the results of the quantitative analysis performed by SEM-EDS and XRD. Therefore, it is predicted that the micro-cracks that occur by the thermal shock are caused by primary M23C6 which was detected in as-cast state.
SEM-EDS micrographs showing micro-crack in base and Nb steels occurred when thermal shocked at 1223 K for 300 cycles.
Microstructures of the base steel and the Nb steel after as-cast and annealing are shown in Fig. 9 and Fig. 10, respectively. Decrease of Cr and C concentration in the center of M23C6 was confirmed as shown in Table 3, when base steel was annealed at 1223 K for 180 ks, which are equal to the heating temperature and the heating time of 300 cycles of thermal shock test. Figure 11 illustrates the line analysis on the cross section of the carbide shown in Fig. 9 in order to see the results of Table 3 more conspicuously. Similarly to the results of the point analysis, decrease of Cr and C concentrations at the center of primary Cr carbide is recognized from the results of the line analysis. Takahashi et al.10) reported that the decomposition of Cr carbide near the surface was attributed to the loss of equilibrium state between the matrix and the precipitated phase (Cr carbide) due to decrease of Cr and C concentration. It is predicted that the decomposed M23C6 became initiation site of micro-cracks when repeating thermal expansion and shrinking from the difference of the thermal expansion coefficient by the difference between the Cr and C concentrations of center part and outer part. On the other hand, as shown in Fig. 10, morphology of the primary NbC observed in the as-cast Nb steel is almost in the same form after annealing, so it is clear that the stability of NbC is thermodynamically higher compared with M23C6. Several µm of fine particles dispersed in the matrix in the both annealed steels are presumed to be secondary carbides M23C6. However, secondary carbide NbC could not be detected.
SEM-EDS micrographs of base steel in as cast state and 1223 K × 180 ks annealed.
SEM-EDS micrographs of Nb steel in as cast state and 1223 K × 180 ks annealed.
EDS line profile of base steel annealed at 1223 K × 180 ks. Note line position in SEM image indicates the analysis position.
Figure 12 shows the STEM image in the vicinity of micro-cracks in the Nb steel after the thermal shock test. In both steels, most of micro-cracks exist in the center of primary M23C6, and cracks are developed by connecting of some of them. Primary NbC which hindered development of cracks was observed as shown in Fig. 12. On the other hand, M23C6 which prevent development of cracks was not recognized. Therefore, in the Nb steel, the net-like NbC which stably existed even under the thermal shock environment prevents temporarily the development of the micro-cracks. As a result, it is presumed that coalescence of micro-cracks is suppressed by NbC in the Nb steel. Ueda et al.2) reported that NbC preferentially crystallized by increasing the amount of Nb addition, and as a result volume of the primary M23C6 decreased. In this study, it was also found that Nb addition reduces volume of M23C6 as well as the initiation site of crack.
Partitioning of alloying elements analyzed by STEM-EDS for micro-crack in Nb steel occurred when thermal shocked at 1223 K for 300 cycles.
From the above results, micro-cracks generated under the thermal shock condition simulating the industrial use environment of jigs for heat treatment is found to be attributed to the decomposition of primary carbide. Addition of Nb is effective for improving the thermal shock resistance.
The relationship between thermal shock resistance and microstructural change was investigated by comparing thermal shock behavior in conventional JIS SCH21 austenitic cast steel and 1.25 mass%Nb added one. The obtained results are as follows.
