Transformations and Microstructures
Simultaneous Optimization of Rigidity and Strength of Super Invar Cast Steel using by Martensitic Reversion
2022 年 62 巻 3 号 p. 586-592
詳細
2022 年 62 巻 3 号 p. 586-592
Super invar cast steel, Fe–32%Ni–5%Co by mass%, with excellent low coefficient of thermal expansion has disadvantages in the both of Young’s modulus and strength, because of coarse columnar solidification structure having <100> texture. To simultaneously overcome these disadvantages, the variations of microstructure and mechanical properties through the novel heat treatment referred to as cryo-annealing, which is consisting of subzero treatment and subsequent annealing, were investigated in a super invar cast steel. The cryo-annealing promoted fcc-bcc martensitic transformation and then bcc-fcc martensitic reversion. The bidirectional martensitic transformations led to the formation of duplex austenitic structure consisting of untransformed and reversed austenite with a coarse-grained structure similar to solidification structure. Furthermore, it is found that the austenitic structure was varied depending on the annealing temperature of the cryo-annealing; reversed austenite was remained at lower annealing temperature, while it recrystallized to fine-grained structure as increasing annealing temperature. The high-density dislocations in reversed austenite and the randomized orientation of recrystallized austenite contributed to the development of strength and Young’s modulus, respectively. Therefore, the simultaneous development of rigidity and strength is not achieved by single cryo-annealing, but can be achieved by two-cycle cryo-annealing. Increasing the first annealing temperature and lowering the second annealing temperature in the two-cycle cryo-annealing are appropriate to randomize crystal orientation through austenite recrystallization and to make volume fraction of reversed austenite higher, respectively. As a result, Young’s modulus and 0.2% strength were simultaneously optimized.
Super invar, a nickel-iron alloy with the composition Fe-32%Ni-5%Co (mass%), exhibits low coefficient of thermal expansion. Owing to its unique features, it has been extensively applied to precision equipment components where sophisticated dimensional tolerance is required. Recent studies have attempted to further increase the size and shape complexity of precision equipment frameworks and the casting of super invar alloys with a near-net shaping technique has become indispensable. However, a <100> texture is developed on the coarse columnar crystal structure of the super invar alloy1) during the casting process which severely compromises the Young’s modulus, thus this low rigidity is one of problem for structural material of large frameworks. The authors proposed a cryo-annealing method2) which consists of subzero treatment and subsequent annealing treatment to improve the rigidity of super invar cast alloys without any plastic deformation.
Owing to its chemical composition, super invar cast alloy exhibits a relatively high martensite transformation start temperature (Ms). Consequently, a large amount of martensite undergoes a transformation (fcc-bcc martensitic transformation) during the subzero treatment. In the subsequent annealing, the transformed martensite reversely transforms to austenite (bcc-fcc martensitic reversion). Therefore, the austenite generated by the cryo-annealing treatment contains high-density dislocations, stemming from the bidirectional martensitic transformations. Additionally, it is confirmed that the austenite is recrystallized by these high-density dislocations.3) In other words, it is possible to recrystallize the austenite in super invar cast alloys via heat treatment, and simultaneously eliminate the <100> texture and improve the Young’s modulus.
Furthermore, super invar cast alloys are also utilized as molding dies of carbon fiber-reinforced plastics which require not only Young’s modulus but also tensile strength. Previous studies on improvement of the mechanical properties of invar alloys reported the dislocation strengthening by plastic deformation4,5) and the particle dispersion strengthening.6,7) However, plastic deformation cannot be applied to cast steel products, and increase of alloy elements to stimulate particle dispersion strengthening can deteriorate the coefficient of thermal expansion.6,7,8) In contrast, the cryo-annealing treatment mentioned above does not alter the basic composition of super invar alloys and can strengthen its microstructure without plastic deformation. Therefore, this method is prospected to facilitate the generation of higher-strength austenite. In particular, it is hypothesized that inducing high-density dislocations in reversed austenite via bidirectional martensitic transformations will enhance the dislocation strengthening. Additionally, remarkable austenite recrystallization will promote grain refinement strengthening.
Herein, we investigate the effects of cryo-annealing treatment on super invar cast alloys and discuss the variations in Young’s modulus and mechanical properties, along with the relationship between these variations and the changes in microstructure. Furthermore, the optimum heat treatment method for the simultaneous improvement of the rigidity and strength of super invar cast alloys was explored.
The chemical composition of the super invar alloy samples used in this study is listed in Table 1. To produce the ingots, the alloy was initially melted in an air atmosphere using a high-frequency induction heating furnace, followed by pouring in sand casting molds with dimensions 400 × 400 × 150 mm3 (Fig. 1). To prevent the occurrence of casting defects such as shrinkage cavities in the slab, a riser of φ 240 mm × 240 mm was installed in the center of the slab (Fig. 1(a)). Dilatometer testing (DIL-402C, NETZECH) using a cylindrical piece (φ 6.0 mm × 25 mm) taken from the cast ingots revealed that the start temperature of the fcc-bcc martensitic transformation, Ms, was 224 K and the start and finish temperatures of the bcc-fcc martensitic reversion, As and Af, were 741 K and 860 K, respectively. Analyses of the microstructure, tensile strength, and Young’s modulus of the alloy were performed on a 25 mm wide area in the center of the cast ingots (Fig. 1(b)). A plate piece (Fig. 1(c)) with a parallel volume of 2.0 × 10 × 25 mm3 was used for the tensile tests and it was confirmed that five or more crystal grains were present on the thick plate surface.
| C | Si | Mn | P | S | Ni | Co | Fe | |
|---|---|---|---|---|---|---|---|---|
| □400×150T | 0.017 | 0.10 | 0.23 | 0.005 | 0.003 | 31.88 | 5.14 | Bal. |
Schematic illustration showing the position and the shape of tensile test piece cut from a cast material.
Next, the samples were subjected to subzero treatment at 77 K in liquid nitrogen for 3.6 ks. Subsequently, the subzero-treated materials were annealed at various temperatures ranging from 873 K to 1473 K for 7.2 ks and then quenched in water. The sequence of the subzero treatment and subsequent annealing treatment was dubbed as cryo-annealing treatment. Some samples were subjected to double cryo-annealing, which comprised two consecutive cycles of cryo-annealing as discussed later in the manuscript (Fig. 2). The annealing temperature for the first and second cryo-annealing treatment cycles were in the range T1 = 873 K–1473 K and T2 = 873 K–1023 K, respectively.
Heat treatment route of cryo-annealing process (C. A.) applied for super invar cast steel.
The microstructure of each sample was observed by optical microscopy (DP20, OLYMPUS). To prepare the samples, the surface parallel to the cutting plane (Fig. 1) was mechanically ground with emery papers up to #1200 grit and subsequently polished with diamond abrasives. The polished mirror surface was etched at ambient temperature with a marble solution, Cu2SO4·5H2O:HCl:H2O = 10 g:50 mL:50 mL. The average austenite grain size was measured using the planimetric method, according to the methodology described in JIS G 0551. The volume fraction of the martensite was evaluated using digital image analysis. Tensile testing was conducted using an Instron testing machine (AG-X, SHIMADZU) at an initial strain rate of 1.7 × 10−3 s−1. The Young’s modulus was evaluated based on the resonance frequency of the 7.0 × 16 × 125 mm3 sample using an oscilloscope (ONO SOKKI). The thermal expansion of the φ 6.0 mm × 25 mm cylindrical piece in the 273 K–343 K temperature range was measured using a dilatometer. Coefficients of thermal expansion in the temperature range of 291 K–301 K were used in the following discussion.
Figure 3 shows the optical micrographs of the as-cast, subzero-treated, and annealed (at various temperatures) materials. The as-cast material exhibited a coarse solidification structure, with only a few observable grain boundaries which are indicated by the black arrows. The subzero-treated material exhibited a duplex microstructure comprising needle-like lenticular martensite (α’) with a volume fraction of 87% and untransformed austenite (γ unt). The lenticular martensite promoted by subzero treatment, was preferentially formed along the dendrite with low Ni concentration, and its distribution was independent of the austenite grain structure.2) When the duplex microstructure was annealed above the Af (860 K), the lenticular martensite reversely transformed to austenite; however, the microstructure of the resulting austenite varied depending on the annealing temperature. The 873 K annealed material (Fig. 3(c)) demonstrated coarse austenite grains, indicated by the black arrows, similar to the as-cast material (Fig. 3(a)). In addition, a needle-like substructure, similar to that of the lenticular martensite (so-called as ghost images),9,10,11) was observed in the austenite grains. These microstructural characteristics indicates that reversely transformed austenite grains having high-density dislocation and the same orientation as untransformed austenite were formed during the sequence of non-thermoelastic martensitic transformation and martensitic reversion. Therefore, when annealing treatment was performed just above the Af, a coarse austenite structure was formed comprising reversed austenite (γ rev) and untransformed austenite. In the 973 K annealed material (Fig. 3(d)), it was discovered that austenite recrystallization occurred driven by the difference in dislocation densities between the untransformed and reversed austenite (indicated by the white arrows).3) In the 1103 K annealed material (Fig. 3(e)), the “ghost images” (generated by the reverse austenite transformation) disappeared completely, and fine recrystallized austenite (γ rex) was observed throughout the microstructure. As the annealing temperature further increased, an equiaxed austenite structure was formed due to grain growth (Fig. 3(f)).
Optical micrographs showing microstructural evolution during cryo-annealing process. (a) as-cast, (b) subzero-treated and cryo-annealed materials, which were annealed at (c) 873 K, (d) 973 K, (e) 1103 K and (f) 1473 K. (Online version in color.)
According to these microstructural observations, the start and finish temperatures of recrystallization, Rs and Rf, were determined to be approximately 950 K and 1050 K, respectively. In the following sections, the 873 K, 1103 K, and 1473 K annealed materials will be referred to as the reversed, fine-recrystallized, and coarse-recrystallized materials, respectively, and their characteristics will be discussed.
3.1.2. Influence of Austenitic Microstructure on Material PropertiesFigure 4 illustrates the nominal strain-stress curves of the as-cast, subzero-treated, reversed, fine-recrystallized and coarse-recrystallized materials. The subzero-treated material containing lenticular martensite was characterized by extremely high strength and low ductility, compared with the other austenitic single-phase materials. However, the nominal stress-strain curves of the four materials with austenite single-phase structures demonstrated substantial differences. The reversed material possessed a dramatically increased tensile strength, exceeding 500 MPa, thus confirming that the reversed austenite formation substantially strengthened the austenitic steel alloy. One of the authors reported12) that the remarkably increased strength of the reversed material was induced by the high-density dislocations in the reversed austenite. Furthermore, it was verified that the strength increase depended on the volume fraction and connectivity of the reversed austenite. However, the fine-recrystallized and coarse-recrystallized materials presented a significant decrease in terms of strength during heterogeneous deformation, suggesting a remarkable reduction of area. Both recrystallized materials exhibited significantly higher strengths than the as-cast material, suggesting the notably grain refinement strengthening that was promoted by recrystallization.
Nominal strain-stress curve of as-cast, subzero-treated and cryo-annealed materials. Cryo-annealed materials were annealed at T1 = 873 K, 1103 K and 1473 K, reversed, fine-recrystallized and coarse-recrystallized materials, respectively.
Figure 5 summarizes (c) the 0.2% proof stress results after tensile testing, (a) the coefficient of thermal expansion, and (b) the Young’s modulus as a function of the annealing temperature for the super invar cast steel during cryo-annealing treatment. The as-cast and subzero-treated materials are also shown for comparison. The Af, Rs and Rf temperatures corresponding to the microstructural change are illustrated by dotted lines. All the remaining samples, except for the subzero-treated material containing bcc lenticular martensite, exhibited severely lower values (below 5.0 × 10−7 K−1) of coefficient of thermal expansion. That is, the austenitic single-phase structure maintained the low thermal expansion characteristics of the super invar alloy. The Young’s modulus was relatively low just above the Af, then significantly increased between the Rs and Rf temperatures, and finally saturated above approximately 1050 K. This result corresponded to the randomization of <100> texture, which was developed through solidification, by recrystallization.2) In contrast to the change in Young’s modulus, the 0.2% proof stress maintained high values up to 950 K, followed by a monotonous decrease as the annealing temperature increased. This decrease in strength indicated that the amount of grain refinement strengthening caused by the formation of the fine-grain structure due to recrystallization was smaller than the amount of dislocation strengthening in the reversed austenite. In other words, it can be understood that the continuous decrease in strength was attributed to the disappearance of the reversed austenite and the subsequent grain-growth of the recrystallized austenite.
Changes in (a) coefficient of thermal expansion (CTE), (b) Young’s modulus and (c) 0.2% proof stress of super invar cast steel as a function of annealing temperature of cryo-annealing process.
From these results, it was demonstrated that the high-strength reversed materials with a high reversed austenite content still exhibited a low Young’s modulus as a result of the innate texture. On the other hand, the recrystallized materials with improved Young’s modulus stemming from the randomized crystal orientation did not achieve a significant increase in strength. It was thus concluded that the simultaneous improvement of the Young’s modulus and strength could not be achieved via a simple cryo-annealing process. Therefore, in the following sections, we will discuss the effectiveness of performing the cryo-annealing treatment twice with the aim of simultaneously improving both properties.
3.2. Variation of Microstructural and Mechanical Properties of Super Invar Cast Steel Caused by Double Cryo-annealing 3.2.1. Effect of Austenitic Microstructure on the Martensite Volume FractionFigure 6 shows the optical microstructures and the volume fractions of the martensite (VM) of the second subzero-treated for (a) reversed, (b) fine-recrystallized, and (c) coarse-recrystallized materials with austenite single phase by cryo-annealing. The VM were measured by microstructural observations over a wider area and it was confirmed to be consistent with the phase identification analysis via X-ray diffraction. When the as-cast material was subzero-treated, the lenticular martensite preferentially formed along the dendrite with low Ni concentration, and the VM was 87% (Fig. 3(b)). The reversed material exhibited the same behavior upon subzero treatment; however, the VM was as small as 47%. The reversed austenite occurred by the first cryo-annealing (γ rev1) were often observed around the newly formed martensite in the reversed material. Krauss9) and Imai et al.13) investigated the transformation behavior of the Fe-(30.5–33.5)%Ni-0.005%C and Fe-(27.6–30.7)%Ni alloys after multiple sequences of martensitic transformation and subsequent martensitic reversion. According to their results, as the number of transformation sequences increased, both the Ms and volume fraction of martensite continuously decreased. It was also revealed that the reversed austenite transformed by martensitic reversion exhibited high thermal stability. In other words, it can be understood that the second martensitic transformation was significantly suppressed, stemming from the distribution of the reversed austenite along the dendrite in the reversed material. On the other hand, the preferential distribution of martensite along the dendrites in the fine-recrystallized and coarse-recrystallized materials was not clearly confirmed (Figs. 6(b), 6(c)). However, compared with the fine-crystallized material (Fig. 6(b)), the transformed martensite in the coarse-recrystallized material (Fig. 6(c)) was coarser and exhibited a higher VM value. This phenomenon suggested that the austenite grain size after recrystallization affected the subsequent martensitic transformation behavior.
Optical microstructure of (a) reversed (T1 = 873 K), (b) fine recrystallized (T1 = 1103 K) and (c) coarse recrystallized (T1 = 1473 K) materials after subzero treatment at 77 K for 3.6 ks. (Online version in color.)
Figure 7 illustrates the effects of the austenitic grain size on the (a) Ms and (b) VM values by the subzero treatment of the three materials shown in Fig. 6, as well as the as-cast material. Owing to the austenite stabilizing effect discussed above, the reversed material that mainly comprised reversed austenite demonstrated not only a decrease in VM but also a lower Ms temperature (by approximately 50 K) than the as-cast material. On the other hand, a comparison of the fine-recrystallized, coarse-recrystallized, and as-cast materials without reversed austenite, revealed that the Ms temperature continuously decreased along with the austenite grain refinement. Umemoto and Owen14) investigated the relationship between the austenite grain size and the Ms temperature using a Fe-31%Ni-0.28%C alloy with an average grain size ranging from 20 to 450 μm. The Ms temperature was reported to continuously decrease owing to the refinement of the austenite grain size, and the Ms temperature decrease rate was particularly noteworthy at grain sizes below 100 μm. Furthermore, regarding the stabilization of austenite caused by the grain refinement, Tsuzaki and Maki highlighted that the grain boundaries acted as barriers to the growth of lenticular martensite.15) For this reason, the fine recrystallized material (Fig. 6(b)) exhibited fine lenticular martensite structures, in contrast to the coarse recrystallized (Fig. 6(c)), and as-cast (Fig. 3(b)) materials, which exhibited coarse lenticular martensite structures. Corresponding to the austenitic grain size dependence of Ms, the VM value was also confirmed to decrease along with the austenite grain refinement. However, the VM of the coarse recrystallized material demonstrated a unique behavior, namely, an approximately 10% increase in value compared to that of the as-cast material. Generally, austenite grain boundaries are considered as the optimum martensite nucleation sites. Therefore, even if the inhibition of martensite growth, which is induced by the austenite grain boundaries at lenticular martensite growth of the initial stage of transformation (also known as burst phenomenon), causes the Ms value to decrease, it is considered that the increase of the VM value was attributed to the increased nucleation density which was induced by the refinement of the austenite grain size at the final stage of transformation.
Effect of austenite grain size on (a) martesite start temperature (Ms) and (b) volume fraction of martesite (VM).
Figure 8 summarizes the 0.2% proof stress and Young’s modulus of austenite single-phase structure transformed by the second cryo-annealing treatment in reversed, fine-recrystallized and coarse-recrystallized materials, as functions of the first and second annealing temperatures, T1 and T2. The corresponding values of the as-cast and recrystallized materials are shown for comparison. It was confirmed that all double cryo-annealing materials exhibited low coefficients of thermal expansion, ranging from 1.0 × 10−7 to 8.0 × 10−7 K−1. Although the Young’s modulus (indicated by the color contour) was insensitive to T2, a clear temperature dependence on T1 was observed, with the value significantly increased with an increase in T1. In contrast, the 0.2% proof stress was greatly affected by the T2 temperature and presented a high value at T2 = 873 K, which was just above the Af. Therefore, it was concluded that by increasing T1 to higher values and decreasing T2 to just above the Af during the double cryo-annealing treatment, the simultaneous improvement of the 0.2% proof stress and Young’s modulus behavior could be realized.
Variations of Young’s modulus and 0.2% proof stress as functions of annealing temperatures, T1 and T2, on double cryo-annealing process.
As for the result that the 0.2% proof stress after the double cryo-annealed materials was strongly affected by T2, when T2 was fixed at 873 K the 0.2% proof stress ceased to change monotonically according to T1. Nagpaul and West,16) and Nakada et al.12) performed tensile tests to evaluate the mechanical properties of the duplex austenitic steel alloys consisting of untransformed austenite and reversed austenite of Fe-25.7%Ni-0.4%C and Fe-28%Ni by cryo-annealing treatment. In both reports, the volume fraction of the reversed austenite was altered by changing the subzero treatment temperature, and it was also discovered that the 0.2% proof stress values of the duplex austenitic steels were highly dependent on the reversed austenitic volume fractions. Figure 9 summarizes the relationship between the 0.2% proof stress behavior and the reversed austenitic volume fraction (Vγ rev) for the super invar cast steel subjected to double cryo-annealing as well as the reversed, fine-recrystallized, and coarse-recrystallized materials. It should be noted that the reversed material subjected to cryo-annealing treatment at T2 = 873 K (873 K + 873 K, denoted by ■ in the figure) contains the reversed austenite transformed by the first (γ rev1) and second (γ rev2) cycle of cryo-annealing (Fig. 6(a)). According to the reports of Krauss et al.9,17) and Alaei et al.,18) when the cryo-annealing treatment, responsible for the martensitic transformation and martensitic reversion, was repeated for multiple cycles the first cryo-annealing cycle remarkably strengthened the austenite via martensitic reversion, whereas the subsequent cycles were relatively ineffective. This behavior suggests the similar strengths of γ rev1 and γ rev2, when the recovery of dislocation during the heat treatment process is neglected. Therefore, the reversed austenite volume fraction (Vγ rev) was determined by the sum of the volume fractions of γ rev1 and γ rev2. Additionally, from the apparent linear relationship between the strength and Vγ rev, it was deduced that the strength after the double cryo-annealing treatment was considerably dependent on the reversed austenite volume fraction.
Relation between 0.2% proof stress and volume fraction of reversed austenite in super invar cast steel subjected to single and double cryo-annealing process.
In view of the above results, the optimum heat treatment processing conditions for enhancing the rigidity and strength of super invar cast steel alloys are schematically illustrated in Fig. 10. First, during the first cycle of the cryo-annealing treatment, the austenite recrystallization induced by martensitic reversion promoted orientation randomization, and thus, improved rigidity. Here, it is important that the recrystallized austenite was coarsened by intentionally raising the annealing temperature aiming to promote the martensitic transformation in the subsequent subzero treatment. Consequently, a significant amount of high-strength reversed austenite with high-density dislocation was generated by the second cryo-annealing treatment. As discussed above, it was finally confirmed that the double cryo-annealing treatment of the super invar cast steel alloy yielded a simultaneous enhancement of 30 and 90% in terms of rigidity and strength, respectively.
Schematic illustration explaining appropriate heat treatment condition of double cryo-annealing process for simultaneous optimization of rigidity and strength.
To simultaneously improve the Young’s modulus and strength of super invar cast steel alloys, we investigated the effects of a subzero treatment followed by annealing treatment (cryo-annealing process) on the variations of its microstructure and mechanical properties. The obtained conclusions are as follows:
(1) When the annealing temperature in the cryo-annealing treatment was just above the Af, the strength significantly increased due to the reversed austenite formation which maintained the high-density dislocations. However, the Young’s modulus was not improved because the texture was inherited owing to crystallographic reversibility. On the other hand, when the annealing temperature was sufficiently high, austenite recrystallization predominantly occurred by the dislocations in the reversed austenite as the driving force. As a result, the Young’s modulus was improved due to orientation randomization; however, the increase in strength due to grain refinement was insignificant, thus a remarkable increase in strength was not achieved.
(2) The start temperature of martensitic transformation by subzero treatment continuously decreased with the progression of austenite grain refinement. However, the austenite grain boundaries were theorized to act as martensite nucleation sites, causing the martensite volume fraction to maximize at a reasonably coarse austenite grain size.
(3) Following double cryo-annealing treatment, the Young’s modulus and 0.2% proof stress behaviors exhibited a strong dependence on the first and second annealing temperatures, respectively. Therefore, the austenite recrystallization and subsequent grain growth were promoted by intentionally raising the first annealing temperature during the double cryo-annealing process. A significant amount of high-strength reversed austenite was also generated by setting the second cryo-annealing temperature just above the martensitic reversion finish temperature. As a result, the simultaneous improvement in the rigidity and strength of the super invar cast steel alloy was achieved.
