Microstructure refinement of Fe 40 Ni 40 B 20 alloy in non-equilibrium solidification: possibility of nanostructure formation

Microstructure refinement in undercooling solidification

The typical cooling and heating curves for the solidification of Fe₄₀Ni₄₀B₂₀ alloys with different ΔT are shown in Fig. 2. The highest ΔT up to 220±10 K was achieved in the undercooling experiments. Two kinds of microstructure refinement take place with increasing ΔT. From Fig. 3a and b, the primary γ-(Fe, Ni) dendrite and lamellar eutectic γ-(Fe,Ni)/(Fe,Ni)₃B are significantly broken and refined for ΔT from 0 to 65 K. The first microstructure refinement occurs at ΔT∼65 K. For ΔT = 80–120 K, the as solidified microstructures are almost occupied by the primary dendrites (Fig. 3c and d). These dendrites are much coarser than those in Fig. 3b. However, with the increasing ΔT, the second microstructure refinement takes place for ΔT>143 K (Fig. 3e and f). After that, only anomalous eutectic and primary dendrites are observed (Fig. 3f). For even larger ΔT (>200±10 K), the anomalous eutectic is very fine, and the size of each phase in the eutectic is ∼2 μm (Fig. 4).

OPEN IN VIEWER

Figure 2

Typical cooling and heating curves of Fe₄₀Ni₄₀B₂₀ alloy corresponding to different ΔT recorded by infrared pyrometer

OPEN IN VIEWER

Figure 3

Microstructures of Fe₄₀Ni₄₀B₂₀ alloy with a ΔT = 0 K, b ΔT = 65 K, c ΔT = 80 K, d ΔT = 123 K, e ΔT = 143 K and f ΔT = 181 K (optical microscope images); inset in f (SEM image) shows anomalous eutectic structures

OPEN IN VIEWER

Figure 4

Anomalous eutectic structures in deep undercooled Fe₄₀Ni₄₀B₂₀ alloys with ΔT = 220 K

Microstructure refinement under different cooling conditions

In these experiments, Fe₄₀Ni₄₀B₂₀ melts with ΔT≈50 K are solidified under different cooling conditions. The microstructures obtained by natural cooling, water quenching along with the crucible and direct water quenching are shown in Fig. 5a–c respectively. One can see that the microstructures are refined gradually with an increasing cooling rate. Meanwhile, the remelting and fragmentation of the dendrite weaken.

OPEN IN VIEWER

Figure 5

Microstructures for Fe₄₀Ni₄₀B₂₀ alloys with ΔT≈50 K subjected to different cooling conditions

The microstructure from iron plate cooling is shown in Fig. 6. It is composed of dendrites at the edge of the sample (Fig. 6a) and lamellar eutectic cells in the middle of the sample (Fig. 6b). The similar microstructure is obtained by Ga–In alloy bath cooling, but they are very fine (Fig. 7). The eutectic cells in the middle of the sample are ∼5 μm in diameter (Fig. 7c). Particularly, the spacing of the eutectic lamellae is <100 nm, which means nanostructure eutectic has been achieved by this method.

OPEN IN VIEWER

Figure 6

Microstructures for Fe₄₀Ni₄₀B₂₀ alloys with ΔT≈50 K subjected to iron plate cooling

OPEN IN VIEWER

Figure 7

Images (SEM) of Fe₄₀Ni₄₀B₂₀ microstructure obtained by Ga–In alloy bath cooling; magnified images of regions marked by capital letters in a are shown in b and c, b dendrite and c lamellar eutectic cells; inset in c is TEM image of lamellar eutectic cell with spacing of lamellae <100 nm

The as spun microstructure from the overheated melt is also shown in Fig. 8. The corresponding X-ray diffraction patterns are given in Fig. 9. From Figs. 8 and 9, it is found that the amorphous structure, along with dispersed large dendrites, appears at r_S = 2500 rev min⁻¹ (with a ribbon thickness h of ∼50 μm), and a totally amorphous structure is obtained when r_S>5000 rev min⁻¹ (with h≈20 μm).

OPEN IN VIEWER

Figure 8

Image (SEM) from cross-section of Fe₄₀Ni₄₀B₂₀ melt spun ribbons obtained at rotation speed of 2500 rev min⁻¹: some dispersed dendrites are shown

OPEN IN VIEWER

Figure 9

X-ray diffraction patterns for as solidified Fe₄₀Ni₄₀B₂₀ ribbons obtained by melt spinning method at different rotation speeds

Microstructure refinement in undercooling solidification

In the present undercooling experiments, two kinds of microstructure refinement take place with increasing ΔT. It is mainly a result of the fragmentation of the primarily formed solid phase during recalescence. The dendrite fragmentation by remelting can be predicted by comparing the characteristic time of dendrite break-up and the time for the interdendritic melt to be completely solidified (i.e. the plateau duration for recalescence).4 The plateau duration is exclusively controlled by the heat transfer from the sample to the environment and can be varied by changing the cooling rate.4 The characteristic time of dendrite break-up depends on the undercooling. It changes non-monotonously with increasing undercooling.4 Therefore, under certain cooling conditions, twice remelting can be predicted in undercooled solidification of many single phase alloys by comparing the two characteristic times.6^–9 In binary alloys Ni–Cu,6 Fe–Ni,7 Fe–Co⁸ and Co–Pd,9 the experimental measured undercooling intervals for first microstructure refinement by single phase dendrite remelting are 27–84, 22–80, 22–83 and 72–95 K respectively. They are compatible with the predicted values calculated by the dendrite fragmentation model.4 At large undercoolings, the stresses caused by solidification shrinkage and fluid flow also favour the fragmentation of the microstructure.6,9 The dendrite fragmentation due to both remelting and high stress together causes microstructure refinement.6,9 The corresponding critical undercoolings are 180, 140, 183 and 142 K for Ni–Cu,⁶ Fe–Ni,⁷ Fe–Co⁸ and Co–Pd⁹ alloys respectively.

In ternary eutectic Fe₄₀Ni₄₀B₂₀ alloys, the primary dendrites and the eutectic structures occupy the final microstructures. The microstructure refinement at low undercooling is similar to that in single phase alloys and due to the primary dendrite remelting. This can also be predicted using the dendrite break-up model,4 but the calculating result is unavailable here for the lack of model parameters of the FeNiB ternary alloy. The experimental obtained undercooling interval for the first refinement is below 65 K. It is comparable with those of binary alloys. It should be noticed, in ternary eutectic alloys, that the primary dendrite break-up is just one cause of microstructure refinement. There are two phases in the microstructure, and the remelting and fragmentation of a regular eutectic phase into an anomalous eutectic are also reasons for refinement. At large undercoolings, the high stress produced during the rapid solidification could also lead to the break-up of the primarily formed structures and cause the second refinement. The corresponding undercooling is ∼140 K in the present case. The critical undercooling is also comparable with those of binary alloys. Moreover, for a ternary eutectic alloy, the plateau time for recalescence is much longer than that for a single phase alloy due to the slower eutectic formation after the primary dendrite growth. Thus, the remelting is more obvious, and there is no integrated dendrite in the undercooling experiments.

Microstructure refinement in rapid cooling solidification

Similar to those in undercooling experiments, the microstructure refinements in rapid cooling solidification are mainly the result of structure fragmentation of the primarily formed solid phase when the cooling rate is not very high. Whereas, at the high cooling rates, some other causes are dominating the microstructure formation. The reasons will be given in the following section. First, some characteristic parameters are defined as follows: the temperature for the beginning of recalescence T_N, the liquidus and solidus temperatures T_L and T_S, the maximum recalescence temperature T_R and the recalescence height ΔT_R (ΔT_R = T_R−T_N = T_R+ΔT−T_L).

In the solidification, the morphologies of the as obtained microstructure are largely influenced by T_R, which is related to both heat production and heat transfer during recalescence. T_R can be obtained from the energy conservation law as20 (1) where f_S is the fraction of solid phase, t_R is the time for recalescence, C_p is the volumetric specific heat and is the heat extraction rate per unit volume, which is related to the contact area between the cooling mediums and melts and the heat conductivity of the cooling mediums.

For Fe₄₀Ni₄₀B₂₀ melts with the same ΔT, the obtained T_R will be different subject to different cooling conditions (i.e. different ) according to equation (1). At the smaller cooling rate, because the latent heat cannot be transferred to the surroundings immediately, the relatively larger T_R will be achieved, which leads to the remarkable remelting and fragmentation of the microstructure. With the increasing cooling rate, the remelting and fragmentation weaken due to decreased T_R. This is consistent with the experimental observation. Only dendrite fragments and/or anomalous eutectics caused by remelting and fragmentation are found in the microstructure by natural cooling or water cooling (Fig. 5). However, the microstructure obtained at very high cooling rates contains well developed dendrites and eutectics with no evidence of remelting or fragmentation (Figs. 6 and 7). Moreover, as mentioned in the above section, the plateau duration of recalescence is exclusively controlled by the heat transfer from the sample to the environment. At a certain undercooling, the plateau duration decreases monotonously with increasing cooling rate.4 According to the fragmentation model,4 in rapid cooling conditions, the high heat transfer rate will lead to a short plateau duration and, thus, a weak fragmentation by remelting.

Nanostructure formation: Combination effects of undercooling and rapid cooling

In all the experimental results, it is worth noting that the eutectic cells with nanostructure have been in formation in the FeNiB alloy with ΔT≈50 K subjected to Ga–In alloy bath cooling. There are three possible mechanisms for the formation of nanostructure in as solidified microstructures: the spinodal decomposition of liquid leads to the final nanostructure solid phase; the decomposition of or precipitation from the primarily formed metastable phases gives birth to the nanostructure; and the nanosized grain solidifies directly from the melts due to the high nucleation rates and low growth rates at a certain condition. In the present case, the nanostructure eutectic is a result of the proper nucleation and growth rates due to the combination effects of undercooling and rapid cooling.

As is known, both nucleation and growth of the crystal phase during solidification are thermal activation processes. The energy barrier of growth is mainly the activation energy for atoms to cross the solid/liquid interface; thus, it is generally smaller than that for nucleation, which consists of the interfacial energy of the nuclei and the activation energy for atom movement. Therefore, in most cases, the nucleation rate has a maximum at the lower temperature T (the larger ΔT) compared with the growth rate.21 The typical curves for T dependence of the nucleation rate I and the growth rate V are plotted in Fig. 10. The temperatures and , which correspond to the maximum I and V, are not equal. This makes it possible to control the as solidified microstructure by different combinations of I and V, which can be fulfilled by changing the cooling rate or the temperature. In Fig. 10, if the melt temperature T is equal to T₁, I is large while V is relatively small, a fine as solidified microstructure could be formed. If T is equal to T₂, both I and V are very small; thus, an amorphous structure could be obtained.

OPEN IN VIEWER

Figure 10

Typical curves for temperature dependence of crystal nucleation (solid line) and growth rates (dashed line): and correspond to maximum I and V respectively

In the case of Ga–In alloy bath cooling with ΔT≈50 K, heterogeneous nucleation occurs at the surface of the sample where a large amount of heterogeneous nucleation sites exist. The primary dendrites originate from these sites and grow directionally toward the centre of the sample due to directional heat transfer (Fig. 7a and b). Meanwhile, as a result of undercooling in the melts, a large amount of eutectic cells nucleate (Fig. 7c). These eutectic cells originate from individual heterogeneous nucleation events.22,23 As discussed in the above section, at a high cooling rate, the remelting of the as solidified structure will not happen. Therefore, the regular lamellar structure is kept, i.e. regular lamellar growth proceeds in the eutectic cells. Moreover, the melt undercooling and the high cooling rate together lead to a high growth rate of eutectic lamellae. According to the models for the growth of regular eutectic,24,25 the spacing between the lamellae is inversely proportional to the square root of the growth rate. The lamella in the eutectic cell is therefore very thin. A huge number of eutectic cells solidify simultaneously and quickly grow into the surrounding melts and touch each other. There is no spacing or remaining melts left for them to grow to a large size. They have a small size at last. All these factors lead to the final small eutectic cells with nanostructure in them. However, in the as spun sample, only an amorphous structure with a few large dendrites in it, rather than nanostructure, is found. The reason is the suppression of the nucleation of the crystal phase from overheated melts subjected to very rapid cooling.

Experimentally, the nanostructure cannot be prepared by either undercooling method or rapid cooling method alone for the Fe₄₀Ni₄₀B₂₀ alloy. By a combination of melt undercooling with rapid cooling, the nanostructure can be obtained. On the one hand, a large nucleation rate can be attained by undercooling. On the other hand, rapid cooling leads to quick heat extraction, which maintains the regular lamellar eutectic growth with a proper growth rate. These two factors lead to a large amount of nuclei growing slowly during the subsequent solidification and the formation of nanostructure.