Soft Ferromagnetic Bulk Metallic Glass with Potential Self-Healing Ability

Abstract

A new concept of soft ferromagnetic bulk metallic glass (BMG) with self-healing ability is proposed. The specific [Fe₃₆Co₃₆B_19.2Si_4.8Nb₄]_100−x(Ga)_x (x = 0, 0.5, 1 and1.5) BMGs prepared by copper mold casting were investigated as a function of Ga content. The Ga-containing BMGs still hold soft magnetic properties and exhibit large plastic strain of 1.53% in compression. Local melting during shearing produces molten droplets of several µm size throughout the fracture surface. This concept of local melting during shearing can be utilized to produce BMGs with self-healing ability. The molten regions play a vital role in deflecting shear transformation zones, thereby enhancing the mechanical properties.

1. Introduction

Among several bulk metallic glass (BMGs) systems, Fe-based BMGs are considered as a class of engineering materials with enormous potential for many fields of application due to their unique combination of excellent magnetic properties, good corrosion resistance, high hardness, high fracture strength, and low cost [1,2,3]. However, these Fe-based BMGs do not exhibit strain hardening and cannot sustain plastic deformation at room temperature. This brittle behavior under mechanical loading severely limits their application as engineering and structural materials. In recent years, huge efforts have been undertaken to improve the plastic deformability of monolithic BMGs or BMG composites [4]. Several researchers have proved that plastic deformability in metallic glasses can be largely enhanced by introducing different microstructural heterogeneities such as nano-crystals, phase separation, or short-/medium-ranged clusters into the glassy matrix by controlling the casting process [5,6,7,8]. In the case of Fe-based BMGs, the room temperature plasticity was improved by the addition of small amounts of soft elements like Ni, Er, and Cu [9,10,11,12]. In recent years, among the several Fe-based glassy alloys Fe-Co-B-Si-Nb BMGs have been regarded as one of the most promising alloys because of its combined advantages of functional and structural materials [1].

Motivated by these results, we considered to dope a Fe₃₆Co₃₆B_19.2Si_4.8Nb₄ base alloy (hereafter named FeCoBSiNb) with Ga. Since the compositional variation of the Ga content in steps of 0.1 at.% is very difficult, we considered adding Ga in steps of 0.5 at.%. In previous works, Ga was employed as a main alloying element to improve the glass-forming ability (GFA) in a few glass-forming alloy systems [13,14,15]. However, in the present study, Ga is considered as a doping element to improve the plastic deformability of Fe-based BMGs without hindering their GFA.

2. Materials and Methods

In this work, Ga-containing glassy samples were produced in 3 steps. An eutectic Fe25Nb75 (wt.%) pre-alloy, together with the crystalline B (99 wt.%), crystalline Si (99.99 wt.%), Fe and Co lumps (99.9 wt.%), were melted together by induction heating under a protective Argon atmosphere. To this master alloy the desired amount of Ga (99.99 wt.%) was added through arc-melting. Cylindrical samples were prepared by remelting the pieces of the master alloys in quartz tubes and subsequently injecting into a water-cooled copper mold. Structural characterization of the as-cast and annealed samples was done by X-ray diffraction (XRD) using a PHILIPS PW 1050 diffractometer with Co-K_α radiation (λ = 0.17889 nm). Microstructural investigation of the as-cast glassy samples was performed using a Carl Zeiss Libra 200MC High Resolution Transmission Microscope (HRTEM). The thermal behavior of the glassy samples was evaluated with a differential scanning calorimeter (DSC) NETZSCH DSC 404 C with an accuracy of ± 2 K. To study the structural evaluation, samples were annealed for 5 min. at different temperatures using the same DSC, with a constant heating and cooling rate of 20 K/min. For compression testing, an INSTRON 8562 device was used in constant strain rate mode at room temperature. The fracture surfaces of the samples were investigated using a scanning electron microscopy (SEM) Zeiss LEO 1525, equipped with energy-dispersive X-ray spectroscopy (EDS). For magnetic measurements, DC M-H hysteresis loops were recorded with a vibrating sample magnetometer (VSM) at ambient temperature.

3. Results and Discussion

3.1. Structural and Thermal Characterization

Figure 1a shows the XRD pattern of as-cast (FeCoBSiNb)_100−x(Ga)_x (x = 0, 0.5, 1, 1.5) glassy samples with maximum achievable diameter for each composition.

All samples exhibit similar diffuse patterns without any distinct crystal diffraction peaks indicating the formation of a glassy structure. However, a second broad diffraction maximum is observed around 98.35° for the samples containing 1 and 1.5 at.% Ga. According to Bernal’s dense random packing hard sphere model [16], the position of the first and second broad diffraction maxima contain information about mean interatomic distances and medium-range order (MRO) in glassy systems. Therefore, it can be assumed that the substitution of Ga atoms leads to some atomic level rearrangements. As reported in the literature, the critical diameter (D_c) for this FeCoBSiNb alloy is 5 mm [1]. If the alloy is fluxed with B₂O₃ prior to casting then D_c can be increased up to 7.7 mm [17]. In our work, following the above described experimental route, D_c is 3 mm. The difference in D_c could be due to the purity of the starting raw materials, the casting technique and the casting parameters used. The addition of Ga does not affect the GFA of this FeCoBSiNb alloy and D_c (3 mm) also remains the same up to 1.5 at.% Ga addition. This may be due to the relatively large atomic radius of Ga (1.53 nm) and negative heats of mixing for Fe-Ga, Co-Ga pairs [18].

The DSC traces of as-cast glassy samples with D_c are shown in Figure 1b. For all compositions, the DSC traces display a clear glass transition event, followed by a supercooled liquid region (SLR) and crystallization, further confirming the glassy structure of these alloys. The thermal properties, such as onset temperature of crystallization T_x, glass transition temperature T_g, crystallization peak temperatures (T_p1, T_p2 and T_p3), liquidus temperature T_liq, extension of the SLR ΔT_x (ΔT_x = T_x − T_g), the GFA parameter γ (γ = T_x/(T_g + T_liq)), and the reduced glass transition temperature T_rg (T_rg = T_g/T_liq) for the present BMGs are listed in

Table 1. Thermal stability parameters T_g, T_x, T_p1, T_p2, T_p3, T_liq, T_rg, γ and ΔT_x as a function of Ga content for [FeCoBSiNb]_100-x(Ga)_x (x =0, 0.5, 1 and 1.5) BMGs.

Compositions	Tg (K)	Tx (K)	Tp1 (K)	Tp2 (K)	Tp3 (K)	Tliq (K)	Trg	γ	ΔTx (K)
FeCoBSiNb	825	858	867	985	1050	1475	0.56	0.373	33
FeCoBSiNb + 0.5Ga	825	863	873	1037	-	1475	0.56	0.375	38
FeCoBSiNb + 1Ga	825	863	873	1038	-	1440	0.57	0.381	38
FeCoBSiNb + 1.5Ga	825	869	873	1038	-	1440	0.57	0.383	41

. The addition of Ga does not affect T_g, but T_x marginally increases from 858 K to 863 K for both 0.5 and 1 at.% Ga, and to 869 K for 1.5 at.% Ga-added glasses. On the other hand, the two small exothermic events T_p2 and T_p3 observed for FeCoBSiNb glass disappears. Instead, a single strong exothermic event T_p2 is observed for Ga-added glasses.

Figure 1c illustrates the XRD patterns of the (FeCoBSiNb)_100−xGa_x (x = 0, 0.5, 1and 1.5) glassy samples after constant-rate heating in the DSC up to the end of their first crystallization event, which is at 902 K. The crystallization sequence of the BMGs containing 0 and 0.5 at.% Ga is almost identical. The BMGs with 1 and 1.5 at.% Ga show distinct bcc-(Fe,Co) peaks along with (Fe,Co)₂₃B₆-type peaks at the end of the first crystallization event. The formation of the bcc-(Fe,Co) phase for BMGs with higher Ga contents may be associated to the presence of Ga atoms in the matrix, which may aid the nucleation of the bcc-(Fe,Co) phase. The absence of bcc-(Fe,Co) phase for lower Ga contents indicates that 0.5 at.% of Ga is not sufficient to trigger the nucleation of this phase.

The M-H hysteresis loops of the different BMGs are shown in Figure 1d. From the hysteresis loops it is clear that saturation occurs at relatively similar magnetic fields for all glassy samples. The saturation magnetization (M_s) decreases marginally from 108 to 105 Am²/kg when the Ga content increases from 0 to 1.5 at.%. This decrease in M_s is attributed to the addition of the non-magnetic element Ga.

High resolution transmission electron microscopy (HRTEM) investigations were carried out to understand the effect of Ga addition in FeCoBSiNb glass. As-cast FeCoBSiNb and FeCoBSiNb + 0.5Ga samples were subjected to wide-angle X-ray diffraction before preparing for HRTEM investigations and the samples were amorphous. HRTEM micrographs for FeCoBSiNb and FeCoBSiNb+0.5Ga samples are shown in Figure 2a,b, together with their corresponding fast Fourier transformed (FFT) images (shown as insets). The as-cast FeCoBSiNb + 0.5Ga glass exhibits some locally ordered regions, which are marked by white circles in the image, but it is not evident whether the Ga atoms cause this ordering/cluster structure. The exact identification of the chemical composition of these locally ordered regions is not easy due to their very small size and the fact that they are dispersed in a chemically complex multicomponent glassy matrix. EDX analysis indicates a faint clustering of Ga atoms.

However, it should be noted that the presence of ordered zones/clusters in a glassy matrix does not necessarily imply that they act as nucleation sites for crystal growth. The ordered embryos must be larger than the critical nucleus size in order for their growth to be thermodynamically favorable [19,20,21]. Since no ordered regions were observed in the Ga- free glass, it can be assumed that these ordered zones are due to the Ga atoms. From the above structural evaluation and thermal stability data it is evident that these ordered zones do not act as heterogeneous nucleation sites, thereby the GFA remains unaffected.

3.2. Mechanical Properties

The room temperature compressive true stress-true strain curves for the as-cast [FeCoBSiNb]_100−x(Ga_x)(x = 0, 0.5, 1 and 1.5) BMGs are shown in Figure 3. All the curves are offset from the origin for better visibility. All samples exhibit an elastic deformation regime followed by a small compressive plastic regime. From the stress-strain curves it is evident that the plastic strain increases gradually with increasing Ga content.

For the 0.5 at.% Ga-added sample, the yield strength σ_y, elastic strain ɛ_y, fracture strength σ_f and plastic strain ɛ_pl are almost the same as for the FeCoBSiNb parent glass, i.e., 3.95 ± 0.02 GPa, 2.80%, 4.06 ± 0.02 GPa and 0.43%, respectively. For higher amounts of Ga addition, the plastic deformation is improved. The yield stress is 3.71and 3.62 ± 0.2 GPa for the 1 and 1.5 at.% Ga-added samples and their corresponding elastic strain values are 2.73% and 3.30%, respectively. The fracture of all Ga-added samples occurs at nearly the same value of true stress, around 4 GPa, but the corresponding fracture strain is different: 3.87% for 1 at.% Ga- and 4.83% for 1.5 at.% Ga-added samples. The plastic deformation of the samples extends up to 1.07% and 1.53% for 1 at.% and 1.5 at.% Ga-added samples. The improvement in the plastic deformability of the glassy rods containing Ga can be explained based on the ordered zones observed in the TEM investigations. The dispersed ordered zones seem to be responsible for promoting plastic deformability in the Ga-containing samples, which is not observed for the Ga-free samples loaded under the same conditions. The ordered zones present in the Ga-containing samples affect the localized deformation in the glassy matrix. They may hinder shear band propagation by acting as pinning centers or by deflecting the shear bands into multiple small branches. The volume and density of these ordered zones also play a crucial role in deflecting the shear bands. If they are distributed densely in a narrow zone, then the deflection will be very less leading to brittle failure. The plasticity increase in the stress-strain curves reveals that the volume fraction and density of these ordered zones are large enough for affecting the shear band propagation through the material but not too high to cause distinct ductile failure. Sarac et al. [22] in their recent work showed that the presence of soft zones/nanocrystals in an amorphous matrix plays a key role in improving the plasticity by controlling the shear propagation rate.

3.3. Fracture Surface Analysis

The fracture surface morphology of the samples after the compression tests was investigated using SEM. The fracture surface of the FeCoBSiNb and FeCoBSiNb + 0.5Ga samples consists of a number of small fracture zones and their zone planes appear to be inclined by about 70°–90° to the direction of the applied load. This kind of failure is mainly due to the simultaneous generation of a large number of small facture events at many sites, at a very high stress level close to 4.0 GPa. The high fracture strength of these glasses is mainly due to the covalent bonding nature of their constituents [1,23]. Although a slight increase in plastic strain is observed for FeCoBSiNb + 0.5Ga samples, the overall appearance of the fracture surface points towards a brittle fracture mode, with many shells indicating the generation and propagation of cracks. Neither shear bands on the lateral surfaces nor typical vein patterns are observed on the fracture surface of these samples.

More interesting features are observed for the samples containing 1 and 1.5 at.% Ga. Figure 4a,b show the fracture surface of the FeCoBSiNb + 1Ga samples after compression testing taken from different regions with different magnifications. The fracture surfaces contain a large number of vein patterns together with several micro-cracks (indicated by green arrows in the image). At first the material starts to deform plastically; after a while brittle fracture starts, because the cracks are superimposed on the vein patterns, indicating a combination of ductile and brittle failure modes.

Figure 4c represents the fracture surface of the FeCoBSiNb + 1.5Ga samples taken from the center of the samples after compression testing and Figure 4d shows a magnified image of the area marked with dotted lines in Figure 4c. The red arrows indicate metal droplets, which are observed throughout the fracture surface. Compared to the 1 at.% Ga samples, the fracture surface of the 1.5 at.% Ga samples displays more vein patterns and less micro-cracks. Again, the micro-cracks are superimposed on the vein patterns. The presence of metal droplets on the fracture surfaces suggests that at the very high stress level of 4.0 GPa, the flowed layers must have melted, as is commonly observed in case of deformable BMGs [24]. In a very recent work Stoica et al. [11] showed that during compression of FeCoBSiNb + 0.5Cu glass the temperature in the shear plane can raise up to 1352 K, by assuming a maximum shear propagation rate, i.e., 0.9Vs (Vs = speed of sound). The raise in temperature (1352 K) is just 30 K above the solidus temperature of the FeCoBSiNb + 0.5Cu alloy and barely enough to melt very thin layers. The solidus temperatures of FeCoBSiNb + 0.5Cu and (FeCoBSiNb]_100−x(Ga)_x (x = 0, 0.5, 1 and 1.5) alloys are ≈1310 K [25] (supplementary Figure S1). In case of FeCoBSiNb + 0.5Cu glass [11] no such molten droplets were found on the fracture surface but for FeCoBSiNb + 1.5Ga glass µm-sized molten droplets are observed throughout the fracture surface (see Figure 4c,d. EDS analysis performed on the molten droplets indicates a minor increase in Ga content compared to the other fractured regions without any melting (supplementary Figure S2 and Table S1). Researchers have already proved that Ga has a tendency to form nano-precipitates with Fe [26,27,28]. From these results it can be assumed that the ordered zones formed due to the addition of Ga atoms play an important role in deflecting shear bands and its contributing to some local melting due to the rise in temperature during shear band propagation. EBSD measurements were carried out to identify the nature of the molten droplets. The absence of Kikuchi bands indicates that they are amorphous (supplementary Figure S3). This proves that the cooling rate is high enough to freeze the molten region instantly.

4. Conclusions

In summary, we have revealed that Ga can be considered as a potential doping element for Fe₃₆Co₃₆B_19.2Si_4.8Nb₄ glass; it improves glass-forming ability as well as plastic deformability to about 1.07% and 1.53% plastic strain for 1 and 1.5 at.% Ga addition without deteriorating the magnetic properties. Furthermore, for the first time we reported the presence of molten regions of several µm size throughout the facture surface in Fe-based BMG. Local melting at crack tips accompanied by fast cooling can be considered as a promising healing mechanism for improving industrial applications of Fe-based BMGs in many ways.