Effect Of Titanium In Cu-Hf Based Bulk Metallic Glasses

Metallic glasses are non-crystalline alloys formed by continuous cooling from the liquid state.  Up until recently [1], such materials could only be formed in thin sections at cooling rates typically in the range 10⁴-10⁶ Ks^-1.  Such alloys, based on iron, nickel and/or cobalt have established numerous applications as ultra soft magnetic materials in power and high frequency cores for transformers, chokes, inductors and other similar devices [2].  In the past decade, researchers had paid more attention to the metallic glasses not only due to their superior chemical and physical properties, but because they can now be fabricated in bulk form.  They are therefore considered as real practical engineering materials and open up new application opportunities.  For instance, glassy alloy can be used for making magnetic tapes [3, 4] as it exhibits better wear resistance.  Extensive experimental investigations have demonstrated that many compositions in these systems can be formed into the glassy state by using suction casting to form cylindrical or slab-shaped ingots.  Maximum section diameters or thicknesses of fully glassy phase range up to 72mm [5], depending on the combination of constituent elements and their precise concentrations.  Recently, Inoue et al. [6, 7] reported Cu-based alloys with high glass forming abilities, high tensile strength of over 2000 MPa and lower material cost bulk metallic glasses which can be prepared by copper mould casting.  Their excellent mechanical properties and glass forming ability (GFA) make the production of precision mechanical parts possible such as high precision gears [6].  Driven partially by interest in engineering applications, there has been an ongoing effort to identify amorphous alloys with greater strength, elastic modulus, hardness, and ductility.  Of particular interest are alloys based on such common metals such as Cu, Al, Co, Fe, Ni, etc.  However, the information of a glassy phase in Cu-Hf-Ti alloys is based on 60 at.% Cu.  The aims of this investigation are to produce bulk amorphous rods of Cu-Hf and Cu-Hf-Ti and determine the critical glassy diameter of these alloy families.

Experimental

Alloy ingots of nominal compositions Cu_100-xHf_x    (x = 50, 45, 40, 35, 30 and 25 at.%)  and Cu₅₅Hf_45-xTi_x (x = 5, 10, 15, 20, 25, 30 , 35 and 40 at.%) were prepared by arc melting mixtures of Hf (crystal bar),  Cu (sheet) and Ti (Rod) having purities of 99.5 at.%, 99.99 at.% and 99.8  at.%, respectively.  The arc melting was performed in a Ti-gettered high purity Argon atmosphere.  Each ingot was re-melted at least four times in the arc melter in order to obtain good chemical homogeneity.  Ribbon samples of mean thicknesses ~ 25 μm and width ~ 2 mm were prepared by melt spinning in a controlled Ar atmosphere.  Copper die suction casting was employed to produce rods with a stepped profile having diameters decreasing from 4 to 3 to 2 mm, each with a total length of 50 mm.  The phase constitutions of the rods were studied by X-ray diffraction (XRD).  The thermal stability, defined by the glass transition temperature (T_g) and the crystallization temperature (T_x), was studied by differential scanning calorimetry (DSC) at a heating rate of 20 K/min.  The solidus temperature (T_m) and liquidus temperatures (T_l) were determined by differential thermal analysis (DTA) at a heating rate of 20 K/min.  Thermal characterization was performed using melt spun ribbons.

Results

The average thickness of the ribbon produced was 25 μm and almost all the ribbons produced were of uniform width.  All the alloy ribbons formed, except for one sample, the binary alloy Cu₇₅Hf₂₅, were found to be nominally fully amorphous by XRD analysis.  They showed high metallic lustre and could easily be bent through 180⁰ without fracture, thus showing very good ductility to a high strain and indicating a fully amorphous or almost fully amorphous thickness.  However, the Cu₇₅Hf₂₅ alloy ribbon had very poor ductility and it was shown to be crystalline by the XRD analysis.  Stepped rods were produced of diameters 2, 3 and 4 mm with a length of 50 mm.  Figure 1 shows a photograph of a 2/3/4 mm diameter stepped rod produced by suction casting.  The rods produced by suction casting showed good metallic lustre and a low incidence of fabrication defects.

The structure results by X- ray diffraction for all the ribbon and rod samples of the Cu_100-xHf_x alloy series with x = 25 – 50 are summarised in Table 1.  All rods were crystalline or largely crystalline as several distinct peaks can be seen for all these traces with no clear evidence of a diffuse halo corresponding to a glassy phase.  The ribbon samples are clearly fully amorphous in each case, with the exception of the composition Cu₇₅Hf₂₅ which shows a small fraction of crystalline structure, with one peak at ~ 43^o 2 θ corresponding to the phase Cu₈Hf₃ (111).  Figure 2 shows DSC curves for the melt-spun ribbons of this alloy series and values of T_g and T_x are plotted as functions of Hf content in Figure 3.  Although no clear glass transition is observed for 30 at.% Hf, the alloys containing 50 at.% to 35 at.% Hf exhibit distinct glass transitions, followed by the glass transitions region before crystallisation, however, the alloy containing 25 at.% Hf was found to be partially crystalline, giving a DSC curve with amorphous + crystalline structure.  The T_g and T_x decrease with increasing Hf content from 35 at.% to 45 at.%, despite the fact that the Hf has a much higher cohesive energy than Cu.  It is also seen that the supercooled liquid region defined by the temperature interval between T_g and T_x, ΔT_x (=T_x − T_g), shows a relative constant value of ~ 30 K over the range 35 – 50 at.% Hf, but appears to narrow considerably for 50 at.% Hf.  The substitution of Hf for Cu in the range 25 – 35 at.% Hf decreases the liquidus temperature, but increasing the Hf content beyond 35 at.% Hf then leads to an increasing T_l.  The values of the reduced glass temperature (T_rg=T_g/T_l) are plotted as a function of Hf content in Figure 4.  T_rg increases when decreasing Hf content from 0.55 up to 0.61 for x = 50 and 35 respectively, then decreasing to the value of 0.59 for x = 30.  Table 2 shows the thermal analysis results for this alloy series.

Table 1. XRD structural results for melt spun ribbons and for 2, 3 and 4 mm diameter rods, where Am and Cr indicate an amorphous or a crystalline structure, respectively.

Table 2. Thermal Analysis Results of Cu_100-xHf_x and Cu₅₅Hf_45-xTi_x alloy series (Heating rate 0.33 K/s).

The effect of Ti addition on the glass transition and the thermal stability was in general accord similar with that obtained by Inoue et al. [6] and by Stewart et al. [8].  The X- ray diffraction results for all the ribbon and rod samples in this alloy series are summarised in Table 1.  Figure 5 shows the diffraction patterns for the cross sections of the 2 and 3 diameter cast rod samples of the Cu₅₅Hf₃₀Ti₁₅, Cu₅₅Hf₁₅Ti_30, Cu₅₅Hf₂₀Ti₂₅ and Cu₅₅Hf₂₅Ti₂₀ alloys, respectively.  It is clearly seen that the maximum diameter showing an amorphous structure is 3 mm corresponding to the compositions Cu₅₅Hf₂₀Ti₂₅ and Cu₅₅Hf₂₅Ti₂₀, while the 2 mm diameter corresponds to samples Cu₅₅Hf₃₀Ti₁₅ and Cu₅₅Hf₁₅Ti₃₀.  Figure 6 shows the DSC curves for the melt-spun Cu₅₅Hf_45-xTi_x alloys.  In comparison with the Cu_100-xHf_x binary alloys, the addition of titanium decreases T_g and T_x, accompanied by a change in crystallisation mode from one to two or three stages.  Figure 7 shows the variation in T_g and T_x with the titanium content for the melt-spun Cu₅₅Hf_45-xTi_x glassy alloys.  It is noted the decrease in T_g from 750 K to 665 K, while T_x has a maximum of 800 K at 5 at.% Ti and then decreases to 693 K at 40 at.% Ti.  ΔT_x has a maximum of 47 at 5 at.% Ti, followed by a decrease of ΔT_x with further increment in Ti content.  Similar behaviour was reported previously in Cu₆₀Hf_40-xTi_x alloys with the same Ti content [6].  T_m and T_l decrease to minimum values at x ~ 25 and then increase up to x = 45 (Figure 7).  The quotient T_g/T_l (T_rg) increases from 0.56 at x = 5 to a maximum of 0.62 at x ~ 25 and then decreases to 0.55 for x = 40 (Figure 8).  In this alloy series the largest fully glassy section thicknesses (3 mm dia.) corresponds to x = 20 and 25.  Table 2 shows the thermal analysis results for this alloy series.

Discussion

As described above, the binary Cu_100-xHf_x alloy series showed a glassy structure for ribbon samples in the composition range 30-50 at.% Hf.  The alloy ribbon containing 25 at.% Hf, had a crystalline + amorphous XRD pattern, evidently due to a reduced glass forming ability since all ribbon samples had approximately the same thickness and has thus been quenched at approximately the same cooling rates. According to the Cu–Hf equilibrium diagram, the crystallisation of this composition is probably due to the phases Cu₃Hf₂ and Cu₅Hf₂ [9].  The alloys with x = 30 - 35 at.% Hf, had values of T_rg > 0.60 and thus has been expected to form a fully glassy thickness to at maximum of 2 mm diameter.  However, attempts to produce fully glassy rods larger than 2 mm diameter were not successful, the XRD patterns showed crystalline structures for 3 and 4 mm diameter rods.  Duan et al. [10], have reported that the alloy Cu₆₆Hf₃₄ can be cast into 2 mm diameter rods and into 2 mm thick fully amorphous strip, this differs by only 1 at.% Hf from one of the present alloys in which we have not yet succeeded in casting as a fully glass 2 mm rod.  These results show that in binary systems the strong composition dependence of the liquidus temperature generally accounts for the relatively narrow composition range of the glass formability.

The addition of Ti to the binary alloys increased drastically the GFA observed experimentally in the bulk samples.  As was stated previously, the XRD results showed a fully amorphous structure for cast rods up to 3 mm diameter.  Coincidentally, the glassy 3 mm diameter alloys had the highest value of T_rg and the lowest liquidus temperature (T_l).  However, according to Davies [1] not only T_g/T_l has strong influence on the GFA, but also other factors such as the magnitude of the enthalpy difference between the liquid and the crystalline phases.  The large differences in chemical character and in the diameters of the constituent atomic species lead to high thermodynamic stability of the liquid phase which tends to promote a reduction in liquidus temperature and also the existence of crystalline phases that have relatively low stabilities.  Consequently, not only an eutectic reaction will occur at a low temperature but also, for a given undercooling below T_l, the driving force for crystallization (AG_v) will tend to be relatively small.  In any case, the dense and more ordered packing of atoms of unlike diameters and chemical characters, associated with high thermodynamic stability of the liquid state will tend to reduce their diffusivities which in effect will reduce T_g.  Moreover, the requirement, on crystallization, to reconstitute to intermetallic phases having complex and ordered structures would retard devitrification unless metastable crystalline phases having simpler structures intervened [1].

Although some bulk glass forming alloys are characterized by large values of (T_x-T_g) in this particular situation it does not appear to be a clear correlation in the present case between the experimental GFA, measured by the critical diameter of fully glassy rod, and the parameter ΔT_x (=(T_x-T_g )), as the Ti content is varied and this has also been concluded for other alloy series in this system [7].  Clearly, it is possible to increase T_rg by decreasing T_l.  This can be achieved by adding further solute atoms of various diameters to the Cu-Hf-Ti alloys.  This makes use of the “confusion principle” that can frustrate the tendency of the liquid phase to crystallise by stabilizing it thermodynamically relative to the crystalline phases [11].  The effect of fourth element additions on the GFA of the Cu-Hf-Ti system is highly pertinent in this context and will be investigated as a part of the future investigations in the present study.

Conclusions

• The binary Cu-Hf alloys series showed a fully amorphous structure for ribbons in the composition range of x = 30 to 50, however the binary alloy with x = 25 had a crystalline + amorphous structure.  The attempts to produce rods with glassy structure were not successful for diameters  ≥ 2 mm in the binary Cu-Hf alloys.
• The relatively strong composition dependence of the liquidus temperature may be a factor governing the GFA.
• Fully glassy rods with diameters of 3 mm have been produced for the x = 20 and 25 alloys and 2mm for the x = 15 and 30 alloys in the 55 at.% Cu ternary series.
• The composition dependence of the GFA’s is consistent with the trends in T_rg but not with the composition dependence of ΔT_x[=(T_x-T_g)]

References

1. H. A. Davies, Metallic Glass Formation Revisited, Proc. Of 4th int. Workshop on Non-Crys. Solids, 3 (1994), Spain, World Scientific.
2. R. W. Chochrane  and J. O. Strom-Olsen, eds, Proc 6th Int. Conf. Rapidly quenched metals, 3 (1989), Amsterdam, Elsevier.
3. K. Ozawa, H. Wakazuki and K. Tanaka, “Friction and Wear of Magnetic Heads and Amorphous Metal Sliding Against Magnetic Tapes”, IEEE Trans. Magn., 20 (1984) 425.
4. O. Kohmoto, K. Ohya and T. Ojima, “Wear-resistant Magnetic Head Using Amorphous Alloy Material”, IEEE Trans. Magn., 25 (1989) 4490.
5. H.W. Kui, A.L. Greer, and D. Turnbull, “Formation of Bulk Metallic Glass by Fluxing”, Appl. Phys. Lett., 45 (1984) 615.
6. A. Inoue, W. Zhang, T. Zhang and K. Kuroaska, “High-strength Cu-based Bulk Glassy Alloys in Cu_Zr_Ti and Cu_Hf_Ti Ternary Systems”, Acta Mater., 49 (2001) 2645.
7. A. Inoue, W. Zhang, T. Zhang and K. Kurosaka, “Cu-based Bulk Glassy Alloys with High Tensile Strength of over 2000 MPa”, J. Non-Cryst. Solids, 304 (2002) 200.
8. P. Stewart, 4th year MEng Final project, “Casting and properties of Copper bulk metallic glasses”, University of Sheffield, 2002.
9. Binary Alloy Phase Diagrams, 2nd ed., T. B.Massalski. ed.,ASM International, Metal Park, OH, 1990.
10. G. Duan, D. Xu and W. L. Johnson, “High Copper Content Bulk Glass Formation in Bimetallic Cu-Hf System”, Met. And Mat. Trans. A, 36A (2005) 455.
11. A.L. Greer, “Metallic glasses”, Science, 267 (1995) 1947