Brief Review on Advanced Metallic Materials

Advanced Nonequilibrium Materials, including

Bulk Glassy Alloys / Bulk Metallic Glasses

Crystallization leading to the formation of :

Metallic Nanomaterials

Metallic Glassy – (Quasi,Nano)Crystal Composites

Nanoquasicrystals

  Partly based on the recently published book: Metallic Glasses and Their Composites

Active research activities on metallic glassy alloys (or metallic glasses) started after the formation of the first Au-Si sample with an amorphous structure in 1960[1] by rapid solidification. This become possible by using a rapid solidification technique for casting of metallic liquids at a very high solidification rate of 10⁶ K/s[2]. For a long time Pd-Cu-Si and Pd-Ni-P were known to be the best metallic glass formers[3] . A large number of bulk glassy alloys (also called bulk metallic glasses) defined as 3-dimentional massive glassy (amorphous) articles with a size of not less than 1 mm in any dimension have been produced during the last 30 years. These alloys become widely studied in 90-th. The high glass-forming ability achieved at some alloy compositions has enabled the production of bulk metallic glasses in the thickness range of 1-100 mm by using various casting processes[4]^,[5],[6].

Production of Glassy Alloys and Their General Features

            Depending upon glass-forming ability (GFA) glassy (amorphous) alloys can be produced using various methods. Alloys having a low GFA can be prepared in an amorphous state (Fig. 1) by condensation from a vapor phase [7] ^{, [8]} . This method is, however, highly power consuming and not efficient. Various alloys can be produced by a solid state reaction using mechanical attrition [9] , in particular ball milling [10] or severe plastic deformation [11] ^{, [12]} . Electrodeposition from a solution is also possible [13] ^{, [14]} . These methods are efficient but also power consuming.

Much more productive is rapid (compared to conventional metallurgical methods) solidification of a liquid phase[15]^,[16] by melt-spinning, Cu-mold casting, liquid forging and so on. For example, Al-RE-TM (RE-rare earth metals, TM-transition metal) system glassy alloys[17]^,[18] were produced in a ribbon shape by the melt spinning technique[19] or as powder using gas atomization technique[20] as well as binary Al-RE alloys[21]^,[22]. An addition of Co partially replacing Y in the Al₈₅Y₁₀Ni₅ (here and elsewhere alloy compositions are given in nominal atomic percents (at.%)) increased tensile fracture strength without worsening of bend ductility[23]. Al₈₅Y₈Ni₅Co₂ alloy exhibits one of the largest supercooled liquid region on heating among the Al-based metallic glasses. A partial substitution of Y by Mischmetal (Mm), a natural mixture of the RE elements, leads to a drastic decrease in the alloy cost without significant deterioration of the mechanical properties[24],[25].

Although ease of devitrification of the Al-based glassy samples connected with a high density of so-called quenched-in nuclei in some glasses as well as a low reduced glass-transition and devitrification temperature in the other glasses[26], in general, impedes a limit on the sample’s critical thickness below 1 mm bulk amorphous samples of high relative density were still obtained by warm extrusion of atomized amorphous powder[27].

Bulk glassy alloys obtained in the late 80-th (except for Pd-Cu-Si and Pd-Ni-P obtained earlier^2,3) with an extraordinary high GFA^{[28],[29],[30]} have larger size (comparable with conventional crystalline alloys) of 1-100 mm. They can be obtained at the cooling rates of the order of 100, 10, 1 K/s and even less[31]^,[32] much lower than 10⁴-10⁶ K/s required for vitrification of marginal glass-formers. The bulk glassy alloys possess three common features summarized by Inoue[33], i.e., belong to multicomponent systems, have significant atomic size ratios of above 12% and exhibit negative heats of mixing among the constituent elements. Cu-based bulk glassy alloys wereobtained not only in the ternary Cu-(Zr or Hf)-Ti [34] but even in the binary Cu-(Zr or Hf)^{34, [35] , [36] , [37]} systems. An addition of the third element like Ti or Al, for example,³⁴ enhances the glass-forming ability of a binary alloy in accordance with the “confusion” principle which postulates that the multicomponent alloys in general posses better GFA than the binary alloys [38] .

Bulk glassy alloys exhibit high strength, hardness, wear resistance and large elastic deformation (Fig. 2) and high corrosion resistance , including passivation in some solutions[39],[40]. Pt-based bulk metallic glass was reported to exhibit a room-temperature strain of 20 % due to a high Poisson ratio of 0.42[41].

Bulk glassy alloys can be thermo-mechanically shaped or welded in the supercooled liquid region. Electromechanical shaping technology allows their rapid shaping by Joule heating at low applied stresses due to the high electrical resistivity of glassy alloys[42]^,[43]. Bonding of glassy alloys can also be achieved by friction welding[44]. Moreover, glassy alloys also exhibit superplasticity[45] including high-strain-rate superplasticity[46].

Glass-Forming Ability and General Properties

In general bulk glassy alloys are formed at the compositions with high T_g/T_l (T_g glass-transition temperature, T_l liquidus temperature) (Fig. 3) ratio exceeding approximately 0.6[47],[48]. At the same time, it has been shown that the width of the supercooled liquid region (dT_x) (defined as T_x-T_g where T_x is onset devitrification temperature) as indicator of the stability of the supercooled liquid against devitrification also correlates quite well with GFA . The parameter gamma=T_x/(T_g+T_l) takes into account both criteria[49] as high T_x and low T_g+T_l values leading to high gamma parameter indicate rather low T_l and high dT_x values.

The best glass-forming compositions are not at the equilibrium eutectic point but somewhat shifted usually towards more refractory component [50] , while T_g is not significantly different in the observed range. This is most likely due to the shift of the eutectic point with undercooling at high enough cooling rate as casting conditions of bulk glassy samples are far from the equilibrium conditions. This may be a result of deep undercooling or existence of the competing crystalline phases in the system. Both factors may cause a shift of the eutectic point.

However, the principles for achieving a good GFA known so far are rather indispensable conditions which sometimes, however, are not sufficient[51]. The addition of Nd caused formation of an amorphous single phase only in the Ge-Ni-Nd alloy while no amorphous phase was formed in the Si-Ni-Nd alloy. It was found that the higher GFA of the Ge-Ni-Nd alloy compared to the Si-Ni-Nd one cannot be explained on the basis of the widely used parameters, geometrical and chemical factors, viscosity and diffusion data. It was suggested that the electronic structure characteristics, for example electronegativity difference, should be taken into consideration⁵¹.

It has been shown that the electronegativity[52] of the constituent elements is an important factor influencing glass formation and the temperature interval of the supercooled liquid region of the glass-forming alloys[53],[54].

Packing density for non-crystalline structures, as a geometrical factor influencing GFA, has been verified using hard spheres model[55]^,[56]. A mixture of atoms with different sizes enables their more dense packing than can be achieved with separate phases. An importance of efficient atomic packing for the formation of metallic glasses was shown recently[57],[58]. It has been emphasized that specific radius ratios are preferred in the compositions of metallic glasses. This features are also closely connected with so called l criterion for good glass-forming ability[59]^,[60].

            It has been also supposed that electron concentration: number of valence electrons per number of atoms (e/a value) affect glass-forming ability[61]^,[62]. By other words good glass-formers have definite electron concentration values. This rule has been proposed by analogy with Hume–Rothery phases related to certain valence electron concentration. However, as many glassy alloys contain transition metals which have multiple valencies, it is difficult to decide which valency value should be taken into consideration in a particular case.

            The glass-transition phenomenon in metallic glasses has been studied extensively. Three kinds of approaches have been formulated (see[63]^,[64], for example, among the other sources): (1) glassy phase is just a frozen liquid, and thus, glass-transition is a kinetic phenomenon and no thermodynamic phase transformation takes place; (2) glass-transition may be a second-order transformation as follows from the shape of the curves for the thermodynamic parameters, for example, specific volume or enthalpy, which exhibit a continuity at the glass-transition temperature while their derivatives like dV/dT or dCp/dT exhibit a discontinuity (in a certain approximation) at the glass-transition temperature; (3) glass-transition may be a first-order transformation as follows from the free-volume model. Moreover, a thermodynamic aspect of glass-transition has been pointed out by Kauzmann [65] . A mode-coupling theory was also used to describe glass transition [66] .

Devitrification/Crystallization of Glassy Alloys Leading to Formation of a Nanostructure

Nanoscale particles of a crystalline or a quasicrystalline phase can be readily formed by devitrification of the glassy alloys. This indirect method of production of the nanostructure requires formation of the glassy phase in the initial stage and its subsequent full or partial devitrification on heating. Such a method leads to the formation of a highly homogeneous dispersion of nanoparticles in various alloys. Nevertheless, a nanostructure can also be obtained directly during rapid solidification or casting at a certain cooling rate used. However, in such a case it is often difficult to adjust an appropriate cooling rate and to obtain a homogeneous structure.

Matrix phase prior to devitrification can be amorphous, glassy or supercooled liquid. Although it might be difficult to establish an intrinsic physical difference between amorphous and glassy alloys such a slightly arbitrary differentiation is useful, especially in relation with devitrification behaviour. Here we define an alloy being “amorphous” if it does not transform to a supercooled liquid before devitrification (Fig. 3). In general glassy alloys exhibiting the supercooled liquid region on heating prior to devitrification have better glass-forming ability compared to amorphous alloys.

Alloys devitrifying from the supercooled liquid exhibit a tendency to form metastable phases and phases with high crystallographic symmetry on devitrification[67]. It may be connected with the change of the local atomic structure in the supercooled liquid region due to higher atomic mobility compared to that in the glassy phase. Below T_g the crystalline products of devitrification inherit the as-solidified structure of the metallic glass.

Four types of phase transformations were found to occur during devitrification of the glassy alloys: polymorphous (a product phase has the same composition as the glassy phase), primary (a product phase has a composition different from that of the glassy phase), eutectic devitrification (two or more phases nucleate and grow conjointly) and spinodal decomposition involving a phase separation of the glassy phase prior to devitrification.

If devitrification occurs by nucleation and growth mechanism (amorphous alloy does not have pre-existing nuclei), then high nucleation rate leading to high number density of the precipitates of the order of more than 10²¹ m^-3 and low growth rate of the precipitating phase are required in order to obtain a nanostructure. The nucleation and growth processes leading to the formation of a nanostructure, including transient nucleation have been described in detail in several earlier works[68]^,[69]. The kinetics of the devitrification process has been also analyzed[70]^,[71].

Devitrification of glassy alloys can be analyzed by Kolmogorov[72] -Johnson-Mehl[73] –Avrami[74] -Kelton[75] general exponential equation for the fraction transformed x(t):

where I and u are nucleation and growth rates, respectively, while n is the Avrami exponent.

Nanostructured alloys are readily obtained on primary devitrification of glasses with a long-range diffusion controlled growth[76].

Another type of phase transformation in an amorphous solid leading to formation of a nanostructure is spinodal decomposition[77]^,[78]. For instance, it has been reported that in some Al–TM–RE glasses devitrification leading to formation of a nanostructure appears to be preceded by the amorphous phase separation[79]^,[80].

Nevertheless, the most common mechanism leading to formation of a nanostructure is primary devitrification. Different Al-RE-TM glasses containing about 85 at.% Al show primary precipitation of the Al solid solution (a-Al) nanoparticles on heating [81] ^{, [82]} with extremely high nucleation rate exceeding 10²⁰ m^-3s^{-1 [83] , [84]} . After reaching a certain size a-Al particles change their morphology from spherical to dendritic [85] . Pressure effects on devitrification of the Al-RE-TM alloys have also been studied [86] .

fcc a-Al lattice parameter measurements and atom probe ion field microscopy investigation[87] showed very low concentration of the alloying elements in nanocrystalline Al in accordance with the phase diagrams[88] of Al-RE and Al-TM. Segregation of the RE metal having low trace diffusivity in Al on the a-Al/amorphous phase interface is considered to be one of the most important reasons for the low growth rate of a-Al. It is also important to note that the primary devitrification requires a long-range diffusion that impedes crystal growth[89].

Clear heterogeneous nucleation was observed during formation of the a-Fe nanocrystals[90]^,[91]. The structure of Fe-based soft magnetic alloys like: Finemet Fe_73.5Cu₁Nb₃Si_13.5B₉[92] and Nanoperm Fe₈₄Zr_3.5Nb_3.5B₈Cu₁[93] after annealing consists of bcc Fe nanocrystals below 20 nm in size finely dispersed in the amorphous matrix. Cu, Nb or Zr elements despite on their low contents are responsible for the formation of a nanostructure upon annealing. For example, devitrification of the Fe_73.5Cu₁Nb₃Si_13.5B₉ alloy, which initially has an isotropic amorphous structure, starts from the formation of Cu-enriched zones[94].

As has been shown by means of atom probe field ion microscopy as well as by high resolution transmission electron microscopy[95] Cu forms nano clusters in the Fe_73.5Si_13.5B₉Nb₃Cu₁ amorphous matrix which act as the sites for heterogeneous nucleation of the bcc Fe particles on devitrification[96]. The density of the clusters estimated by three dimensional atom probe is in the order of 10²⁴ m^-3 at the average cluster size of about 2-3 nm[97].

Not only pure metals and limited solid solutions but intermetallic compounds can be formed with a nanoscale size. For example, the devitrification of nanocrystal-forming Ti-based alloys, for example, the Ti₅₀Ni₂₀Cu₂₃Sn₇ alloy begins from the primary precipitation of a nanoscale equiaxed, almost spherical particles of cF96 Ti₂Ni solid solution phase (other alloying elements are partially dissolved in this phase) with a lattice parameter of 1.138 nm[98]^,[99]. Formation of such a nanoscale cF96 phase having a large cubic unit cell has also been observed in the Zr- and Hf-based alloys[100]^,[101].

An extremely small size and low growth rate of cF96 crystals were observed in Hf₅₅Co₂₅Al₂₀ glassy alloy. This alloy undergoes a double-stage devitrification forming the primary cubic cF96 Hf₂Co phase from the supercooled liquid by the steady state nucleation and diffusion-controlled growth of nuclei followed by the subsequent devitrification of the residual glassy matrix forming Hf₂Al and an unidentified phase. The cF96 nanoscale phase has a very low average growth rate of the order of 10^-11 m/s. Very small cF96 Hf₂Co clusters of 2-5 nm in size are observed in a HRTM image of the annealed sample (Fig. 4). Diffusional redistribution of Al within the glassy matrix may be the reason for the small growth rate of the clusters. It is interesting to note that although this alloy exhibits devitrification by nucleation and growth mechanism the local order of the cF96 clusters is close to that of the glassy phase[102].

In many cases redistribution of the alloying elements on a short scale precedes devitrification. For example, Mg-Ni-Mm and Mg-Ni-Y-Mm glassy alloys show a multistage devitrification behaviour[103].

The amorphous Ge-Cr-Al-RE alloys also exhibit local order changes prior to nanoscale devitrification. For example, some structure changes occur in the Ge₇₀Cr₁₆Al₁₀Nd₄ alloy prior to devitrification¹²⁴. Primary Ge nanoparticles are also formed in many Ge-Al-Cr-RE alloys[104]. Anomalous X-ray scattering studied shows that Ge-Ge atomic pair in Ge-Al-Cr alloy has a different coordination compared to metal-metal and Ge-metal pairs[105].

The as-solidified microstructure of the rapidly solidified Ge-Al-Cr-RE alloys examined by TEM showed the presence of finely dispersed zones homogeneously distributed in the amorphous matrix[106]. Their size is estimated to be about 1 nm[107]. These zones are responsible for the split of the first diffraction maximum in the SAED and XRD patterns. However, as the ordered zones are as small as 1 nm in size. An intermediate Ge-Ni-La compound which precipitates prior to the equilibrium phases in the Ge₆₀Ni₃₅La₅ amorphous alloy may inherit the local atomic order of the amorphous phase [108].

However, in many cases a nano-dispersed structure can be obtained directly from the melt upon rapid solidification by proper alloying. For example, the microstructure of the Si₄₈Al₂₀Fe₁₀Ge₇Ni₅Cr₅Zr₅ alloy having an amorphous type X-ray diffraction pattern is inhomogeneous and contains Ge particles of less than 5 nm in size embedded in an amorphous matrix (Fig. 5)[109]. Ge having a lower mixing enthalpy with the other alloying elements than Si is rejected from the amorphous matrix and precipitates forming fine nanoparticles, though it has an unlimited solubility in Si.

Here one should mention that the factors leading to nano-devitrification of metallic glasses and amorphous alloys in many cases are not fully understood while several reasons are listed in the literature. They can be connected with the occurrence of heterogeneities such as oxygen impurity-enriched clusters[110], spinodal decomposition in the liquid or glass[111] and homogeneous nucleation in partitioning systems[112].

The time-temperature-transformation diagrams[113] created in the isothermal mode or under continuos heating are useful for comparison of thermal stabilities of different glasses against devitrification as well as for the selection of the heat treatment regimes. Such diagrams have been created for different metallic glasses[114] [115] [116] [117] [118] [119].

Comparison of the long-term thermal stabilities of different metallic glasses has been done using continuos heating transformation CHT diagrams[120]. CHT diagrams have been constructed by applying a corollary from the Kissinger analysis method using the DSC data at different heating rates.

CHT diagrams also can be recalculated from the isothermal ones[121] using a method close to that used for steels[122]. This method was proved to provide a good correspondence between the data at least in the mid-temperature range.

            Not only thermal activation but plastic deformation can cause nanoscale devitrification of a glassy phase. For example, deformation of some Al-RE-TM amorphous alloys at room temperature causes precipitation of deformation-induced Al particles of 7-10 nm in diameter within the shear bands on bending[123] or nano-indentation[124]. It has been suggested that a local temperature rise can play a role in mechanically-induced devitrification[125]. However, further studies on nano-indentation of glassy alloys at low strain rates also showed the formation of the nanoparticles when local rise of the temperature due to adiabatic heating is not supposed to take place[126]. This postulate has been confirmed by bending test at low temperature using dry ice[127].

Nano-devitrification leading to formation of quasicrystals

An icosahedral quasicrystalline phase having a long-range quasiperiodic translational and an icosahedral orientational order, but with no three-dimensional translational periodicity was initially discovered in Al-Mn alloys[128] and later in the other binary Al-TM- based alloys containing Cr[129] and V[130]^,[131] and different ternary alloys[132]. The Al-TM base icosahedral structure has been presumed[133] to consist of Mackay icosahedral clusters. Later the icosahedral phase was observed in Ga-, Ti-, Mg- and Pd-based alloys ([134], for example) as well as Cd-[135], Rare Earth-[136] and Zn-based[137] alloys.

Quasicrystals can coexist with rational approximant crystalline structures[138]. A multiple twinning around an irrational axis of the approximants has been reported in an aggregate of fine size cubic crystallites[139].

A nanomaterial consisting of almost spherical icosahedral particles with a diameter below 10 nm was obtained in Zr-Pt alloy[140] by casting. An enhancement of the icosahedral short-range order in the supercooled Ti-Zr-Ni liquids decreases the barrier for the nucleation of the metastable icosahedral phase, even in the alloys in which the stable crystalline phases are formed according to the equilibrium phase diagram[141]^,[142]. Reduced supercooling (undercooling) before crystallization from the melt was found to be the smallest for quasicrystals, larger for crystal approximants and the largest for crystal phases[143]. A low energy barrier for nucleation of the icosahedral phase may explain the fact that only growth of the pre-existed icosahedral nuclei was observed in the Zr₆₅Ni₁₀Al_7.5Cu_7.5Ti₁₀Ta₁₀ alloy [144] .

Formation of the nanoscale icosahedral phase was observed in the devitrified Zr-Cu-Al, Zr-Al-Ni-Cu[145] and Zr-Ti-Ni-Cu-Al[146] glassy alloys containing an impurity of oxygen above about 1800 mass ppm, although no icosahedral phase is formed if oxygen content is lower than 1700 mass ppm. Recently, the nanoscale icosahedral phase was obtained in devitrified Zr-Al-Ni-Cu-Ag[147], Zr-Al-Ni-Cu-Pd[148], Zr-Ni-NM[149] (NM-noble metals), Ti-Zr-Ni-Cu[150], Zr-Al-Ni-Cu-(V, Nb or Ta)[151], Zr-Pd[152] and Zr-Pt[153]^,[154] system alloys at much lower (typically about 800 mass ppm) oxygen content. The icosahedral phase is also formed in the Zr₇₀Cu₂₉Pd₁ alloy at the very low Pd content[155].

It is also noted that the icosahedral phase precipitates from the supercooled liquid in Zr-based amorphous alloys containing the NM elements. The nanoscale icosahedral phase has been produced recently in the NM-free Zr-Cu-Ti-Ni[156] and Zr-Al-Ni-Cu glassy alloys with low oxygen content below 500 mass ppm.

The amorphous (glassy) -> icosahedral phase transformation, for example in a Zr₆₅Al_7.5Ni₁₀Cu_7.5Ag₁₀ alloy, is not a polymorphous but rather a primary type reaction[157]. This explains the observed low growth rates.

The local atomic structures around Pt as well as Zr in the amorphous and quasicrystalline Zr₇₀Al₆Ni₁₀Pt₁₄ alloys have been determined by the anomalous X-ray scattering method[158]. Devitrification of several other Zr-based bulk metallic glasses has been also studied in situ using a high energy monochromatic synchrotron beam diffraction in transmission on heating[159].

Transformation from glassy+beta-Zr to glassy+icosahedral structure was observed in Zr₆₅Ni₁₀Al_7.5Cu_7.5Ti₅Nb₅ alloy on heating by a single-stage transformation with diffusion control. b-Zr solid solution particles were found to dissolve in the glassy phase while the nanoscale particles of the icosahedral phase precipitate after the completion of the first exothermic reaction[160], which is considered to be a single-type reaction in a certain sense similar to peritectic one.

In the rapidly solidified Ti_xZr_yHf_zNi₂₀ system alloys the nanoscale icosahedral phase is formed in the composition ranges close to the cI2 beta solid solution phase and complex cF96 phase formation ranges[161]. Moreover, cI2 b solid solution and icosahedral phases were found to have close stochiometric compositions. This is not a surprising result as late-transition metals (LTMs) have a high solubility in the cI2 b phase of early-transition metals (ETMs).

The nanoscale icosahedral quasicrystalline phase has been also produced upon heating glassy Hf₆₅Pd_17.5Ni₁₀Al_7.5[162], Hf₆₅Al_7.5Ni₁₀Cu_12.5Pd₅[163], Hf₅₉Ni₈Cu₂₀Al₁₀Ti₃[164] and Hf₆₅Au_17.5(Ni,Cu)₁₀Al_7.5[165] alloys (here and elsewhere in the paper the (Ni,Cu) symbol means Ni or Cu). Formation of the nanoscale icosahedral phase has been also found at an early devitrification stage of a binary Hf₇₃Pd₂₇ alloy[166]. According to the quasilattice constant value, the icosahedral phase in the Hf₆₅Pd_17.5Ni₁₀Al_7.5, Hf₆₅Au_17.5Ni₁₀Al_7.5 and Hf₆₅Au_17.5Cu₁₀Al_7.5 alloys consists of 137-atom Bergmann rhombic triacontahedra.

However, Hf-based alloys have a higher tendency to form a cubic cF96 phase compared to Zr-based ones. The alloys in the systems in which an equilibrium Hf-based cF96 phase exists do not show the formation of the icosahedral phase from the amorphous matrix[167].

The formation of the nanoscale icosahedral phase was recently observed in the Cu-based alloys alloyed with Pd[168] and Au[169] while Ag- and Pt-bearing alloys did not form the icosahedral phase[170]. Replacing 5-10 at.% Cu with Pd in the Cu₆₀Zr₃₀Ti₁₀ glass-former changes its devitrification pathway[171]^,[172],[173], inducing nucleation and diffusion-controlled growth of a nanoicosahedral phase of about 3-10 nm in size (Fig. 6) consisting of rhombic triacontahedra from the supercooled liquid region in the initial stage of the devitrification process. A single initial devitrification reaction takes place in the Cu₅₅Zr₃₀Ti₁₀Pd₅ alloy while in the Cu₅₀Zr₃₀Ti₁₀Pd₁₀ alloy a combination of two reactions forming the icosahedral and crystalline phase takes place.

The addition of 5 at.% Pd expands significantly the supercooled liquid region in Cu-Zr-Ti alloy[174]. Cu-Zr-Ti[175] alloy contains nanoscale crystalline particles (about 5 nm size) in as-solidified state whereas the alloy containing 5 at.% Pd is amorphous¹⁷⁴.

Nanoscale particles of the cP2 CuZr phase were observed in the as-solidified Cu₅₀Zr₃₀Ti₁₀Pd₁₀ alloy. The dissolution of the CuZr nanoparticles took place on heating up to supercooled liquid region owing to the instability of the CuZr phase below 988 K^[176]. According to the Cu-Zr phase diagram, the CuZr phase undergoes eutectoid transformation at 988 K which is above the supercooled liquid region of the Cu₅₀Zr₃₀Ti₁₀Pd₁₀ alloy (about 750-800 K). The nanoscale CuZr phase becomes thermodynamically unstable and dissolves in the supercooled liquid when atomic diffusivity is enhanced by temperature. One of the reasons is that the energy to form an eutectoid interface in nanoparticles is higher than the energy for dissolution of the nanocrystals.

High-Strength Nanocomposites Produced by Devitrification

            Devitrification of the glassy alloys is the way of production of composite nanomaterials. For example, Zr₆₅Al_7.5Cu_7.5Ni₁₀Pd₁₀ alloy having nanoscale icosahedral phase particles embedded in the glassy matrix shows better combination of the mechanical properties compared to the as-cast glassy sample [177] . At the same time, it is well known that stoichiometric icosahedral phase-based alloys have high hardness and are extremely brittle [178] . It is difficult to consider that the icosahedral phase itself in the present alloys has plastic deformability. The good mechanical properties are attributed to the existence of the residual intergranular glassy phase, though the volume fraction of the glassy phase is estimated to be less than 5%. The icosahedral particles can act as a resistant medium against the shear deformation.

Nanocrystalline precipitates increase the room temperature mechanical strength of the Zr-Al-Cu-Pd[179], Zr-Al-Cu-Pd-Fe[180] and (Zr/Ti)-Cu-Al-Ni[181] bulk alloys produced by crystallization of cast glassy samples. The deformation behavior of the partially crystallized Zr-based glassy alloys consisting of a metallic glassy matrix with crystallite precipitates of different shape and size has been studied recently[182].

            Some bulk glassy-crystal composites with enhanced ductility have been produced by proper alloying with the elements having zero or positive mixing enthalpy with main component, for example, in Zr-[183]^,[184] and Cu-based alloys [185]^,[186].

Al-RE-TM amorphous alloys posses a high tensile strength exceeding 1200 MPa[187] and good bend ductility [188]^,[189].  The homogeneous dispersion of the nanoscale fcc a-Al particles in an amorphous matrix causes a drastic increase in tensile fracture strength to 1560 MPa[190]^,[191] . These particles can be formed by controlling the cooling rate upon solidification or by annealing glassy alloys. An optimum strength value was obtained when the volume fraction of a-Al phase reached 25%[192]^,[193]. The significant decrease in tensile fracture strength by the further increase in V_f is due to the embrittlement of the remaining amorphous phase by the progress of structural relaxation and the enrichment in the solute elements[194]^,[195].

However, in some cases formation of the primary Al particles was found to deteriorate mechanical properties. Partial substitution of Ni by Cu in the Al₈₅Y₈Ni₅Co₂ metallic glass caused formation of the nanoscale Al particles. However, the tensile strength and hardness values drastically decreased[196]. Cu has much lower absolute value of heat of mixing with Al, Y and Co than Ni[197] that leads to decrement of the interatomic constraint force. Thus, Cu may weaken the interaction needed for the stability of the glass, thus resulting in the disappearance of T_g and precipitation of Al nanocrystals. In addition, the volume fraction of the Al nanocrystals in Al₈₅Y₈Ni₃Co₂Cu₂, for example, is much lower than that in the primarily devitrified Al₈₅Y₈Ni₅Co₂ metallic glass[198]^,[199].

Difference in devitrification pathways of amorphous alloys depending on heating rate, temperature and other conditions

Al₈₅RE₈Ni₅Co₂ glassy alloys showed precipitation of the a-Al nanoparticles after continuous heating using DSC at high enough heating rates (0.67 K/s and higher) or isothermal annealing at the temperature above T_g[200]^,[201]. At the same time, Y-, Gd- and Dy-bearing metallic glasses[202] as well as the Al₈₅Y₄Nd₄Ni₅Co₂[203] showed simultaneous formation of the intermetallic compound(s) and a-Al nanoparticles, or primary formation of the intermetallic compound after annealing below T_g. For example, the Al₈₅Y₈Ni₅Co₂ alloy shows formation of an unknown intermetallic compound conjointly with a-Al nanoparticles after isothermal annealing up to the completion of the primary phase transformation below T_g (Fig. 7). The intermetallic compounds are metastable and have a multicomponent composition. The volume fraction of the intermetallic compound is higher than that of a-Al and the fraction of a-Al depends on annealing temperature below T_g.

Two crystalline phases: the equiaxed nanoscale Ge particles with the size of about 3-20 nm and relatively large dendrites of a multicomponent Ge₁₈Cr₄AlCeSm crystalline phase of about 0.5 mm in size precipitate simultaneously in the Ge₆₈Cr₁₄Al₁₀Ce₄Sm₄ amorphous alloy on heating by a single-type exothermic reaction[204]. At the same time no eutectic-type coupled growth was observed. Ge and Ge₁₈Cr₄AlCeSm phases in many cases have no common interface and are separated by the amorphous matrix. Moreover, the size and the volume fraction of the multicomponent phase depend on the heating rate whereas the size and distribution of the Ge particles are independent of it. It can be suggested that Ge nanoparticles grew from the Ge–enriched zones formed during structural relaxation.

Nanocomposites produced in-situ by rapid solidification.

The Al-Y-Ni-Co-Pd ribbon samples have been produced by the melt-spinning technique. The addition of Pd to Al-Y-Ni-Co alloys caused disappearance of the supercooled liquid region as well as the formation of the highly dispersed primary a-Al nanoparticles about 3-7 nm in size homogeneously embedded in the glassy matrix upon solidification (Fig. 8). An extremely high density of precipitates of the order of 10²⁴ m^-3 is obtained. The resulted d-spacings of a-Al in Al₈₅Y₄Ni₅Co₂Pd₄ correspond very well to that of FCC Al.

The first direct observation of micro-strain and dislocations quenched in nanoparticles with a size below 7 nm was provided.

Summary

A large number of metallic glassy alloys undergo nano-devitrification on heating leading to the formation of nanoscale crystalline or quasicrystalline particles. In some alloys this effect leads to the formation of the nanocomposites with enhanced mechanical properties compared to fully glassy and fully crystalline alloys. Such alloys are being extensively studied at present and are believed to be important structural materials in the future.

References:

Back to Main Menu