Amorphous metal membranes

Amorphous metal membranes

9 Amorphous metal membranes A. Paolone,1 D. Chandra2 1 Consiglio Nazionale delle Ricerche, Istituto dei Sistemi Complessi, U.O.S. La Sapienza, Roma, ...

946KB Sizes 0 Downloads 0 Views

9 Amorphous metal membranes A. Paolone,1 D. Chandra2 1

Consiglio Nazionale delle Ricerche, Istituto dei Sistemi Complessi, U.O.S. La Sapienza, Roma, Italy; 2Department of Chemical and Materials Engineering, University of Nevada, Reno, NV, United States

Introduction The Pd and Pd alloys are considered to be the benchmark active materials for hydrogen separation and purification [1,2]. However, Pd is an expensive (wV 37/g1) and strategic material. Therefore, alternative materials are under consideration for its replacement. Among the pioneering reports, Hara et al. [3] proposed noble-metal-free hydrogen-permeable amorphous metal membranes, w30 mm thick, with a nominal composition Zr36Ni64 that achieved a practical level of hydrogen permeation rate higher than 1 cm3 H2(STP)/cm2 min, even without the use of any noble metal coating [3e5]. In the last 20 years, the amorphous alloys mainly composed of Ni, Nb, and Zr have proven to possess a permeability comparable or even higher than that of palladium [6,7]. In this review, we will report the findings of investigation of the main physical properties of amorphous membranes that are potentially useful for hydrogen purification and separation, from the studies conducted in the last 20 years. The review of Hu et al. [8] reports a comprehensive explanation of different amorphous material synthesis; further details about the synthesis of the amorphous materials for hydrogen purification are reported in Refs. [2,6]. Due to these excellent works, we will not cover this issue in the present review. However, we will devote the present review to the most important physical properties of samples for this kind of applications. The amorphous state of these materials is an out-of-equilibrium phase, usually obtained by rapid quenching of the alloys. The amorphicity, however, seems to be the critical parameter to obtain the hydrogen permeability useful for applications. Thus, crystallization should be avoided at operating temperatures that should be well below Current Trends and Future Developments on (Bio-) Membranes. © 2020 Elsevier Inc. All rights reserved.



Chapter 9 Amorphous metal membranes

the crystallization temperatures (see Section 2). The hydrogen permeability of any material can be conceptually written as the product of the solubility and diffusivity. Section 3 reports an overview of the solubility properties, while Section 4 illustrates the permeability of different amorphous alloys. Section 5 is devoted to the microstructure of these materials which is intriguing also from a fundamental point of view; due to the lack of long-range order typical of the crystalline state, it poses challenges for its investigation. Section 6 discusses the mechanical properties of amorphous membranes that are known to be less sensitive to hydrogen embrittlement than their crystalline counterparts. Finally, the developments of computational methods inherent to the structure or the properties of these local atomically ordered compounds are reviewed in Section 7.

Crystallization of amorphous membranes A notable review of the studies about the crystallization in amorphous membranes and the effect of hydrogen on this process can be found in Ref. [9], including works published before 1999. A novel impulse to these studies was given by the pioneering work of Hara et al. [3e5] on the amorphous membranes for hydrogen purification. Hara et al. [4,5] reported that the crystallization process, measured by differential scanning calorimetry (DSC), gives rise to at least two exothermic peaks, suggesting the occurrence of a first stage of nanocrystallization at lower temperatures and a subsequent bulk crystallization at higher T, Tc. They extended their studies to (Zr36Ni64)1-a(Hf36Ni64)a and in (Zr36Ni64)1-a (Ti39Ni61)a alloys and reported that the DSC peaks are visible above 800 K. However, in the Hf-containing alloy, the crystallization process shifts toward higher temperatures, while in the Ticontaining alloy with a > 0.6, the Tc tends to shift toward lower temperatures [4]. The crystallization process was studied in detail for the Nb20Ti40Ni40 alloy membrane by Ishikawa et al. [10] by means of DSC and X-ray diffraction (XRD). Five exothermic peaks were reported between 800 and 1050 K. The first two peaks did not induce changes in the XRD patterns and the material continued to show an amorphous structure. The other exothermic peaks in the DSC pattern corresponded to crystallization of various phases: B2 Tie Ni, body-centered cubic (BCC) (Nb, Ti) and Ti2Ni [10]. Also the Nb30Ti35Ni35 alloy membrane displayed multiple peaks in the DSC curve between 700 and 1000 K [11]. In the (Ti45Zr16Be20Cu10Ni9)100-xNbx alloy, the crystallization is

Chapter 9 Amorphous metal membranes

witnessed by two DSC peaks above 400 C that shift to higher temperatures as the Nb content, x, increases [12]. Paglieri et al. [13] investigated the dependence of the crystallization temperature of a series of amorphous ribbons produced by melt spinning with chemical composition (Ni0.6Nb0.4)100xZrx and (Ni0.6Nb0.3Ta0.1)100xZrx (where x ¼ 0, 10, 20 or 30) from the Zr content. In the case of the membranes without Ta, the Tc decreases from w650 C for x ¼ 0 to w520 C for x ¼ 0.3. Similar results were obtained for Ta-containing ribbons, except for the x ¼ 0; indeed, Ni0.6Nb0.3Ta0.1 displays a Tc much higher than the Tafree counterpart by w100  C [13]. For compositions around x ¼ 20 and 30, Palumbo et al. [14] reported a limited variation of the crystallization temperature with the Ta content, while they observed a strong dependence on the Zr concentration [15]. In all cases, more than one peak in DSC curves were observed for the crystallization, suggesting a multistep process. Kim et al. [16] studied the crystallization of Ni63.7Zr36.3 and Ni30Zr70 under both isothermal and nonisothermal conditions. In isothermal experiments, the Johnson-Mehl-Avrami (JMA) equation was applied to derive the Avrami exponent, n. For Ni63.7Zr36.3 alloy, the value of the Avrami exponent n > 2.5 suggested that the growth occurs by small particles with an increasing nucleation rate. On the other hand, 1.5 < n < 2.5 for Ni30Zr70 alloy indicated growth of small particles with a decreasing nucleation rate [16]. In both cases, crystallization was governed by three dimensional growths. From the same experiments, it was found that the activation energy (Eact ¼ 411.9 kJ mol1) of the Ni63.7Zr36.3 alloy is higher than that of the Ni30Zr70 alloy (Eact ¼ 383.7 kJ mol1), implying easier crystallization transformation of Ni30Zr70 alloy than that of Ni63.7Zr36.3 alloy. In nonisothermal conditions, Kissinger and Ozawa models were used to derive the activation energy of the crystallization of (Ni0.6Nb0.4)100xZrx (0  x  0.3) that ranged between w530 and w660 kJ mol1 (see Fig. 9.1) [16]. It must be noted, however, that the various DSC peaks correspond to the steps of the crystallization process that have different activation energies. Indeed, Palumbo et al. [14] reported that (Ni0.6Nb0.3Ta0.1)80Zr20 has three DSC peaks with activation energies between 400 and 580 kJ mol1. In a Ni32Nb28Zr30Cu10 membrane, the Kissinger plot for the three peaks gave activation energies of 370  30, 441  2, and 470  70 kJ mol1 [17]. Only in the (Ni0.6Nb0.4)90Zr10 and (Ni0.42Nb0.28Zr0.30)98B2 [15,17], the crystallization occurred in a single step. Ding et al. [18] produced Nb42Ni40Co18-xZrx (x ¼ 0, 4, 12) and Nb42Ni32Co6Zr12M8 (M ¼ Ta, Ti, Zr) amorphous alloy ribbons by the melt-spinning technique. These amorphous alloys



Chapter 9 Amorphous metal membranes

Figure 9.1 Kissinger and Ozawa plots of amorphous (Ni0.6Nb0.4)100-x Zrx alloys, where x ¼ 0, 10, 20, or 30 at% [16].

crystallization temperatures exceeded 850 K. It has been found that the Tc of the Nb42Ni40Co18 alloy decreased with Zr addition, while hydrogen permeability increased linearly with Zr content [18]. The effect of coaddition to Nb-based alloys was investigated also by Kim et al. [19] in Ni45-xCoxNb30Zr25 (x ¼ 0, 7.5 and 15). The first crystallization peak in DSC does not change its position with x, while the maximum of the second peak shifts to higher temperatures, as x increases. In the meanwhile, the activation energies of the crystallization reactions decrease with increasing x [19]. Lai et al. [20] studied 45-mm thick splat-quenched Ni60Nb35M5 (M ¼ Sn, Ti and Zr) amorphous metallic membranes and found a decrease of the glass transition temperature, Tg, by w10 K, after hydrogenation. Whereas, the crystallization enthalpy reported increased in the samples containing Sn and Ti [20]. These authors hypothesized that the observed decrease in measured Tg is the result of hydrogen occupying free volume within the alloy membrane. Nayebossadri [21] produced 3e6 mm thick Zr30Cu57.5Y12.5, Zr54Cu46 and Zr40.5Ni59.5 ribbons by closed field unbalanced magnetron sputter ion plating. The crystallization temperature is reported as lower than those of similar samples produced by other methods. Moreover, the DSC traces show a single exothermic peak which may correspond to a different crystallization mechanism/path in the deposited and melt-spun alloys [21]. The Zr40.5Ni59.5, Zr56.2Cu43.8, Zr30Cu57.7Y12.3, and Zr32Cu57.3Ti10.7 fabricated by the magnetron sputtering method exhibit stable

Chapter 9 Amorphous metal membranes

thermal properties above 400  C, under an inert atmosphere. Nevertheless, their thermal stability is reduced by more than 150  C under hydrogen [22]. Conversely, Dandana et al. [23] reported that in the Zr57Al10Cu15.4Ni12.6Nb5 alloy, the thermal stability improved by hydrogen absorption, preventing crystallization. With slight addition of hydrogen, the onset crystallization temperature increased by 6 degreese10 C. Lai et al. [24] reported synthesizing amorphous CueZr alloys with Zr ¼ 37, 54, 60 at% by splat quenching and showed a decrease in Tg and Tc with increased concentration of Zr. However, the glass transition temperatures reported did not change after Pd deposition and annealing. After annealing the membranes at a temperature lower than Tg by 11 K, the structure of the membranes changed slightly, whereas there were no changes observed after annealing at lower Tc [24].

Hydrogen solubility in amorphous membranes General aspects of interactions of hydrogen with amorphous materials were already known for over 20 years (see the review of Eliaz et al. [9]). One typical method to measure the solubility of hydrogen in materials is by means of pressureecomposition curves obtained under isothermal conditions. These curves describe the relation between the content of hydrogen in the solid sample and the equilibrium pressure between the hydrogen gas and the solid. In crystalline materials, the chemical potential of hydrogen (or its pressure) remains constant within the twophase region, although the total hydrogen concentration increases. Therefore, the progressive transformation of the a-solid solution into the b-hydride phase occurs at almost constant pressure, and one usually observes the so-called “pressure plateau” in crystalline materials [9,25]. On the contrary, amorphous materials do not show constant pressure plateau and their pressuree composition isotherms consist in monotonously increasing concave curves, as evidenced in various studies [9,10,14,15,17,26e28]. These curves usually show deviations [7,9,14,15,17] from the empirical Sieverts’ law which is valid for crystalline metals at sufficiently low pressures:   c¼K

 pffiffiffi p ¼ K0 e

Ek= RT

pffiffiffi p

where, c is the atomic hydrogen concentration in the solid, K is the hydrogen solubility in the material, p is the hydrogen pressure in the gas, K0 is the solubility constant, Ek is the activation energy of



Chapter 9 Amorphous metal membranes

solution, R is the constant of ideal gases, and T is the absolute temperature. In general, the solubility of hydrogen in amorphous membranes decreases as temperature increases [6,9,10,14,15,17,26,28]. In the case of (Ni0.6Nb0.4)70Zr30 at a fixed pressure of 14 bar, one obtains H/Mw0.52 for T ¼ 300  C and H/Mw0.42 at T ¼ 400  C [14]. Another aspect that is common to all amorphous membranes investigated so far is the dependence of the hydrogenation energy on the hydrogen concentration in the solid samples [9,14,15,17]. Indeed, it was observed that as the hydrogen content in the solid sample increases, the hydrogenation enthalpy, DHhyd, decreases. For example, in (Ni0.42Nb0.28Zr0.30)98B2, the DHhyd decreases from 89  9 kJ mol1 H2 at H/M ¼ 0.10 to 43  3 kJ mol1 H2 at H/M ¼ 0.20 [15]. The same effect was also reported for (Ni0.6Nb0.4-yTay)100-xZrx with y ¼ 0, 0.1 and x ¼ 20, 30 [14]. This effect, observed about 40 years ago in amorphous samples with composition Pd77.5Cu6Si16.5 and Ni49.9Pd31.8P18.3, was explained in accordance with the Kirchheim’s model [29e31]. Accordingly, different interstitial sites are available for hydrogen trapping during the progression of the absorption process: at the beginning, the deepest energy levels are occupied (higher hydrogenation enthalpy), while only shallower energy levels (lower hydrogenation enthalpy) are available at higher hydrogen content [29]. Jayalakshmi et al. [32] obtained similar results from studies conducted by thermal desorption spectroscopy in some NieNbeZreTa amorphous ribbons. They showed that the dehydrogenation occurred at T > 500 C in samples containing <40 at%H, whereas at higher hydrogen concentrations, the dehydrogenation was initiated at 200  C [32]. As NieNbeZr membranes are among the most promising amorphous materials for hydrogen purification and separation, some systematic studies of the hydrogen solubility in these alloys above 300 C are available [6,14,15,17]. In (Ni0.6Nb0.4)100-xZrx, the hydrogen solubility increases as the Zr content, x, increases [6,14,15]. Substituting part of the Nb with Ta did not significantly influence the absorption properties [14]. In (Ni0.42Nb0.28Zr0.30)98B2, the absorption pressureecomposition isotherms could be measured only between 440 and 485 C as the kinetics of the boroncontaining membrane were very slow at T < 420  C [15]. On the contrary, the kinetics is much faster in Ni32Nb28Zr30Cu10 membranes, and the p-c isotherms could be measured between 158  C and 400  C [17]. An abrupt decrease of the hydrogen absorption was observed between 242 and 300 C. A combined X-ray diffraction study indicated that this behavior is due to an increase in the mean interatomic distance, from w2.31 Å at 200  C to

Chapter 9 Amorphous metal membranes

w2.38 Å around 300  C. This abrupt expansion was evidenced only when the samples were heated in a hydrogen atmosphere, whereas a much lower and nonabrupt thermal expansion was observed for samples heated in an inert atmosphere [15]. Hara et al. [33] obtained the pressureecomposition isotherms of Zr36xHfxNi64 between 473 and 573 K that showed the solubility of hydrogen decreased as the Hf concentration increases. From these measurements, the parameters of the hydrogen distribution sites as a function of temperature or the Hf composition could be obtained [33]. Quite recently, the occurrence of isotope effect was evidenced in Ni32Nb28Zr30Cu10 by means of hydrogen and deuterium solubility measurements [34]. The equilibrium pressure between the gas and the solid was reported to be higher for deuterium than for hydrogen, with a mean ratio P(D2)/P(H2)z1.5. Moreover, the absorption of H2 was faster than that of D2. Both properties closely resemble those of the PdeAg alloys that are currently used in membrane reactors and separators [35e37]. Some studies reported the hydrogenation of amorphous membranes by means of electrolytic charge in a 2:1 glycerinee phosphoric acid electrolyte at 25 C and at a current density of i ¼ 5e10 A m2. Some samples of amorphous and quasicrystalline Zr69.5Cu12Ni11Al7.5 alloys reached H/M z 0.8 and H/M z 1.4, respectively, after 70 h of charging [38]. Other examples of Zr41.2Ti13.8Cu12.5Ni10Be22.5 that reached H/M z 0.95 in 70 h have also been shown [39]. Higher hydrogen contents (up to 1.8 H/ M) were reached by Jayalakshmi et al. [40] in some NieNbebased amorphous alloys. In some cases, the hydrogenation or dehydrogenation of amorphous materials was enhanced by a surface coating of samples by either palladium or nickel; these are both hydrogen catalysts [41]. The samples plated with Pd show a faster kinetics than those with Ni plating [41]. The kinetics of hydrogenation was also investigated by Singh et al. [42] in (Zr69.5Al7.5Cu12Ni11)100 xTix (x ¼ 0, 4 and 12) in quasicrystaleglass composites.

Hydrogen permeability Hara et al. [3e5] were the first to propose a membrane for hydrogen separation made of nonprecious metals that could satisfy the following four requirements: high hydrogen solubility, high diffusivity, catalytic activity of the surface to dissociate hydrogen molecules into atoms, and mechanical strength in a hydrogen atmosphere These authors demonstrated that



Chapter 9 Amorphous metal membranes

amorphous Zr36Ni64 displayed a permeability at 623 K of 1.2  109 mol m1 s1 Pa1/2, with a practical level of permeation rate higher than 1 cm3 H2(STP)/cm2 min. This permeability value is comparable to that of Pd77Ag23. Before hydrogen permeation, an activation procedure was needed; it consisted in an exposure of both sides of the membrane to 0.3 MPa hydrogen for an hour at 573 K and a permeation measure at 653 K. The lack of such pretreatment led to the need of a long time to reach a stable permeation rate [3]. Permeability was found to decrease when Ti or Hf was added to the alloy, and this fact was attributed to the increased activation energy for permeation [4]. In some cases, Hara et al. [4,5] had to coat the surfaces of the membranes with a thin layer of Pd in order to increase the activity toward hydrogen. Hydrogen selectivity of at least 104 was reported, thanks to checks with helium gas [4]. Hara et al. [5] reported that the permeation rate was stable for a few days at 573 K or less and was approximately proportional to the square root difference of hydrogen partial pressures across the membranes. Soon after these first reports, Yamaura et al. [43,44] displayed that the addition of Nb to the original ZreNi alloy exhibited beneficial effects in terms of permeability, ductility of the materials, and resistance to embrittlement. Melt-spun (Ni0.6Nb0.4)100-xZrx (x ¼ 0 to 40 at%) and other amorphous alloy membranes with composition Ni45Nb45Zr10, Ni65Nb25Zr10 and Ni44Nb43Zr10Pd3, all coated with a thin layer of Pd, were examined and a maximum hydrogen permeability of 1.3108 (mol m1s1Pa1/2) at 673 K for the (Ni0.6Nb0.4)70Zr30 amorphous alloy was reported; this value is higher than that of pure Pd [43]. Moreover, these authors noticed that the hydrogen permeation rate is proportional to the difference of the square roots of pressures applied to the two sides of the membranes, and, therefore, the hydrogen diffusion though the amorphous solid is the rate-controlling factor for hydrogen permeation [43]. Possibly, at very low pressures, the rate-controlling mechanism is controlled by surface reactions [43]. In subsequent studies by Yamaura [44], the effect of small addition of various elements to (Ni0.6Nb0.4)45Zr50 X5 (X ¼ Al, Co, Cu, P, Pd, Si, Sn, Ta or Ti) was reported. The maximum values of hydrogen permeability were found in (Ni0.6Nb0.4)45Zr50Co5 and (Ni0.6Nb0.4)45Zr50Cu5 (2.46108 and 2.34108 mol m1s1Pa1/2 at 673 K, respectively). On the contrary, (Ni0.6Nb0.4)45Zr50P5 displayed the lowest permeability (1.36108 mol m1s1Pa1/2 at 673 K) among the alloys, due to a preferential development of ZreP atomic pairs [44]. The hydrogen permeability of the melt-spun Ni70-x/2Nb30-x/2Zrx (x ¼ 10, 20, 30, 40 and 60 at%) amorphous alloys increased with increasing Zr content in the ternary system [45]. However, it was

Chapter 9 Amorphous metal membranes

noticed that Nb plays a central role, because the binary system Ni40Zr60 has a lower permeability [45]. Moreover, the hydrogen permeation increased with increasing content of substituted elements; one can compare the (Ni0.6 X0.4)70Zr30 (X ¼ Ti, Hf and Nb) and the (Ni0.9X0.1)70Zr30 (X ¼ Y, Ti, Hf, V, Nb and Ta) amorphous alloys [45]. Shimpo et al. [46] evidenced that the diffusion occurring at 673 K of the Pd layer into the bulk of (Ni0.6Nb0.4)55Zr40Co5 and (Ni0.6Nb0.4)45Zr50Co5 amorphous alloys was detrimental for the time dependence of permeability. However, the same alloys could sustain a 100 h permeation test at 573 K, with only a slight decrease of permeability [46]. It must be noted that the amorphous ribbons with a width of 100 mm were produced for this work [46]. In Ni65Nb25Zr10 and Ni60Nb20Zr20 amorphous alloys, the hydrogen permeation was reported to increase significantly upon the transition from the glassy solid state to the supercooled liquid state [47]. However, the permeability decreased rapidly with time, when measured at higher temperature, because of the tendency of the amorphous alloys to crystallize [47]. Kim et al. [48] investigated amorphous Ni60Nb30Ta10 and reported a maximum permeability of 4.13  108 mol m11s1Pa1/ 2 at 673 K. Moreover, this alloy displayed a good resistance to hydrogen embrittlement. Alloys with composition Ni42Zr30Nb28xTax (x ¼ 0; 7; 14; 21; 28) were studied by Qiang et al. [49]. The addition of Ta did not deteriorate the permeation properties but slightly increases the resistance to crystallization [49]. Dolan et al. [50] reported the hydrogen permeability of noneutectic amorphous (Ni0.6Nb0.4)70Zr30 and (Ni0.6Nb0.3Ta0.1)70Zr30 membranes in pure hydrogen or in a simulated coal-derived synthesis gas, with high levels of CO, CO2, and H2O. They showed a degradation of permeability above 325 C, due to a partial crystallization of the alloys. On the contrary, the eutectic alloys are able to operate at higher temperatures [50]. The rate of degradation of the permeability was reported to be related to the Zr:Nb ratio; the lower contents of Zr led to slower degradation, except in the case of Ni64Zr36. This suggests that the optimal chemical composition of the membranes is a compromise between absolute permeability (that increases with Zr content) and durability (that decreases with Zr content). It was suggested that 10 at %


Chapter 9 Amorphous metal membranes

Similarly Paglieri et al. [13] investigated the permeability of (Ni0.6Nb0.4)70Zr30 and (Ni0.6Nb0.3Ta0.1)70Zr30 membranes above 400 C that showed a degradation over time due to the diffusion of the Pd-coating layer. Membranes derived by NieNbeZr alloys with various elemental composition were furtherly investigated by Refs. [16,18e20,37,52]. Fig. 9.2 [7] shows a comparison of the permeation values obtained for various amorphous membranes derived from the original Zr36Ni64 amorphous material proposed by Hara et al. [3]. Some other aspects of the same amorphous materials or novel amorphous alloys can be cited. Recently, Adibhatla et al. [53] suggested a novel method to enhance the activity toward hydrogen of the surface of NieNbeZr membranes, avoiding the necessity to use a thin Pd layer as a catalyst. These authors displayed that a thermal treatment at a temperature close to the crystallization temperature led to the development of nanoscale Ni particles at

Figure 9.2 Hydrogen permeation of several membranes of various nonprecious metals and alloys with different compositions. Permeation rates of Pd and Pd alloys are also included as reference plots [7].

Chapter 9 Amorphous metal membranes

the surface, resulting in a dramatic increase in the catalytic activity of the alloy surface to the dissociation and recombination of H2. Yamaura et al. [45] reported in their review some experiments about the hydrogen production by methanol steam reforming using melt-spun NieNbeTaeZreCo amorphous alloy membranes. Ishikawa [10] compared the permeability of amorphous and recrystallized Nb20Ti40Ni40. The latter compound has half the value of the amorphous one but is stable as a function of temperature and displays a good resistance to embrittlement. Similar results were obtained also in Nb30Ti35Ni35 and Nb40Ti30Ni30.[11] In Ni42Nb28Zr25Ta5 Chin et al. [54] observed that the crystalline membranes have superior permeability in comparison with the amorphous state; however, they were more subject to embrittlement. Geiller et al. [12] proposed (Ti45Zr16Be20Cu10Ni9)100-xNbx with x ¼ 0, 5, 10, 15 at% membranes composed of a dendritic crystalline phase rich in Nb and Ti that grows within an amorphous matrix and showed permeability even five times higher than that of Pd at T ¼ 350 C. Such high values of permeability were ascribed to the presence of the crystalline phases that act as a path for the fast diffusion of hydrogen; indeed, the hydrogen permeability was found to be proportional to the volume fraction of the dendritic phase [12]. Park et al. [55] investigated the hydrogen permeability of amorphous Cu50Zr50 and Cu65Zr35 membranes and found values between 1.5  and 2.0108 mol m11s1Pa1/2 between 150 and 310  C. Amorphous metallic membranes with composition CueZr (Zr ¼ 37, 54, 60 at%) were studied also by Lai et al. [24]. There is no significant difference in the permeabilities of uncoated and palladium-coated Cu63Zr37 and Cu46Zr54; however they were found much lower than that of Pd, possibly due to oxides developed on the ribbon surface [24]. Finally, Pd33Ni52Si15 amorphous alloy membrane was produced by Prochwicz et al. [56,57]. Hydrogen permeability ranges between 2.5 and 4.01010 mol m11s1Pa1/2 that unexpectedly decreased with the increasing thickness of the surface of Pdcoating layer. These authors argued that this phenomenon was due to an amorphous to polycrystalline transition of the membrane structure. The surface of the amorphous membrane provides hydrogen atoms with a larger number of active centers as compared to the partially ordered Pd layer. This leads to an increase in hydrogen concentration on the inner membrane layer [56].



Chapter 9 Amorphous metal membranes

Local atomic order Due to the amorphous nature of these type of membranes, X-ray diffraction measurements show only a broad peak centered around 40 degrees 2q, whose exact location can provide only the mean atomic distance in the sample. Therefore, more local probes, such as extended X-ray absorption fine structure (EXAFS) or pair distribution function, should be used to investigate the atomic order and the microstructure. Yamaura [27] investigated (Ni0.6Nb0.4)100xZrx (x ¼ 0, 30 and 50 at%) alloys both in the as-prepared and in the hydrogenated state, by radial distribution function analysis derived from XRD measurements. The main peaks of the scattering intensity profiles moved toward smaller Q for the hydrogenated (Ni0.6Nb0.4)70Zr30 and (Ni0.6Nb0.4)50Zr50 samples, indicating that the atomic distance increases by hydrogen absorption. Moreover, these authors suggested the growth of the short-range order induced by hydrogenation [27]. From the pair distribution functions, Yamaura et al. [27] were able to calculate the atomic distance between couples of atoms and the coordination numbers. The ZreZr atomic distance drastically increases with hydrogen absorption, and the coordination numbers of the NieNb and ZreZr atomic pairs are much larger than those of other pairs in the (Ni0.6Nb0.4)70Zr30 and (Ni0.6Nb0.4)50Zr50 amorphous alloys [27]. Oji et al. [58,59] investigated the EXFAS spectrum of Ni42Nb28Zr30 and Ni36Nb24Zr40, and their hydrogenated counterparts (Ni42Nb28Zr30)0.91H0.09 and (Ni36Nb24Zr40)0.89H0.11 obtained by electrolytic charging. The hydrogenation did not alter the local structure around the three metal atoms in Ni42Nb28Zr30, but in Ni36Nb24Zr40 the interatomic distances ZreZr, ZreNb and NbeNi elongated. Oji et al. [58,59] suggested the occurrence of two types of distorted icosahedral Zr5Ni5Nb3 structure in the sample. The coordination numbers obtained from EXAFS data are well explained considering a random assembling of these two structures. The EXAFS analysis, moreover, suggests that in Ni36Nb24Zr40, hydrogen resides in the tetrahedral ZreNb sites and in the sites near the Nb atoms and are surrounded by (Nb and Ni) or (Nb, Ni, and Zr) atoms [59]. Further support to the distorted icosahedral model was given by the XAFS investigation of as-prepared and hydrogenated (Ni0.6Nb0.4)0.65Zr0.35 reported by Matsuura et al. [60]. Fukuhara et al. [61] investigated the local structure of Ni42Nb28Zr30 and Ni36Nb24Zr40 glassy alloys by X-ray absorption near edge structure (XANES), these results will be reported in

Chapter 9 Amorphous metal membranes

the next section together with the density functional theory (DFT) calculations needed for their interpretation. Finally, Sarker et al. [62] very recently reported a combined study of (Ni0.60Nb0.40)70Zr30 using means of atom probe tomography and neutron scattering experiments, coupled with the DFTe based molecular dynamics. The atom probe tomography was used to investigate the local chemical composition, and it evidenced the occurrence of three-dimensional structures composed of Nb-rich (with a mean composition of Ni32Nb46Zr22) or Zr-rich (Ni39Nb18Zr44) nanometric clusters embedded in a ternary matrix whose compositions deviated from the nominal composition of the membrane [62]. Also the measurements of Sarker et al. [62] supported the distorted icosahedral model.

Mechanical properties The mechanical properties of amorphous materials are extremely important for their potential use, especially when they are exposed to a hydrogen atmosphere. It has long been recognized that hydrogen may embrittle many crystalline metals and alloys. In crystalline materials, several mechanisms for embrittlement have been proposed (see Eliaz et al. [9] for a review in which they report high-pressure bubble formation, reduction in surface energy [adsorption mechanism], reduction in the lattice cohesive force [decohesion mechanism], hydrogen interaction with dislocations, and hydride formation). It was already known 20 years ago that all mechanisms related to enhancement or inhibition of dislocations transport cannot act in amorphous alloys [9]. On the other hand, some different mechanisms, such as filling of free volume, might be responsible for the embrittlement of amorphous alloys. These, however, may be accompanied by traditional mechanisms such as high-pressure bubble formation and decohesion [9]. Yamaura et al. [63] reported that in Ti50Ni25Cu25 metallic glasses, the fracture strength decreases as the hydrogen content increases gradually up to w25 at% and more significantly for higher hydrogen content, becoming close to zero for H/M z 40%. These authors attributed the hydrogen embrittlement to the filling of Ti4-tetrahedral sites [63]. Examining the fractured samples with low H/M showed a vein-like structure, whereas, at higher H/M, cleavage-like structures were observed [63]. Kimura et al. [64] investigated the crystallization temperature, Tc, the Vickers hardness (Hv), and tensile fracture strength (sf) as a function of Zr content for the amorphous (Ni0.5Nb0.5)100-xZrx (x ¼ 10, 20, 30, and 40 at%) alloys. All the three physical quantities,



Chapter 9 Amorphous metal membranes

Tc, Hv, and sf increase almost linearly with decreasing Zr content and the highest values are obtained for the 10% Zr-containing alloy [64]. Du et al. [65] reported good hydrogen brittleness resistance of Pd71.5Cu12Si16.5 metallic glass ribbons. Kawashima et al. [66] reported that there was virtually no embrittlement of the (Ni0.6Nb0.4)70Zr30 alloy even at high hydrogen content. These authors reported that both the tensile fracture strength and the Vickers hardness increases with increasing H/M. However, these same authors noted that the morphology of the amorphous ribbons changed after electrolytic hydrogenation. In fact, the samples were found to have developed a curvature soon after hydrogenation, but after some time (864 ks) the original shape was restored [66]. Later on Horikawa et al. [67] observed the same bending curvature in Zr36Ni64 membrane with a thickness of 30 mm that suggested deformation was due to the gradient of the hydrogen concentration in the thickness direction of the membranes. On the contrary, Jayalakshmi et al. [68] reported some evidence of embrittlement of Zr50Ni27Nb18Co5 and Ni59Zr16Ti13Nb7Sn3Si2 that occurred only for H/M > 13 at%. In particular, these authors observed a reduction of the bending strain with increasing hydrogenation, obtained by electrolytic charging. The Ni-rich alloy was more prone to embrittlement than the Zr-rich. While XRD continues to see only an amorphous structure of the sample, TEM measurements and electron diffraction measurements indicate that at high hydrogen content, some small crystallites of g-ZrH or Ni2H form, with typical dimensions of 2e5 nm [68]. The leading mechanism for embrittlement was found to be the decohesion between metal atoms [68]. Jayalakshmi [40] also investigated the bending fracture of a large number of NieNbebased amorphous membranes. Among them, the alloys containing Ta (both ternary and quaternary) exhibit excellent resistance to embrittlement (up to H/M z 0.8), while the binary Ni60Nb40 alloy showed early embrittlement. These authors, as many others, observed a large variation of the mean interatomic distance upon hydrogenation and therefore, suggested again that the embrittlement is dominated by decohesion between transition metal atoms due to lattice dilatation [40]. Palumbo et al. [69] attributed a strong decrease of the tensile modulus around 473 K of (Ni0.6Nb0.4)1-xZrx membranes as Zr hydrides formed. Nishida et al. investigated the effect of annealing on the ductility and fracture toughness of Nb20Ti40Ni40. The asprepared melt-spun samples show bending ductility, but after annealing in the temperature range from 798 to 923 K, they

Chapter 9 Amorphous metal membranes

become brittle and grains with dimension of less than 50 nm develop. Annealing above 948 K for prolonged periods increases the grain size to above 150 nm, associated with the decrease in hardness and the ductility, while the fracture toughness is recovered [70]. Similarly, Ishikawa [10] pointed out that upon heating Nb20Ti40Ni40, the increase of the diameter of the BCC (Ti,Ni) phase led to a poor resistance to hydrogen embrittlement. Chin et al. [54] investigated Ni42Nb28Zr25Ta5 membranes and showed partial crystallization into a Ni10Zr7 matrix with enhanced permeabilities. However, they reported drastic decrease in their mechanical strength, so that all samples failed during the permeability tests. Singh et al. [42,71] studied quasicrystalline alloys (Zr69.5Al7.5Cu12Ni11)100xTix (x ¼ 0, 4 and 12). These alloys exhibited lower microhardness in the amorphous state than in the quasicrystalline state; hydrogenation only slightly changed it. Quite recently, nonoindentation measurements were performed on a series of NieNbeZr amorphous metallic ribbons [72]. For indentations into the uncharged samples, the nanoindentation hardness values, Hind, were estimated according to the OliverePharr method, that showed decreases in hardness as the Zr content increased, consistent with the increasing interatomic distance with Zr content [72]. In the hydrogenated state, Hind decreases in samples with low Zr contents, while it increases for higher concentrations of Zr. These authors explain this peculiar behavior as due to the different concentration of weakly bound mobile and strongly bound immobile states for hydrogen. The amount of the latter increases as the Zr content increases, and this fact explains the dependence of Hind on the amount of Zr [72]. Dandana et al. [73] measured hardness and elastic modulus by means of nanoindentation of uncharged and hydrogenated Zr57Al10Cu15.4Ni12.6Nb5 samples. After hydrogenation, both physical quantities decrease. Dandana et al. [73] attributed this to the formation of large void structures that facilitate the local atomic displacement. On the contrary, Lai et al. [20], who studied splat-quenched Ni60Nb35M5 (M ¼ Sn, Ti and Zr) amorphous metallic membranes by nanoindentation, observed changes in the mechanical properties induced by hydrogen. They [20] also observed that after hydrogen permeation testing, all compositions showed an increase in Young’s modulus and hardness. Lai et al. [20] examined samples after the permeation tests and reported that these effects are a result of a decrease in free volume due to the presence of atomic hydrogen and not due to the formation of crystalline hydrides.



Chapter 9 Amorphous metal membranes

Computational studies on amorphous membranes One of the first computational investigations of the structure of amorphous membranes was performed on Fe3B by Hao et al. [74]. To create a sample of amorphous Fe3B suitable for use in subsequent DFT calculations, a supercell containing 100 atoms was created by applying ab initio molecular dynamics to a liquid-like state at 1200 K, followed by a rapid temperature quench followed by energy minimization. The resulting amorphous structures were composed of Voronoi polyhedra, mostly trigonal prisms of type (0, 3, 6) and (0, 5, 4) for the boron environment and a large variety of icosahedra of Voronoi type (0, 0, 12), and distorted icosahedra around iron atoms [74]. Hao et al. [74] also investigated the potential interstitial sites for hydrogen and calculated the binding energies for each of them. Computationally, these authors could obtain the results already known from experimental studies: a broad range of binding energies was calculated, including a large number of sites that bind H much more favorably than the crystalline material. Finally, Hao et al. [74] calculated the hydrogen solubility in the amorphous materials by considering individual binding sites that exist with binding energies that are highly favorable relative to gaseous H2. The interactions among interstitial atoms could play a central role in the material’s net solubility due to appreciable concentrations of interstitial H. According to Hao et al.‘s [74] calculations, deuterium and tritium are expected to have a lower solubility than hydrogen in amorphous Fe3B due to the different zero-point energies of the hydrogen isotopes. Fukuhara et al. [61] investigated the structure of an amorphous hydrogenated Ni36Nb24Zr40 alloy by means of EXAFS. They compared the experimental data with those obtained with the calculations of the structure of various possible icosahedral Ni5Nb3Zr5 clusters that fairly reproduce the real composition. Fukuhura et al. [61] identified two stable and four metastable H sites, as reported in Fig. 9.3. In an alternative approach to the molecular dynamics used by Hao et al. [73], Fujima et al. [75] proposed a bottom-up approach. These authors [75] examined a series of 60 Ni5Zr5Nb3 isolated clusters (isomers) whose stoichiometry mimicked that of Ni36Zr40Nb24 and identified specific clusters as the building blocks of the amorphous structure. The optimized cluster has the following characteristics, (i) Ni-centered clusters have lower energies than Nb- and Zr-centered clusters because of the small

Chapter 9 Amorphous metal membranes

Figure 9.3 (A) Icosahedral structures of the NiNbZr alloys with the positions of the hydrogen atoms [61]. (B) coordination polyhedra with distances between the atoms [61].

atomic radius of Ni and (ii) the clusters with smaller NbeNi coordination numbers have lower energies [75]. Further, Fujima et al. [75] studied a system composed of eight clusters with different orientations and alignments and obtained two distinct structures: (i) an amorphous phase in which icosahedra maintain the initial shapes and (ii) a crystalline phase with a lower energy in which the icosahedra are distorted to form face-centered cubic (FCC)like cuboctahedra. The clustering of Nb and Ni atoms plays an important role to prevent the crystallization [75]. Kang et al. [76] reported a comprehensive review of the computational efforts in the field of the prediction of hydrogen flux through crystalline or amorphous membranes. They proposed prediction of flux as a function of operating conditions by DFT calculations. The solubility and diffusion coefficients of interstitial H are known provided that the H transport through the bulk



Chapter 9 Amorphous metal membranes

film is rate limiting. Once the energy of each interstitial site and a description of HeH repulsion are available, then the net solubility of H in the amorphous samples can be calculated as a function of temperature and H2 pressure using Grand Canonical Monte Carlo simulations [76]. In the case of amorphous Zr30(Ni0.6Nb0.4)70 at lowest H concentrations, Kang et al. [76] showed that the theoretical predictions for hydrogen solubility are in good agreement with the experimental data, whereas at higher H concentrations, the calculations systematically underestimated the solubility of H. Results that are in better agreement with experiments could be obtained considering the volume expansion of the amorphous sample due to hydrogenation [76]. The diffusivity of hydrogen is calculated not by means of the Fick’s law but more properly by means of the single-component MaxwelleStefan diffusivity, obtained using kinetics Monte Carlo simulations [76]. The most obvious feature is that the corrected diffusivity of H in these amorphous metals has strong concentration dependence due to broad distribution of interstitial site energies that exist for these materials [76]. Finally, Kang et al. [76] computed the expected flux across the amorphous membrane, and it turned out to be reasonably in agreement with experimental values. Following the same lines, the expected permeability of several amorphous alloys (Ta40Ni60, Ta25Ni60Ti15, Ti33Co67, Hf44Cu56, Hf25Cu60Ti15, Zr45Cu45Al10, Zr30Cu60Ti10, Zr54Cu46 and Nd60Fe30Al10) was calculated [77]. Although the permeability is dependent on both solubility and diffusivity, Hao et al. [77] reported that there is no simple correlation existed between the permeability and the solubility values for the investigated materials. Later on, the same authors pointed out that even though DFT calculations have been successful in applying quantitative methods to predict the performance of amorphous membranes, this approach is extremely time consuming [78]. Therefore, they reported a simpler method that used DFT calculations on smaller clusters. This allowed them to optimize the membrane performance with respect to their alloy composition and provided qualitative information similar to those obtained when larger clusters were considered as in their previous work [78]. All the results reported by Hao et al. [78] were obtained considering a Gaussian distribution of the site energy of the interstitial hydrogen atoms whose parameters were determined by the authors. This method was applied to a series of amorphous alloys of composition Zr30Cu60T10 (T ¼ Sc, Ti, Y, Nb, Mo, Tc, Ru, Rh, Pd, Ag, Ta, W, Re, Os, Ir, Pt, Au). The three alloys with highest predicted permeability were identified as Zr30Cu60T10, where T ¼ Sc, Ta, Y. They have similar permeabilities as those of Pd, for T > 600 K. However, Zr30Cu60Mo10 and

Chapter 9 Amorphous metal membranes

Zr30Cu60Ir10 have permeabilities that are 2e4 orders of magnitude lower than Pd [78]. A different approach for the prediction of the hydrogen permeability of amorphous membranes was proposed by Lee et al. [79]. This methodology is based on less time and resource-consuming calculations, performed by means of the combination of CALPHAD (CALculation of PHAse Diagram) type thermodynamic calculations for the hydrogen solubility and molecular dynamics simulations based on the second nearestneighbor modified embedded-atom method interatomic potential for the calculation of hydrogen diffusivity [79]. This type of methodology was used for the prediction of hydrogen permeability in amorphous Cu50Zr50 and Cu65Zr35 alloys. The comparison with the available experimental data suggested that this method can be of great help in designing the chemical composition of amorphous membranes for hydrogen purification [79]. Park et al. [80,81] investigated the change in the local atomic structures induced by hydrogen charging and its correlation with the embrittlement occurring during hydrogen permeation of amorphous membranes by means of molecular dynamics calculations. These authors investigated the short-range order of the Ni90Al10 alloy by calculating the fraction of Voronoi polyhedra using Voronoi tessellation method, and they found that the icosahedron indexed by (0,0,12,0) is the most abundant. Upon hydrogen charging, the volume of the membrane increases. The initially stable and densely packed icosahedral structures reported to have low ability to accommodate hydrogen atoms collapses, while prism or prism-like structures of low coordination number are created. These structural changes in the hydrogenated samples are proposed to be responsible for hydrogen embrittlement and hydrogen-induced plasticity of amorphous membranes [81]. Both phenomena occur according to Park et al. [81] because of the sudden increase in the degree of strain localization in the prism and prism-like polyhedra. Another aspect of metallic glasses investigated by numerical methods is the response to shear deformations [81]. Delogu [82] evidenced that induced atomic displacements perturbed small groups of atoms and became irreversibly rearranged. Amorphous metals are prone to mechanical instability due to the observation of deformed regions. The shear deformation induces a gradual decrease of their mechanical stability and finally their rearrangement. Moreover, deformations also promote the formation of new regions with low mechanical stability [82].



Chapter 9 Amorphous metal membranes

Conclusions ad future trends The research of the last 20 years conducted toward the development of amorphous NieNbeZr ribbons shows permeability values close to or higher than that of the benchmark Pd and Pd alloys membranes. The amorphous structure of such materials poses many challenges for their investigation and the interpretation of their properties, such as (i) suppression of crystallization occurring at high temperature, (ii) the decrease in hydrogenation enthalpies as the hydrogen content increases, (iii) the hydrogen embrittlement seems to be less dramatic that in crystalline materials but the physical mechanisms behind should be more clearly defined, and (iv) computational studies should greatly help in the comprehension of the short-range local structure, and some models connecting the structural and the functional properties started to be developed. In some cases, contradictory results were reported for membranes with similar compositions, and this fact could be also due to differences in synthesizing and slight changes in the microstructure as a function of the synthesis conditions. Many properties of these materials still need to be investigated, such as the long-term stability in a hydrogen environment, the resistance to poisoning due to gases different from hydrogen, and the time evolution of the microstructure. This information could be the subject of future research in this field.


Body-centered cubic CALculation of PHAse Diagram Density functional theory Differential scanning calorimetry Face-centered cubic Extended X-ray absorption fine structure Johnson-Mehl-Avrami equation Standard temperature and pressure Transmission electron microscopy X-ray absorption near edge structure X-ray diffraction

List of symbols c Eact Ek H/M Hind

atomic hydrogen concentration activation energy of crystallization activation energy of solution hydrogen over metal atomic ratio indentation hardness

Chapter 9 Amorphous metal membranes

Hv i K K0 n P Q R T Tc Tg DHhyd sf

Vickers hardness current density hydrogen solubility solubility constant Avrami exponent hydrogen pressure scattering vector constant of ideal gases absolute temperature crystallization temperature glass transition temperature hydrogenation enthalpy tensile fracture strength

References [1] N.W. Ockwig, T.M. Nenoff, Membranes for hydrogen separation, Chem. Rev. 107 (2007) 4078e4110. [2] M.D. Dolan, N.C. Dave, A.Y. Ilyushechkin, L.D. Morpeth, K.G. McLennan, Composition and operation of hydrogen-selective amorphous alloy membranes, J. Membr. Sci. 285 (2006) 30e55. [3] S. Hara, K. Sakaki, N. Itoh, H.-M. Kimura, K. Asami, A. Inoue, An amorphous alloy membrane without noble metals for gaseous hydrogen separation, J. Membr. Sci. 164 (2000) 289e294. [4] S. Hara, N. Hatakeyma, N. Itoh, H.-M. Kimura, A. Inoue, Hydrogen permeation through palladium-coated amorphous Zr-M-Ni (M ¼ Ti, Hf) alloy membranes, Desalination 144 (2002) 115e120. [5] S. Hara, N. Hatakeyma, N. Itoh, H.-M. Kimura, A. Inoue, Hydrogen permeation through amorphous Zr36-xHfxNi64 alloy membranes, J. Membr. Sci. 211 (2003) 149e156. [6] S.-I. Yamaura, A. Inoue, Preparation and characterization of amorphous alloy membranes, in: A. Basile, F. Gallucci (Eds.), Membranes for Membrane Reactors: Preparation, Optimization and Selection, John Wiley & Sons, Ltd, 2011, ISBN 978-0-470-74652-3 (Chapter 20). [7] S. Sarker, D. Chandra, M. Hirscher, M. Dolan, D. Isheim, J. Wermer, D. Viano, M. Baricco, T.J. Udovic, D. Grant, O. Palumbo, A. Paolone, R. Cantelli, Developments in the Ni-Nb-Zr amorphous alloy membranes, Appl. Phys. A 122 (2016) 168. [8] Z.-Q. Hu, A.-M. Wang and H.-F. Zhang, Amorphous materials, in modern inorganic synthetic chemistry, (Chapter 22), [9] N. Eliaz, D. Eliezer, An overview of hydrogen interaction with amorphous alloys, Adv. Perform. Mater. 6 (1999) 5e31. [10] K. Ishikawa, Y. Seti, K. Kita, M. Matsuda, M. Nishida, K. Aoki, Hydrogen permeation in rapidly quenched amorphous and crystallized Nb20Ti40Ni40 alloy ribbons, Int. J. Hydrogen Energy 36 (2011) 1784e1792. [11] K. Ishikawa, Y. Seti, K. Kita, M. Nishida, K. Aoki, Hydrogen permeability and microstructure of rapidly quenched Nb-TiNi alloys, J. Alloy. Comp. 509S (2011) S790eS793. [12] T. Geiller, M.-Y. Kim, Y.-C. Kim, J.-Y. Suh, E. Fleury, IEEE Nanotechnology Materials and Device Conference, Hydrogen Permeation Properties of InSitu Ti-Based Bulk Metallic Glass Matrix Composite Membranes, 2011.



Chapter 9 Amorphous metal membranes

[13] S.N. Paglieri, N.K. Pal, M.D. Dolan, S.-M. Kim, W.-M. Chien, J. Lamb, D. Chandra, K.M. Hubbard, D.P. Moore, Hydrogen permeability, thermal stability and hydrogen embrittlement of Ni-Nb-Zr and Ni-Nb-Ta-Zr amorphous alloy membrenes, J. Membr. Sci. 378 (2011) 42e50. [14] O. Palumbo, F. Trequattrini, N. Pal, M. Hulyalkar, S. Sarker, D. Chandra, T. Flanagan, M. Dolan, A. Paolone, Hydrogen absorption properties of amorphous (Ni0.6Nb0.4-yTay)100-xZrx membranes, Prog. Nat. Sci.-Mater. 27 (2017) 126e131. [15] O. Palumbo, F. Trequattrini, S. Sarker, M. Hulyalkar, N. Pal, D. Chandra, M. Dolan, A. Paolone, New studies of the physical properties of metallic amorphous membranes for hydrogen purification, Challenges 8 (2017) 4. [16] S.-M. Kim, D. Chandra, N.K. Pal, M.D. Dolan, W.-M. Chien, A. Talekar, J. Lamb, S. Paglieri, T.B. Flanagan, Hydrogen permeability and crystallization kinetics in amorphous Ni-Nb-Zr alloys, INt. J. Hydrogen Energ. 37 (2012) 3904e3913. [17] O. Palumbo, F. Trequattrini, M. Hulyalkar, S. Sarker, N. Pal, D. Chandra, T. Flanagan, M. Dolan, A. Paolone, Hydrogen induced abrupt structural expansion at high temperatures of a Ni32Nb28Zr30Cu10 membrane for H2 purification, Membranes 6 (2016) 48. [18] H.Y. Ding, W. Zhang, S.I. Yamaura, K.F. Yao, Hydrogen permeable Nbbased amorphous alloys with high thermal stability, Mater. Trans. 54 (2013) 1330e1334. [19] M.-H. Kim, M.R. Debnath, Y.-I. Wang, J.-Y. Suh, E. Fleury, D.-H. Kim, Effect of Co on the degradation of hydrogen permeability of Ni-Nb-Zr amorphous membranes, Met. Mater. Int. 20 (2014) 215e219. [20] T. Lai, S.S. Singh, A.S.S. Singaravelu, K.S. Vadari, A. Khosravi, N. Chawla, M.L. Lind, Hydrogen permeability and mechanical properties of NiNb-M (M ¼ Sn, Ti and Zr) amorphous metallic membranes, J. Alloy. Comp. 684 (2016) 359e365. [21] S. Nayebossadri, D. Smith, J. Speight, D. Book, Amorphous Zr-Based thin films fabricated by magnetron sputtering for potential application in hydrogen purification, J. Alloy. Comp. 645 (2015) 556e560. [22] S. Nayebossadri, C.J. Greenwood, J. Speight, D. Book, Thermal and structural stability of Zr-based amorphous thin films for potential application in hydrogen purification, Separ. Purif. Technol. 187 (2017) 173e183. [23] W. Dandana, M.A. Yousfi, K. Hajlaoui, F. Gamaoun, A.R. Yavari, Thermal stability and hydrogen-induced softening in Zr57Al10Cu15.4Ni12.6Nb5 metallic glass, J. Non-Cristalline Solids 456 (2017) 138e142. [24] T. Lai, H. Yin, M.L. Lind, The hydrogen permeability of Cu-Zr binary amorphous metallic membranes and the importance of thermal stability, J. Membr. Sci. 489 (2015) 264e269. [25] Hydrogen in metals II, in: G. Alefeld, J. Viilkl (Eds.), Top. Appl. Phys. 29 (1978). [26] V.T. Huett, D. Zander, L. Jastrow, E.H. Majzoub, K.F. Kelton, U. Köster, Gaseous hydrogen charging of ZreCueNieAl glasses and quasicrystals, J. Alloy. Comp. 379 (2004) 16e22. [27] S.-I. Yamaura, M. Sakurai, M. Hasegawa, K. Wakoh, Y. Shimpo, M. Nishida, H. Kimura, E. Matsubara, A. Inoue, Hydrogen permeation and structural features of melt-spun NieNbeZr amorphous alloys, Acta Mater. 53 (2005) 3703e3711. [28] S. Hao, D.S. Sholl, Comparison of first principles calculations and experiments for hydrogen permeation through amorphous ZrNi and ZrNiNb films, J. Membr. Sci. 350 (2010) 402e409.

Chapter 9 Amorphous metal membranes

[29] R. Kirchheim, Acta Metall. 30 (1982) 1069e1078. [30] R. Kirchheim, F. Sommer, G. Schluckebier, Acta Metall. 30 (1982) 1059e1068. [31] R. Kirchheim, Acta Metall. 21 (1973) 1233e1242. [32] S. Jayalakshmi, Y.G. Choi, Y.C. Kim, Y.B. Kim, E. Fleury, Intermetallics 18 (2010) 1988e1993. [33] S. Hara, H.-X. Huang, M. Ishitsuka, M. Mukaida, K. Haraya, N. Itoh, K. Kita, K. Kato, Hydrogen solution properties in a series of amorphous ZreHfeNi alloys at elevated temperatures, J. Alloy. Comp. 458 (2008) 307e312. [34] D. Chandra, M. Dolan, A. Paolone, Hydrogen isotope effect in a Ni32Nb28Zr30Cu10 amorphous membrane, J. Alloys Compds 803 (2019) 172e175. [35] E. Serra, M. Kemali, A. Perujo, D.K. Ross, Hydrogen and deuterium in Pd-25 Pct Ag alloy: permeation, diffusion, solubilization, and surface reaction, Metall. Mater. Trans. A 29 (1998) 1023e1028. [36] A. Paolone, S. Tosti, A. Santucci, O. Palumbo, F. Trequattrini, Hydrogen and deuterium solubility in commercial Pd-Ag alloys for hydrogen purification, ChemEngineering 1 (2017) 14. [37] S. Pati, R.A. Jat, S.K. Mukerjee, S.C. Parida, Hydrogen isotope effect on thermodynamic and kinetics of hydrogen/deuterium absorptiondesorption in Pd0.77Ag0.10Cu0.13 alloy, J. Phys. Chem. C 119 (2015) 10314e10320. [38] D. Zander, H. Leptien, U. Köster, N. Eliaz, D. Eliezer, Hydrogenation of Zrbased metallic glasses and quasicrystals, J. Non-cryst. Solids 250e252 (1999) 893e897. [39] D. Peng, J. Shen, J. Sun, Y. Chen, Kinetic characteristic of hydrogenation ZrTi-Cu-Ni-Be bulk amorphous alloys, J. Mater. Sci. Technol. 20 (2004) 157e159. [40] S. Jayalakshmi, V.S. Vasantha, E. Fleury, M. Gupta, Characteristics of Ni-Nbbased metallic amorphous alloys for hydrogen-related energy applications, Appl. Energy 90 (2012) 94e99. [41] S.-I. Yamaura, A. Inoue, Effect of surface coating element on hydrogen permeability of melt-spun Ni40Nb20Ta5Zr30Co5 amorphous alloy, J. Membr. Sci. 349 (2010) 138e144. [42] D. Singh, R.R. Shahi, T.P. Yadav, R.K. Mandal, R.S. Tiwari, O.N. Srivastava, Hydrogenation of (Zr69.5Al7.5Cu12Ni11)100xTix quasicrystalline alloys and its effect on their structural and microhardness behavior, J. Non-cryst. Solids 380 (2013) 11e16. [43] S.-I. Yamaura, Y. Shimpo, H. Okouchi, M. Nishida, O. Kajita, H. Kimura, A. Inoue, Hydrogen permeation characteristics of melt-spun NieNbeZr amorphous alloy membranes, Mater. Trans., JIM 44 (2003) 1885e1890. [44] S.-I. Yamaura, Y. Shimpo, H. Okouchi, M. Nishida, O. Kajita, A. Inoue, The affect of additional elements on hydrogen permeation properties of meltspun Ni-Nb-Zr amorphous alloys, Mater. Trans., JIM 45 (2004) 330e333. [45] S.-I. Yamaura, S. Nakata, H. Kimura, Y. Shimpo, M. Nishida, A. Inoue, Hydrogen permeation of the melt-spun NieXeZr amorphous membranes, Mater. Trans. 46 (2005) 1768e1770. [46] Y. Shimpo, S.-I. Yamaura, M. Nishida, H. Kimura, A. Inoue, Development of melt-spun NieNbeZreCo amorphous alloy for high-performance hydrogen separating membrane, J. Membr. Sci. 286 (2006) 170e173. [47] S.-I. Yamaura, S. Nakata, H. Kimura, A. Inoue, Hydrogen permeation of NiNb-Zr metallic glasses in a supercooled liquid state, Mater. Trans., JIM 47 (2006) 2991e2996.



Chapter 9 Amorphous metal membranes

[48] K.B. Kim, K.D. Kim, D.Y. Lee, Y.C. Kim, E. Fleury, D.H. Kim, Hydrogen permeation properties of Pd-coated Ni60Nb30Ta10 amorphous alloy membrane, Mater. Sci. Eng. A 449e451 (2007) 934e936. [49] J.B. Qiang, W. Zhang, S. Yamaura, A. Inoue, Thermal stability and hydrogen permeation of Ni42Zr30Nb28-xTax amorphous alloys, Mater. Trans., JIM 50 (2009) 1236e1239. [50] M.D. Dolan, N.C. Dave, L.D. Morpeth, R. Donelson, D. Liang, M.E. Kellam, S. Song, Ni-based amorphous alloy membranes for hydrogen separation at 400  C, J. Membr. Sci. 326 (2009) 549e555. [51] M.D. Dolan, S. Hara, N.C. Dave, K. Haraya, M. Ishitsuka, K. Kita, K.G. McLennan, L.D. Morpeth, M. Mukaida, Thermal stability, glass-forming ability and hydrogen permeability of amorphous Ni64Zr36XMX (M ¼ Ti, Nb, Mo, Hf, Ta or W) membranes, Separ. Purif. Technol. 65 (2009) 298e304. [52] Y.-I. Wang, J.-Y. Suh, Y.-S. Lee, J.-H. Shim, E. Fleury, Y.W. Cho, S.-U. Koh, Direct measurement of hydrogen diffusivity through Pd-coated Ni-based amorphous metallic membranes, J. Membr. Sci. 436 (2013) 195e201. [53] A. Adibhatla, M.D. Dolan, W. Chien, D. Chandra, Enhancing the catalytic activity of Ni-based amorphous alloy membrane surfaces, J. Membr. Sci. 463 (2014) 190e195. [54] H.-S. Chin, J.-Y. Suh, K.-W. Park, W. Lee, E. Fleury, Hydrogen permeability of glass-forming Ni-Nb-Zr-Ta crystalline membranes, Met. Mater. Int. 17 (2011) 541e545. [55] K.-W. Park, J.-P. Ahn, H.-K. Seok, Y.-C. Kim, Relationship between activation energy for hydrogen permeation and hydrogen permeation properties of amorphous Cu50Zr50 and Cu65Zr35 membranes, Intermetallics 19 (2011) 1887e1890. [56] W. Prochwicz, W. Macherzynski, B. Paszkiewicz, Z.M. Stepien, Hydrogen permeation properties of Pd-coated Pd33Ni52Si15 amorphous alloy membrane, Mater. Sci.-Poland 33 (2015) 56e58. [57] W. Prochwicz, W. Macherzynski, B. Paszkiewicz, Z.M. Stepien, Effect of surface coating with palladium on hydrogen permeability of Pd33Ni52Si15 amorphous alloy membrane, Opt. Appl. 46 (2016) 173e179. [58] H. Oji, K. Handa, J. Ide, T. Honma, N. Umesaki, S. Yamaura, M. Fukuhara, A. Inoue, S. Emura, Local atomic structure around Ni, Nb, and Zr atoms in Ni-Nb-Zr-H glassy alloys, J. Phys. Conf. Ser. 190 (2009) 012075. [59] H. Oji, K. Handa, J. Ide, T. Honma, S. Yamaura, A. Inoue, N. Umesaki, Local atomic structure around Ni, Nb, and Zr atoms in NieNbeZreH glassy alloys studied by x-ray absorption fine structure method, J. Appl. Phys. 105 (2010) 113527. [60] M. Matsuura, M. Fukuhara, K. Konno, T. Fujita, M.W. Chen, N. Fujima, A. Inoue, J. Non-cryst. Solids 357 (2011) 3357e3360. [61] M. Fukuhara, N. Fujima, H. Oji, A. Inoue, S. Emura, Structures of the icosahedral clusters in NieNbeZreH glassy alloys determined by firstprinciples molecular dynamics calculation and XAFS measurements, J. Alloy. Comp. 497 (2010) 182e187. [62] S. Sarker, D. Isheim, G. King, Q. An, D. Chandra, S.I. Morozov, K. Page, J.N. Wermer, D.N. Seidman, M. Dolan, Icosahedra clustering and short range order in Ni-Nb-Zr amorphous membranes, Sci. Rep. 8 (2018) 6084. [63] S.-I. Yamaura, M. Hasegawa, H. Kimura, A. Inoue, Effects of hydrogen on the mechanical properties of Ti50Ni25Cu25 metallic glass, Mater. Trans. 43 (2002) 2543e2547. [64] H. Kimura, A. Inoue, S.-I. Yamaura, K. Sasamori, M. Nishida, Y. Shinpo, H. Okouchi, Thermal stability and mechanical properties of glassy and

Chapter 9 Amorphous metal membranes









[73] [74]



[77] [78]


[80] [81]


amorphous NieNbeZr alloys produced by rapid solidification, Mater. Trans. 44 (2003) 1167e1171. X. Du, X. Ye, Q. Ren, The preparation and hydrogen brittleness resistance of Pd71.5Cu12Si16.5 metallic glass ribbons, IOP Conf. Ser. Earth Environ. Sci. 100 (2017) 012132. A. Kawashima, S. Yamaura, N. Ohtsu, H. Kimura, A. Inoue, Mechanical properties of melt-spun amorphous Ni-Nb-Zr alloys after hydrogen charging, Mater. Trans. 47 (2006) 1523e1526. K. Horikawa, K. Yanaue, H. Kobayashi, Response of hydrogen-induced bending deformation in ZrNi amorphous membranes, Mater. Trans. 51 (2010) 2181e2187. S. Jayalakshmi, S.O. Park, K.B. Kim, E. Fleury, D.H. Kim, Studies on hydrogen embrittlement in Zr- and Ni-based amorphous alloys, Mater. Sci. Eng. A 449e451 (2007) 920e923. O. Palumbo, S. Brutti, F. Trequattrini, S. Sarker, M. Dolan, D. Chandra, A. Paolone, Temperature dependence of the elastic modulus of (Ni0.6Nb0.4)1-xZrx membranes: effects of thermal treatments and hydrogenation, Energies 8 (2015) 3944e3954. M. Nishida, M. Matsuda, Y. Shimada, K. Takashima, K. Ishikawa, K. Aoki, Microstructural and mechanical characterizations of rapidly solidified NbTiNi hydrogen permeation alloy, J. Phys. Conf. Ser. 144 (2009), 012106. D. Singh, R.S. Tiwari, O.N. Srivastava, Hydrogenation of Zr-based quasicrystals, in: M.T. Ravanchi (Ed.), New Advances in Hydrogenation Processes, IntechOpen, Rijeka, 2017. Chap. 7 Y. Zhao, I.-C. Choi, M.-Y. Seok, U. Ramamurty, J.-Y. Suh, J.-I. Jang, Hydrogen-induced hardening and softening of NieNbeZr amorphous alloys: dependence on the Zr content, Scr. Mater. 93 (2014) 56e59. W. Dandana, K. Hajlaoui, Effect of hydrogen addition on the mechanical properties of Zr-based metallic glass, J. Alloy. Comp. 742 (2018) 563e566. S. Hao, M. Widom, D.S. Sholl, Probing hydrogen interactions with amorphous metals using first-principles calculations, J. Phys. Condens. Matter 21 (2009) 115402. N. Fujima, T. Hoshino, M. Fukuhara, Local structures and structural phase change in Ni-Zr-Nb glassy alloys composed of Ni5Zr5Nb3 icosahedral clusters, J. Appl. Phys. 114 (2013) 063501. S. Kang, S. Hao, D.S. Sholl, Using first-principles models to advance development of metal membranes for high temperature hydrogen purification, Membr. Sci. Technol. 14 (2011) 309e331. S. Hao, D.S. Sholl, Computational prediction of durable amorphous metal membranes for H2 purification, J. Membr. Sci. 381 (2011) 192e196. S. Hao, D.S. Sholl, Rapid prediction of hydrogen permeation through amorphous metal membranes: an efficient computational screening approach, Energy Environ. Sci. (2013) 232e240. B.-M. Lee, J.-H. Shim, J.-Y. Suh, B.-J. Lee, A semi-empirical methodology to predict hydrogen permeability in amorphous alloy membranes, J. Membr. Sci. 472 (2014) 102e109. K.-W. Park, Y. Shibutani, E. Fleury, Hydrogen-induced structural change in Ni90Al10 metallic glass, J. Alloy. Comp. 509S (2011) S456eS459. K.-W. Park, Y. Shibutani, Effect of hydrogenation on the mechanical property of amorphous Ni90Al10 membranes, Int. J. Hydrogen Energy 36 (2011) 9324e9334. F. Delogu, Irreversible rearrangements and fertile regions in deformed metallic glasses, Intermetallics 22 (2012) 203e209.