Pressure-induced reversible amorphization and an amorphous-amorphous transition in Ge₂Sb₂Te₅ phase-change memory material

Ge₂Sb₂Te₅ (GST) is a technologically very important phase-change material that is used in digital versatile disks-random access memory and is currently studied for the use in phase-change random access memory devices. This type of data storage is achieved by the fast reversible phase transition between amorphous and crystalline GST upon heat pulse. Here we report pressure-induced reversible crystalline-amorphous and polymorphic amorphous transitions in NaCl structured GST by ab initio molecular dynamics calculations. We have showed that the onset amorphization of GST starts at approximately 18 GPa and the system become completely random at approximately 22 GPa. This amorphous state has a cubic framework (c-amorphous) of sixfold coordinations. With further increasing pressure, the c-amorphous transforms to a high-density amorphous structure with trigonal framework (t-amorphous) and an average coordination number of eight. The pressure-induced amorphization is investigated to be due to large displacements of Te atoms for which weak Te–Te bonds exist or vacancies are nearby. Upon decompressing to ambient conditions, the original cubic crystalline structure is restored for c-amorphous, whereas t-amorphous transforms to another amorphous phase that is similar to the melt-quenched amorphous GST.     

Suppression of phase separation and giant enhancement of superconducting transition temperature in FeSe1?xTex thin films

We demonstrate the successful fabrication on CaF₂ substrates of FeSe_1−xTe_x films with 0 ≤ x ≤ 1, including the region of 0.1 ≤ x ≤ 0.4, which is well known to be the “phase-separation region,” via pulsed laser deposition that is a thermodynamically nonequilibrium method. In the resulting films, we observe a giant enhancement of the superconducting transition temperature, T_c, in the region of 0.1 ≤ x ≤ 0.4: The maximum value reaches 23 K, which is ∼1.5 times as large as the values reported for bulk samples of FeSe_1−xTe_x. We present a complete phase diagram of FeSe_1−xTe_x films. Surprisingly, a sudden suppression of T_c is observed at 0.1 < x < 0.2, whereas T_c increases with decreasing x for 0.2 ≤ x < 1. Namely, there is a clear difference between superconductivity realized in x = 0 ? 0.1 and in x ≥ 0.2. To obtain a film of FeSe_1−xTe_x with high T_c, the controls of the Te content x and the in-plane lattice strain are found to be key factors.Since the discovery of superconductivity in LaFeAs(O,F) (1), many studies concerning iron-based superconductors have been conducted. FeSe is the iron-based superconductor with the simplest crystal structure (2). The T_c of FeSe is ∼8 K, which is not very high in comparison with other iron-based superconductors. However, the value of T_c strongly depends on the applied pressure, and the temperature at which the resistivity becomes zero, Tczero, reaches as high as  ～ 30 K at 6 GPa (3). This suggests that FeSe samples with higher T_c are available by the fabrication of thin films because we can introduce lattice strain. Indeed, we have previously reported that FeSe films fabricated on CaF₂ substrates exhibit T_c values ∼1.5 times higher than those of bulk samples because of in-plane compressive strain (4). On the other hand, superconductivity with T_c of 65 K has recently been reported in a monolayer FeSe film on SrTiO₃ (5, 6). It is unclear whether this superconductivity results from the characteristics of the interface. However, this finding indicates that FeSe demonstrates potential as a very-high-T_c superconductor.The partial substitution of Te for Se in FeSe also raises T_c to a maximum of 14 K at x = 0.5 ? 0.6 (7). In FeSe_1−xTe_x, it is well known that we cannot obtain single-phase samples with 0.1 < x < 0.4 because of phase separation (7). Here, we focus on this region of phase separation. Generally, the process of film deposition involves crystal growth in a thermodynamically nonequilibrium state. Thus, film deposition provides an avenue for the synthesis of a material with a metastable phase. In this paper, we report the fabrication of epitaxial thin films of FeSe_1−xTe_x with 0 ≤ x ≤ 1 on CaF₂ substrates, using the pulsed laser deposition (PLD) method. We demonstrate that single-phase epitaxial films of FeSe_1−xTe_x with 0.1 ≤ x ≤ 0.4 are successfully obtained and that the maximum value of T_c is as large as 23 K, which is higher than the previously reported values for bulk and film samples of FeSe_1−xTe_x (7–12), except for those of the monolayer FeSe films (5, 6). Our results clearly show that the optimal Te content for the highest T_c for FeSe_1−xTe_x films on CaF₂ is different from the widely believed value for this system.Fig. 1A presents the X-ray diffraction patterns of FeSe_1−xTe_x films for x = 0 ? 0.5 on CaF₂. Here and hereafter, the Te content x of our films represents the nominal Te composition of the polycrystalline target. With the exception of an unidentified peak in the FeSe_0.5Te_0.5 film, only the 00l reflections of a tetragonal PbO-type structure are observed, which indicates that these films are well oriented along the c axis. Fig. 1 B and C presents enlarged segments of these plots near the 001 and 003 reflections, respectively. The 2θ values of the peak positions decrease with increasing x in a continuous manner, which is consistent with the fact that the c-axis length increases with increasing x. It should be noted that the values of the full widths at half maximum (FWHM), δ(2θ), of the FeSe_1−xTe_x films with x = 0.1 ? 0.4, which is known as the region of phase separation in the bulk samples (7), are δ(2θ) = 0.2°?0.3° for the 001 reflection and δ(2θ) = 0.4°?0.6° for the 003 reflection, which are nearly the same as the values for the FeSe and FeSe_0.5Te_0.5 films. This result is in sharp contrast to the previously reported result that the FWHM was broad only in films of FeSe_1−xTe_x with x = 0.1 and 0.3 (13), where phase separation has been believed to occur. The results presented in Fig. 1 A–C indicate the formation of a single phase in our FeSe_1−xTe_x films with x = 0.1 ? 0.4.Open in a separate windowFig. 1.(A) Out-of-plane X-ray diffraction patterns of FeSe_1−xTe_x thin films for x = 0 ? 0.5 with film thicknesses of 120–147 nm. The # symbols represent peaks associated with the substrate. The * represents an unidentified peak. Enlarged segments of the plots presented in A near the 001 and 003 peaks are shown in B and C, respectively. (D) The c-axis lengths of FeSe_1−xTe_x films, where the x value indicated on the horizontal axis is the nominal Te content. (E) Relations between the a-axis and c-axis lengths in FeSe_1−xTe_x films. The colors and shapes of the symbols correspond to the Te content x, as shown. The dashed lines are guides for the eye. The data for x = 0 and 0.5 presented in A–E are cited from refs. 4, 9, and 14.In Fig. 1D, the c-axis lengths of 29 films of FeSe_1−xTe_x are plotted as a function of x. The values of the c-axis lengths vary almost linearly with the nominal Te contents of the targets in the whole range of x, including both end-member materials. The evident formation of a single phase and the systematic change in the c-axis length strongly indicate that the nominal Te content of the polycrystalline target is nearly identical to that of the final FeSe_1−xTe_x film. Note that the compositional analysis of grown films using scanning electron microscopy/energy-dispersive X-ray (SEM/EDX) analysis is impossible for FeSe_1−xTe_x films on CaF₂ substrates because the energies of the K edge of Ca and the L edge of Te are very close to each other. The above-mentioned features indicate that phase separation is suppressed in our FeSe_1−xTe_x films with x = 0.1 ? 0.4 on CaF₂ substrates. To our knowledge, this result is the first manifestation of the suppression of phase separation in FeSe_1−xTe_x with x = 0.1 ? 0.4.Fig. 1E presents the relations between the a-axis and c-axis lengths in films of FeSe_1−xTe_x. At first glance, there seem to be no relations between the a-axis length and x, in sharp contrast to the behavior of the c-axis length. The a-axis and c-axis lengths of films with the same x show a weak negative correlation. This behavior cannot be explained by a difference in Te content of a film, which should result in a positive correlation. By contrast, if variations in c are caused by a difference in in-plane lattice strain, this behavior can be explained in terms of the Poisson effect. Indeed, the a-axis lengths of films of FeSe and FeSe_0.5Te_0.5 are smaller than those of bulk samples with the same composition. Thus, we consider that the a-axis length predominantly depends on the in-plane lattice strain rather than the Te content x. One might think that this behavior looks strange, because the lattice constant of CaF₂(aCaF2/2) is longer than the a of FeSe_1−xTe_x, which usually leads to a tensile strain. In a previous paper, the penetration of F⁻ ions from the CaF₂ substrates into the films was proposed as a possible mechanism for nontrivial compressive strain in FeSe_1−xTe_x films on CaF₂ substrates (15). Because of the smaller ionic radius of F⁻ than that of Se²⁻, this peculiar compressive strain can be explained by the partial substitution of F⁻ for Se²⁻ near the interface between a film and a substrate.Fig. 2 A–D presents the temperature dependences of the electrical resistivities, ρ, of 16 films of FeSe_1−xTe_x for x = 0.1 ? 0.4. The value of T_c depends on the film thickness, even in films with the same x. The highest Tconset, which is defined as the temperature where the electrical resistivity deviates from the normal-state behavior, and the Tczero of the FeSe_1−xTe_x films are 13.2 K and 11.5 K, respectively, for x = 0.1; 22.8 K and 20.5 K, respectively, for x = 0.2; 20.9 K and 19.9 K, respectively, for x = 0.3; and 20.9 K and 20.0 K, respectively, for x = 0.4. Compared with the results for bulk samples, a drastic enhancement of T_c is observed in these FeSe_1−xTe_x films. Surprisingly, the values of Tczero in the films with x = 0.2 and 0.4 exceed 20 K. These values are larger than those reported for FeSe_0.5Te_0.5 films (8, 10, 11, 14). In particular, the T_c of the FeSe_0.8Te_0.2 film with a thickness of 73 nm is ∼1.5 times as high as those of bulk crystals of FeSe_1−xTe_x with the optimal composition, x ≈ 0.5 (7). Based on the measurement of the ρ of the FeSe_0.8Te_0.2 film under a magnetic field applied along the c axis, we estimate an upper critical field at 0 K of μ₀H_c2 = 55.4 T, using the Werthamer–Helfand–Hohenberg (WHH) theory (16), which yields a Ginzburg–Landau coherence length at 0 K of ξ_ab(0) ～ 24.4 Å (SI Appendix). This value of μ₀H_c2 is approximately half the value for an FeSe_0.5Te_0.5 film on CaF₂ with a T_c of ∼16 K (9).Open in a separate windowFig. 2.Temperature dependences of the electrical resistivities, ρ, of FeSe_1−xTe_x thin films for (A) x = 0.1, (B) x = 0.2, (C) x = 0.3, and (D) x = 0.4 with different film thicknesses. Insets present enlarged views of the plots near the superconducting transition.Using the data shown above, we present the phase diagram of FeSe_1−xTe_x films on CaF₂ substrates in Fig. 3. For comparison, the data for bulk samples of FeSe_1−xTe_x (7, 17) are also plotted in Fig. 3. In bulk crystals, the optimal Te content to achieve the highest T_c is considered to be x ≈ 0.5, and phase separation occurs in the region of 0.1 ≤ x ≤ 0.4 (7). However, our data clearly demonstrate that this phase separation is absent and that the optimal composition for an FeSe_1−xTe_x film on a CaF₂ substrate is not x ≈ 0.5 but x ≈ 0.2. It should be noted that the dependence of T_c on x suddenly changes at the boundary defined by 0.1 < x < 0.2. Unlike the “dome-shaped” phase diagram that is familiar in iron-based superconductors, the values of T_c in films with 0.2 ≤ x ≤ 1 increase with decreasing x, whereas the strong suppression of T_c is observed at 0.1 < x < 0.2. The behavior in films with x ≥ 0.2 can be explained by the empirical law that shows the relation between T_c and structural parameters. In iron-based superconductors, it is well accepted that the bond angle of (Pn, Ch)-Fe-(Pn, Ch) (Pn, ?Pnictogen; Ch, ?Chalcogen), α (18, 19), and/or the anion height from the iron plane, h (20), are the critical structural parameters that determine the value of T_c. In bulk samples of FeSe_1−xTe_x, α and h approach their optimal values, i.e., α = 109.47° (18, 19) and h = 1.38 Å (20), with decreasing x (down to x = 0), which should be the same in FeSe_1−xTe_x films. Therefore, the increase of T_c in films with 0.2 ≤ x ≤ 1 with decreasing x can be explained by the optimization of α and/or h based on the empirical law. However, the sudden suppression of T_c in films with 0 ≤ x < 0.2 is not consistent with this scenario, and its origin should be sought among other factors. We consider there are two candidates for this origin from the structural analysis of bulk samples of FeSe_1−xTe_x. One is the effect of the orthorhombic distortion. In a bulk sample of FeSe, a structural phase transition from tetragonal to orthorhombic occurs at 90 K (21). However, in bulk samples of FeSe_1−xTe_x with x ～ 0.4 ? 0.6 where T_cs take optimum values, there are papers with different conclusions on the presence/absence of a similar type of structural transition to that of FeSe (22–24). It should be noted that a structural transition temperature is lower and that the orthorhombicity is much smaller than those of FeSe even in the report where the structural transition is present (24). These results on crystal structures suggest that the orthorhombic distortion results in a suppression of T_c. This scenario is applicable to the behavior of our films, if a large orthorhombic distortion is observed only in films with x = 0 ? 0.1. The other candidate is the change in the distance between the layers of Fe-Ch tetrahedra, δ. As shown in SI Appendix, in polycrystalline samples of FeSe_1−xTe_x, the δ value of FeSe is much smaller than those of FeSe_1−xTe_x with x ≥ 0.5 where δ is nearly independent of x (22). We speculate that the decrease of δ in FeSe is related with the suppression of T_c. Indeed, in polycrystalline samples, FeSe exhibits smaller values of δ and T_c than does FeSe_0.5Te_0.5 (7, 22), and the intercalation of alkali metals and alkaline earths into FeSe results in the c-axis length as large as ∼20 Å and T_c as high as 45 K (25, 26). At this moment, the origin of the suppression of T_c at 0.1 < x < 0.2 is unclear. Regardless of its origin, we believe that it is reasonable to distinguish between superconductivity in x = 0 ? 0.1 and in x ≥ 0.2. In other words, our phase diagram in Fig. 3 provides a previously unidentified view for superconductivity in FeSe_1−xTe_x, that is, a discontinuity in superconductivity of FeSe_1−xTe_x. We are able to come to this picture only after the data for x = 0.1 ? 0.4 become available in this study. If we remove a cause for the suppression of T_c in x=0,0.1 in some way, a further increase in T_c can be expected because of the optimization of structural parameters.Open in a separate windowFig. 3.Dependence of T_c on x. The red and blue circles represent the Tconset and Tczero values of the FeSe_1−xTe_x thin films, respectively. The black triangles represent the Tconset values obtained in measurements of the magnetic susceptibility of bulk samples (7, 17). The dashed curve is a guide for the eye.In conclusion, we prepared high-quality epitaxial thin films of FeSe_1−xTe_x on CaF₂ substrates, using the pulsed laser deposition method. We successfully obtained FeSe_1−xTe_x films with 0.1 ≤ x ≤ 0.4, which has long been considered to be the “phase-separation region,” using a thermodynamically nonequilibrium growth of film deposition. From the results of electrical resistivity measurements, a complete phase diagram is presented in this system, in which the maximum value of T_c is as high as 23 K at x = 0.2. Surprisingly, a sudden suppression of T_c is observed at 0.1 < x < 0.2, whereas T_c increases with decreasing x for 0.2 ≤ x < 1. This behavior is different from that of the dome-shaped phase diagram that is familiar in iron-based superconductors.