Aarhus University Seal / Aarhus Universitets segl



Phase-change materials (PCMs) are at the cutting edge of material research for non-volatile memories and neuromorphic computing. They are one of the new types of non-volatile memory being studied to meet the world’s ever-increasing demand for digital information, the volume of which is doubling every two years. Such memories could potentially work a thousand times faster than current flash computer memory.

PCMs such as Ge-Sb-Te alloys can be rapidly and reversibly switched between their amorphous and crystalline states within a few nanoseconds by electrical pulses. The strong electrical contrast between the two states can be encoded as “0” and “1” for fast nonvolatile data storage applications.

Fast switching between amorphous and crystalline states

By applying an electrical or laser pulse, the amorphous state ("off"-state) is rapidly heated up to a temperature below the melting point, resulting in fast crystallization. The crystal, with an ordered structure and much higher electrical conductivity, can generate the "on" state. When a heat pulse is sufficiently large, the crystalline state can be heated up above the melting temperature, followed by a rapid quenching. The material is vitrified by avoiding crystallization and returns to the amorphous solid (glass) state. Such a complex sequence can be repeated millions of times with high reproducibility, which allows for fabricating nonvolatile memory chips for fast data storage.

Neutrons shed light on materials' dynamics

The phase switching consists of crystallization and vitrification processes, both of which involve the metastable (supercooled) liquid state of the material. Thus, the liquid-state behavior is non-trivial and is closely related to the crystallization kinetics and phase stabilities.

The high resolution of the neutron time-of-flight spectrometer is a powerful tool to understand the details of the atomic movements. With quasi-elastic neutron scattering at the Heinz Maier-Leibnitz Zentrum in Garching, we are able to make these movements "visible". Using neutron facilities at MLZ, Munich, we can determine the diffusivity and relaxation times of atoms in the liquid state of phase-change materials. Our work showed a striking breakdown of the Stokes-Einstein relation at high temperature even more than a hundred degrees above the melting point Tm. This questioned the widely-used assumption that the Stokes-Einstein relation is valid near Tm in PCMs, and raised the possibility that the breakdown is related to the thermodynamic transition in the supercooled liquid below Tm. The thermodynamic transition is also a fragile-strong transition in viscosity and metal-semiconductor transition, which plays a crucial role in the functionalities of phase-change materials.

No-man's land of Phase-Change Materials

The challenge with identifying the key factor for determining the fast phase switching and the stability is that the supercooled liquid below melting temperature Tm and above glass transition Tg is extremely difficult to access by experiments. This is because the crystallization rates of liquid phase-change materials, in contrast to the chalcogenide glassformers of earlier studies, are extremely high. The temperature regime of supercooled PCMs, obscured by fast crystallization, establishes the existence of a “no-man’s land”, in which no observations can be made except by ultrafast probes or deduction from external studies such as crystallization rates. Recent studies of crystallization kinetics, liquid dynamics, and densities of PCM-related chemical compounds all imply anomalous behaviors below Tm in the no-man’s land of phase-change materials.

Discovering the fragile-strong transition

The drastic change in the temperature dependence of viscosity from a high-temperature fragile liquid to a low-temperature strong liquid near Tg is the so-called fragile-strong transition (FST). According to the liquid fragility concept, some liquids, exhibiting a near-Arrhenius rise in viscosity on approaching Tg, are classified as “strong” liquids (e.g. SiO2), while others, showing a range of non-Arrhenius behavior, are referred to “fragile” liquids. Fragility is commonly characterized by measuring the slope of the Tg-scaled Arrhenius plot (Angell plot) at Tg, called “steepness index” or “m-fragility”.

A clear-cut fragile-strong transition is demonstrated in Ge15Te85 as a “double-kink” in the viscosity(T)-curve just above its eutectic melting temperature, which is verified by a direct differential scanning calorimetry (DSC) measurement near Tg. The corresponding structural change is shown as a stepwise rise in the second to first peak position ratio r2/r1 of reduced pair distribution functions G(r) from in situ x-ray scattering. There is increasing evidence for the existence of a fragile-strong transition in those fast switching phase-change materials such Ge-Sb-Te, which is likely located below the melting temperature obsecued by fast crystallization.

Probing liquid-liquid transitions in Phase-Change Materials

It is very recent that a pump-probe femtosecond x-ray diffraction experiment, which resolves structural changes within nanosecond timescales, has provided the first direct evidence of a structural transition in PCM Ag-In-Sb-Te (AIST) and Ge15Sb85 below Tm in the supercooled liquid. The liquid-liquid transition (LLT) is associated with a metal-semiconductor (M-SC) transition and a fragile-strong transition (FST) in viscosity. The latter is relevant to switching kinetics, as the FST controls the kinetic factor of nucleation and growth of crystals. In the fragile liquid state, a high kinetic factor facilitates crystallization (fast switching) at an elevated temperature during a “set” pulse, while, in the strong liquid state, a low kinetic factor hinders crystallization at ambient temperature, making the state favorable for data retention. Thus, the existence of this transition appears to be essential to overcome the time-temperature dilemma.

Uncovering fast relaxation dynamics

The amorphous solid (glass) inherently undergoes relaxation processes toward lower-energy states, accompanied by changes in structure and almost all physical properties such as entropy, density, diffusion, and electrical resistivity. A particularly relevant process is the so-called Johari-Goldstein (β-) relaxation, which depicts local fast atomic motions involving a group of atoms in the low-mobility matrix.

β-relaxations have been extensively discussed in metallic, molecular and polymer glasses due to their relevance for many important properties. In metallic glasses, they are responsible for the fastest atomic diffusion. They promote accelerated aging and rapid crystallization. These phenomena are of particular interest for memory applications of PCMs.

Characterizing β-relaxations in PCMs was difficult due to the strict requirements of sample geometry which are hard to meet for poor glass-forming PCMs. Thanks to the recently developed powder approach, we have succeeded in uncovering β-relaxations in amorphous PCMs such as GeTe, Ge2Sb2Te5 and AIST. By contrast, β-relaxations are vanishingly small in amorphous chalcogenides of similar compositions such as GeSe, Ge15Te85 and GeSe2 which have rather slow crystallization kinetics making them unsuitable for memory applications (i.e. non- PCMs). The striking difference in relaxation dynamics raises the urgent questions how β- relaxations play a role in the stabilities of amorphous phases and crystallization behaviors in PCMs, and what is the structural origin of the β-relaxation processes in PCMs.


Amorphous metals (also called metallic glasses) are advanced metallurgical materials with disordered (amorphous) structures and offer the alluring combination of a high hardness, large elastic limit and the ability to be molded like thermo-plastics. They can be more durable and tougher than conventional alloys, while having the potential for a light weight, bio-compatibility and high corrosion resistance. All this makes metallic glasses particularly attractive for addressing societal challenges regarding energy efficiency, health and sustainability.

3D-printing of Amorphous Metals

The glass-forming ability has been a long-standing issue, which limits the size of the fully amorphous metals processed by conventional quenching and casting techniques.

In a very recent development, 3D-Printing or Additive Manufacturing has been used as a powerful tool to process metallic glasses. We are collaborating with the research unit of Heraeus Holding GmbH, to understand the structure-and-property relations in fully amorphous large-size complex-shaped metallic parts.

Characterizing defects in 3D-printed Amorphous Metals

In the 3D printing (i.e. powder bed fusion) process, pores may result from incorrect powder deposition or inconsistent laser energy density and scan rates. Under external loading, pores may induce high stress-concentrations and facilitate crack initiation in the microstructure of materials, which leads to an uncontrollable variation of quality and reliability for  powder bed fusion -processed parts. Thus, characterizing the structural defects and understanding their relation to mechanical properties are crucial for designing materials. Using X-ray micro-CT, we can reconstruct the 3D pore structures, which reveal clear structural heterogeneities with obvious differences in pore shape and porosity.

Modelling structural defects and shear bands

Finite element modelling shows the interaction between structural defects and shear band in 3D-printed amorphous metals. Under tensile loading, the formation of shear transformation zones is observed, which initiated at weak points around the pores and connected to each to form large shear bands.

Brilliant X-rays shed light on the structure of amorphous metals

While phase-change materials crystallize in a few nanoseconds, bulk metallic glasses are designed to be good glassformers with extraordinary mechanical properties. It was little known how bulk metallic glasses can retain amorphous structure and gain a high viscosity unlike those conventional metals during vitrification. Our research characterized thermodynamic and kinetic properties of these materials and used high-energy synchrotron X-rays to observe the in-situ structural changes of highly viscous metallic glass-forming liquids on the atomic-scale. We developed an empirical model that links the kinetic properties (e.g. liquid fragility) of bulk metallic glasses to the temperature dependence of their atomic-scale structural evolution. In addition, we took an approach that combines an electrostatic levitator with synchrotron X-ray scattering, which allowed us to in-situ observe the atomic structural evolution of a metallic droplet during the entire vitrification process by avoiding heterogeneous nucleation. It revealed that Zr-based metallic liquids undergo a transition in the medium-range-order structure just before vitrification, leading to highly viscous behavior in the BMGs.