Materials World Network - Understanding and exploiting mixed-mode ultra-fast optical-electrical behavior in nanoscale phase change materials

Phase-change materials, such as GeSbTe or AgInSbTe alloys, exhibit some remarkable properties; they can be amorphized in femtoseconds and crystallised in picoseconds, yet can remain stable against spontaneous changes of state for many years. They show hugely contrasting properties between phases, including an electrical conductivity difference of up to five orders of magnitude and a large refractive index change; properties that have led to their application in electrical (phase-change RAM or PCM devices) and optical (DVD and Blu-Ray disks) memories. The origin of such remarkable properties has been a source of much recent research. Kolobov showed that, contrary to conventional expectations, the short-range order in Ge2Sb2Te5 is higher in the amorphous than in the crystal phase. This was explained by an 'umbrella flip' of Ge atoms from primarily tetrahedral to octahedral bonding in the amorphous to crystalline transition, and was put forward as the potential origin of ultra-fast switching. While this simple 'umbrella-flip' model has since turned out not to be a truly realistic model of the phase-transition, and cannot explain the behavior of phase-materials that do not contain germanium, it sparked a world-wide 'quest' for an accurate understanding of the nature of switching processes in this important class of materials. Part of the answer was revealed by the 'discovery' that the crystalline phase of phase-change alloys is also rather unusual, exhibiting strong resonance bonding, with such bonding being suggested as a 'necessary condition' for technologically useful phase-change properties. Most recently a metal-insulator type disorder induced transition in the crystalline phase has also been reported, and it has also been suggested that distortions in the crystalline phase may trigger a collapse of long-range order, generating the amorphous phase without going through the liquid state.
The scientific and technological importance of phase-change materials is clearly extremely high. However, many of their remarkable properties remain poorly understood, and the ways in which such properties might be exploited to deliver exciting applications going way beyond simple binary memories is largely 'uncharted territory'. For example we have, very recently, shown that by crystallizing GeSbTe alloys using femtosecond optical pulses we can perform reliable arithmetic processing, so providing a form of 'phase-change processor', Furthermore, we showed that a fundamental advantage of phase-change materials over other common electronics materials is that they have readily accessible and usable electrical and optical responses, and signals can be transferred relatively simply between these two domains. This mixed-mode behavior of phase-change materials provides a (as yet unused) powerful means to understand the fundamental switching properties of these materials. There are also several potentially very important applications of mixed-mode behavior, such as ultra-fast optically-gated switching for example (or, more speculatively, optically-active memristors - or 'memflectors'). However, this mixed-mode behavior of phase-change materials has never before been explored. Our proposal therefore combines a new route to addressing key scientific questions that remain unanswered, along with an exploration of entirely new ways in which to exploit the remarkable properties of phase-change materials; specifically we ask:
1. exactly how fast are these phase-change (crystallization and amorphization) processes?
2. does amorphization always involve melting in phase-change materials?
3. what are the precise dynamics of switching events; are they different in optically-excited and electrically excited cases; do they remain the same on the nanocale?
4. what are the key materials drivers for ultra-fast switching?
5. can we scale mixed-mode behavior to the nanoscale?
6. can we exploit mixed-mode behavior to provide advanced functionality?

A recent IDC study estimated that the amount of data stored globally by the end of 2011 will be 80% higher than in 2010 (1800 Exabytes cf. 1000 Exabytes), and that by 2015 there will be a staggering 700% increase in the amount of data stored in comparison to 2010. Much of this memory requirement is currently met by magnetic hard disk (HDD) and CMOS Flash memories, and the current total market size for HDD and (NAND) Flash is estimated to be around $50 billion in terms of revenue. Unfortunately, these current mass data storage technologies face difficult technical barriers to progress in order to meet this exponentially increasing storage requirement. These limitations come about due to the well-known super-paramagnetic limit in conventional magnetic recording and due to cell scaling problems in floating-gate Flash memories. Research is of course ongoing to try to overcome these technical barriers, but progress is difficult and costly, and the time is ripe for new, high-performance memory technologies to emerge.
An equally important consideration is the sustainability of today's ICT. Electricity produced to power the world's data centres generates greenhouse gases on the scale of the outputs of entire countries such as the Netherlands or Argentina, and these emissions could increase fourfold by 2020 if no action is taken. Memory is the number one power consumer in servers today, and the ratio of memory power use to processor power use is growing. Clearly therefore memories that consume significantly less power than today's HDDs and NAND Flash memories are desperately needed.
There is thus a pressing need and an exciting opportunity for the development of new data storage/memory materials and technologies. The general requirements of any new approach should include non-volatility, solid-state implementation (no moving parts), low write/read latency, high endurance, very low cost per bit and excellent scalability to future technology nodes. Resistive phase-change memory materials offer just such capabilities, and it is therefore most timely and pertinent to concentrate research efforts in this area. Although the first generation of electrical phase-change memories are close to commercialization, for them to make a real long-term impact a number of key breakthroughs are needed - in power consumption, in speed and in storage density. The key issues are all materials-related, and are comprehensively addresssed by our MNW project. By fundamental studies of the dynamics of basic switching mechanism we will gain key insights into exactly how fast these materials really are. By reducing materials dimensions down to true nanometre level we will determine exactly how far we can really scale. By significantly increasing the speed of the switching operation, and by significantly scaling down the size of the switching volume, we can drastically reduce the power requirements.
However, the use of phase-change materials for simple binary memories barely begins to exploit their remarkable properties to the full. Phase-change materials should also be capable of non-binary arithmetic processing, multi-value logic and bio-inspired type processing. Phase-change materials also have an untapped potential for mixed-mode optical-electrical operation, that may lead to new devices and applications (such as ultra-fast optically triggered gates). We will also investigate such exciting possibilities as part of our MNW project.
Thus, our work is likely to have far-reaching impact on academic and industrial research and development in fields ranging from materials to electronics to photonics and computing. The potential economic, environmental and societal impacts of our work are also very significant; it may eventually lead to exploitation opportunities for new, improved and 'added functionality' phase-change materials and devices; it may lead to memory systems with much reduced power consumption; it may lead to entirely new ways of computing.