The study of crystal growth in phase-change thin films is of crucial importance to improve our understanding of the extraordinary phase transformation kinetics of these materials excellently suited for data storage applications. Here, we developed and used a new method, based on isothermal heating using laser illumination in combination with a high-speed optical camera, to measure the crystal growth rates, in a direct manner over 6 orders of magnitude, in phase-change thin films composed of several GeSb alloys. For Ge 8 Sb 92 and Ge 9 Sb 91 , a clear nonArrhenius temperature dependence for crystal growth was found that is described well on the basis of a viscosity model incorporating the fragility of the supercooled liquid as an important parameter. Using this model, the crystal growth rate can be described for the whole range between the glass transition temperature of about 380 K and the melting temperature of 880 K, excellently explaining that these phasechange materials show unique and remarkable behavior that they combine extremely low crystal growth rates at temperatures below 380 K required for 10 years of data retention and very fast growth rates of 15 m s −1 at temperatures near the melting point required for bit switching within tens of nanoseconds. ■ INTRODUCTIONCrystallization of phase-change materials is a temperatureactivated process, generally characterized by relatively large activation energies. 1−3 Methods for studying the crystal growth properties from room temperature up to and slightly above the glass-transition temperature are well established, and can be applied by, for example, placing a phase-change film on a hot plate or in a heating holder or furnace and using optical, atomic force, or electron microscopy to study the crystal growth. 4,5 Only microscopy-based techniques are able to directly determine crystal growth rates, while many other techniques such as resistance measurements, differential scanning calorimetry, X-ray diffraction, etc. can be used to monitor the overall crystallized fraction but cannot make a distinction between nucleation and growth. These techniques are generally also combined with isochronal (for a range of heating rates) instead of isothermal (for a range of temperatures) measurements making the analysis of the crystallization kinetics more approximate. However, a drawback of isothermal measurements is that they generally limit the maximum growth rate that can be accurately measured in situ. The reason for this limitation is mostly imposed by the (relatively slow) heating rates used to reach the isothermal temperature in combination with the requirement that crystallization only starts when the sample has become stabilized at the isothermal temperature.Here we demonstrate our work observing crystal growth at higher temperatures, and thus higher growth rates, by using a laser to additionally heat the film and by monitoring the growth using a high speed optical camera. As explained above, the benefit of optical microscopy over other methods to measure crystal grow...

Abstract-Analyzing data centers with thermal-aware optimization techniques is a viable approach to reduce energy consumption of data centers. By taking into account thermal consequences of job placements among the servers of a data center, it is possible to reduce the amount of cooling necessary to keep the servers below a given safe temperature threshold. We set up an optimization problem to analyze and characterize the optimal setpoints for the workload distribution and the supply temperature of the cooling equipment. Furthermore under mild assumptions we design and analyze controllers that drive the data center to the optimal state without knowledge of the current total workload to be handled by the data center. The response of our controller is validated by simulations and convergence to the optimal setpoints is achieved under varying workload conditions.

No abstract

Analyzing data centers with thermal-aware optimization techniques is a viable approach to reduce energy consumption of data centers. By taking into account thermal consequences of job placements among the servers of a data center, it is possible to reduce the amount of cooling necessary to keep the servers below a given safe temperature threshold. We set up an optimization problem to analyze and characterize the optimal setpoints for the workload distribution and the supply temperature of the cooling equipment. Furthermore under mild assumptions we design and analyze controllers that drive the data center to the optimal state without knowledge of the current total workload to be handled by the data center. The response of our controller is validated by simulations and convergence to the optimal setpoints is achieved under varying workload conditions.

No abstract