The properties of elemental metals under ultra high pressure and temperature conditions are a subject of continuous interest. A salient example is iron, the main constituent of the Earth’s core; its physical properties between 140 GPa and 360 GPa are needed to build a reliable model of the Earth’s interior. Apart from traditional static (laser-heated diamond anvil cell [1]) and dynamic (gun or laser induced shock waves [2]) pressurizing techniques, new experimental concepts have appeared [3-5] which allow access to a broad range of thermodynamic states. On the theoretical side, a variety of techniques have been developed which aim at accurately predicting the equilibrium state of systems under HP/HT or even at directly simulating real dynamic compression or release paths at the atomic scale. The most accurate of these approaches are based on density functional theory or Quantum Monte Carlo [6], and various statistical mechanics approaches to phase transitions. Because of the small energy scales involved, high accuracy is required for the methods to be usefully predictive.

Yet, at the same time, unacceptable discrepancies remain on an apparently simple property as melting. The experimental melting line of iron, but also tantalum, molybdenum, tungsten… differ by up to 6000K around 300 GPa [2,7,8] for the worst case, depending on the compression technique (static/dynamic [2,7,8]), but also the melting diagnostic (optical/X-ray for static compression [9]). Indeed, contradictory reports have been/will be presented at different general high pressure conferences in the last two years: EHPRG2009 – 2010 High pressure Gordon Conference – APS SCCM 2011 – Z-CAM workshop 2011. Several models have been proposed to reconcile all observations, which mainly involve an additional phase transition [7,10,11], but none appears unambiguously conclusive. So far, only a few works have attempted to take into account the specificity of experiments to explain the differences between observations (e.g. shear in diamond anvil cell, [12]). We believe that such a strategy, implying a systematic critical analysis, should be pushed forward vigorously, and that metrology and diagnostics should be questioned as well, in close interplay between theoreticians and experimentalists. 

Our goal is to organize a focused meeting which will help to address the ambiguities and technical challenges which have resulted in these controversies. Enhanced collaborations between theoreticians and experimentalists are required if significant progress is expected.  Experimental diagnostics: temperature and pressure measurements and melting criteria need to be improved and unified. For that purpose, all materials constituting the high pressure device and their interactions have to be studied. Their transport, electronic, optical and mechanical properties have to be determined. Important issues in current simulation methods have also to be discussed: how to be really predictive at high P and high T?

The properties of elemental metals under ultra high pressure and temperature conditions are a subject of continuous interest. A salient example is iron, the main constituent of the Earth’s core; its physical properties between 140 GPa and 360 GPa are needed to build a reliable model of the Earth’s interior. Apart from traditional static (laser-heated diamond anvil cell [1]) and dynamic (gun or laser induced shock waves [2]) pressurizing techniques, new experimental concepts have appeared [3-5] which allow access to a broad range of thermodynamic states. On the theoretical side, a variety of techniques have been developed which aim at accurately predicting the equilibrium state of systems under HP/HT or even at directly simulating real dynamic compression or release paths at the atomic scale. The most accurate of these approaches are based on density functional theory or Quantum Monte Carlo [6], and various statistical mechanics approaches to phase transitions. Because of the small energy scales involved, high accuracy is required for the methods to be usefully predictive.
Yet, at the same time, unacceptable discrepancies remain on an apparently simple property as melting. The experimental melting line of iron, but also tantalum, molybdenum, tungsten… differ by up to 6000K around 300 GPa [2,7,8] for the worst case, depending on the compression technique (static/dynamic [2,7,8]), but also the melting diagnostic (optical/X-ray for static compression [9]). Indeed, contradictory reports have been/will be presented at different general high pressure conferences in the last two years: EHPRG2009 – 2010 High pressure Gordon Conference – APS SCCM 2011 – Z-CAM workshop 2011. Several models have been proposed to reconcile all observations, which mainly involve an additional phase transition [7,10,11], but none appears unambiguously conclusive. So far, only a few works have attempted to take into account the specificity of experiments to explain the differences between observations (e.g. shear in diamond anvil cell, [12]). We believe that such a strategy, implying a systematic critical analysis, should be pushed forward vigorously, and that metrology and diagnostics should be questioned as well, in close interplay between theoreticians and experimentalists. 
Our goal is to organize a focused meeting which will help to address the ambiguities and technical challenges which have resulted in these controversies. Enhanced collaborations between theoreticians and experimentalists are required if significant progress is expected.  Experimental diagnostics: temperature and pressure measurements and melting criteria need to be improved and unified. For that purpose, all materials constituting the high pressure device and their interactions have to be studied. Their transport, electronic, optical and mechanical properties have to be determined. Important issues in current simulation methods have also to be discussed: how to be really predictive at high P and high T?

Various statistical mechanics methods have been used in the past few decades to theoretically address melting properties, and are now reaching maturity. It is time for an in depth analysis of these methods. These include: 

i) thermodynamic integration; 

ii) phase coexistence within molecular dynamics either classical or First Principles MD; iii) MD homogeneous melting and Z method. 

It is expected that, when applied consistently, all these methods should deliver the same results. However, significant discrepancies persist in the literature (e.g. the case of molybdenum [10,18,19]). One of the important objectives of the proposed workshop will be to bring together computational physicists to confront these issues. Specific focus will be given to accuracy needed on free energy calculation, and to time and length scale issues in molecular dynamics calculations. 

Among current challenges of HP-HT experiments we note the following: 

i) Extract reliable temperatures from pyrometry measurements in shock and static compression devices; this involves a precise knowledge of thermal and optical properties of materials around the shock front/laser heated area (sample and pressure medium or window). Theoretical work that will directly apply to the interpretation of experimental measurements (and enable to get rid of a black or grey body hypothesis) needs to be performed, such as dynamic emissivity calculations. New experiments which aim at determining these properties need to be performed. We intend to devote an entire session to this issue, which seems to us extremely important. In the absence of accurate temperature measurements, it is risky to plot the results of dynamic experiments on the phase diagrams derived from static compression experiments.

ii) Set standard diagnostics for melting in shock and static compression experiments. To date no objective and unified bulk melting diagnostic exists. Shock experiments measure a change in sound velocity which is interpreted as a change in shear modulus, a bulk property.  X-rays scan the crystallographic order at a smaller scale; an unambiguous melting criterion is the recording of the signal scattered by the liquid sample [9,13], but due to its weakness it can be only rarely recorded. As a result, different criteria measured with different time scales have been used by different groups (with x-ray: recrystallization at the second-scale [14] or disappearance of the signal; temperature plateaux [14]; change of the sample shape or optical properties [7]). Similar observations have led to different conclusions. They urgently need to be confronted and unified; transition metals, on which several groups have data (as a sample or a laser absorber), are the ideal candidates. 

iii) Develop new diagnostics/models for dynamic compression experiments. Dynamic compression experiments can now reach off -Hugoniot states, for instance through ramp-loading paths and/or using pre-heated/compressed samples [3].Time-resolved diagnostics have opened new perspectives to characterize transitory and transport phenomena. For instance, crystallization [4,15], transition paths, thermal conduction [5], mechanical properties [16] have been studied recently. These phenomena are fundamental to understand the observations made in traditional high pressure experiments: for example, thermal conduction governs the temperature distribution in laser-heated diamond anvil cells. Moreover, the interplay between experiments and large MD simulations [17] can help to understand the response of metals to stress and temperature and to address the role of kinetic effects in dynamic experiments. 

We need to develop optical or x-ray methods to determine crystal structure in dynamic experiments that are time-resolved, that can determine low-symmetry structures, and work in gun experiments, i.e. with single direction access.