arXiv subjects

Dario Alfè

Publications and source records attributed to Dario Alfè.

46 records · Page 3Linked to original sources

arXiv2014

On the Accuracy of van der Waals Inclusive Density-Functional Theory Exchange-Correlation Functionals for Ice at Ambient and High Pressures

Density-functional theory (DFT) has been widely used to study water and ice for at least 20 years. However, the reliability of different DFT exchange-correlation (xc) functionals for water remains a matter of considerable debate. This is particularly true in light of the recent development of DFT based methods that account for van der Waals (vdW) dispersion forces. Here, we report a detailed study with several xc functionals (semi-local, hybrid, and vdW inclusive approaches) on ice Ih and six proton ordered phases of ice. Consistent with our previous study [Phys. Rev. Lett. 107, 185701 (2011)] which showed that vdW forces become increasingly important at high pressures, we find here that all vdW inclusive methods considered improve the relative energies and transition pressures of the high-pressure ice phases compared to those obtained with semi-local or hybrid xc functionals. However, we also find that significant discrepancies between experiment and the vdW inclusive approaches remain in the cohesive properties of the various phases, causing certain phases to be absent from the phase diagram. Therefore, room for improvement in the description of water at ambient and high pressures remains and we suggest that because of the stern test the high pressure ice phases pose they should be used in future benchmark studies of simulation methods for water.

cond-mat.mtrl-sci↗

arXiv2014

Benchmarking the performance of Density Functional Theory and Point Charge Force Fields in their Description of sI Methane Hydrate against Diffusion Monte Carlo

High quality reference data from diffusion Monte Carlo calculations are presented for bulk sI methane hydrate, a complex crystal exhibiting both hydrogen-bond and dispersion dominated interactions. The performance of some commonly used exchange-correlation functionals and all-atom point charge force fields is evaluated. Our results show that none of the exchange-correlation functionals tested are sufficient to describe both the energetics and the structure of methane hydrate accurately, whilst the point charge force fields perform badly in their description of the cohesive energy but fair well for the dissociation energetics. By comparing to ice Ih, we show that a good prediction of the volume and cohesive energies for the hydrate relies primarily on an accurate description of the hydrogen bonded water framework, but that to correctly predict stability of the hydrate with respect to dissociation to ice Ih and methane gas, accuracy in the water-methane interaction is also required. Our results highlight the difficulty that density functional theory faces in describing both the hydrogen bonded water framework and the dispersion bound methane.

cond-mat.mtrl-sci↗

arXiv2013

Melting curve of face-centred-cubic nickel from first principles calculations

The melting curve of Ni up to 100 GPa has been calculated using first principles methods based on density functional theory (DFT). We used two complementary approaches: i) coexistence simulations with a reference system and then free energy corrections between DFT and the reference system, and ii) direct DFT coexistence using simulation cells including 1000 atoms. The calculated zero pressure melting temperature is slightly underestimated at $1637 \pm 10$ K (experimental value is 1728 K), and at high pressure is significantly higher than recent measurements in diamond anvil cell experiments [Phys. Rev. B {\bf 87}, 054108 (2013)]. The zero pressure DFT melting slope is calculated to be $30 \pm 2$ K, in good agreement with the experimental value of 28 K.

cond-mat.mtrl-sci↗

arXiv2012

Transport properties for liquid silicon-oxygen-iron mixtures at Earth's core conditions

We report on the thermal and electrical conductivities of two liquid silicon-oxygen-iron mixtures (Fe$_{0.82}$Si$_{0.10}$O$_{0.08}$ and Fe$_{0.79}$Si$_{0.08}$O$_{0.13}$), representative of the composition of the Earth's outer core at the relevant pressure-temperature conditions, obtained from density functional theory calculations with the Kubo-Greenwood formulation. We find thermal conductivities $k$ =100 (160) W m$^{-1}$ K$^{-1}$, and electrical conductivities $σ= 1.1 (1.3) \times 10^6 Ω^{-1}$ m$^{-1}$ at the top (bottom) of the outer core. These new values are between 2 and 3 times higher than previous estimates, and have profound implications for our understanding of the Earth's thermal history and the functioning of the Earth's magnetic field, including rapid cooling rate for the whole core or high level of radiogenic elements in the core. We also show results for a number of structural and dynamic properties of the mixtures, including the partial radial distribution functions, mean square displacements, viscosities and speeds of sound.

cond-mat.mtrl-sci↗

arXiv2012

Thermal and electrical conductivity of iron at Earth's core conditions

The Earth acts as a gigantic heat engine driven by decay of radiogenic isotopes and slow cooling, which gives rise to plate tectonics, volcanoes, and mountain building. Another key product is the geomagnetic field, generated in the liquid iron core by a dynamo running on heat released by cooling and freezing to grow the solid inner core, and on chemical convection due to light elements expelled from the liquid on freezing. The power supplied to the geodynamo, measured by the heat-flux across the core-mantle boundary (CMB), places constraints on Earth's evolution. Estimates of CMB heat-flux depend on properties of iron mixtures under the extreme pressure and temperature conditions in the core, most critically on the thermal and electrical conductivities. These quantities remain poorly known because of inherent difficulties in experimentation and theory. Here we use density functional theory to compute these conductivities in liquid iron mixtures at core conditions from first principles- the first directly computed values that do not rely on estimates based on extrapolations. The mixtures of Fe, O, S, and Si are taken from earlier work and fit the seismologically-determined core density and inner-core boundary density jump. We find both conductivities to be 2-3 times higher than estimates in current use. The changes are so large that core thermal histories and power requirements must be reassessed. New estimates of adiabatic heat-flux give 15-16 TW at the CMB, higher than present estimates of CMB heat-flux based on mantle convection; the top of the core must be thermally stratified and any convection in the upper core driven by chemical convection against the adverse thermal buoyancy or lateral variations in CMB heat flow. Power for the geodynamo is greatly restricted and future models of mantle evolution must incorporate a high CMB heat-flux and explain recent formation of the inner core.

physics.geo-ph↗

arXiv2012

High-temperature behaviour of supported graphene: electron-phonon coupling and substrate-induced doping

One of the salient features of graphene is the very high carrier mobility that implies tremendous potential for use in electronic devices. Unfortunately, transport measurements find the expected high mobility only in freely suspended graphen. When supported on a surface, graphene shows a strongly reduced mobility, and an especially severe reduction for temperatures above 200 K. A temperature-dependent mobility reduction could be explained by scattering of carriers with phonons, but this is expected to be weak for pristine, weakly-doped graphene. The mobility reduction has therefore been ascribed to the interaction with confined ripples or substrate phonons. Here we study the temperature-dependent electronic structure of supported graphene by angle-resolved photoemission spectroscopy, a technique that can reveal the origin of the phenomena observed in transport measurements. We show that the electron-phonon coupling for weakly-doped, supported graphene on a metal surface is indeed extremely weak, reaching the lowest value ever reported for any material. However, the temperature-dependent dynamic interaction with the substrate leads to a complex and dramatic change in the carrier type and density that is relevant for transport. Using ab initio molecular dynamics simulations, we show that these changes in the electronic structure are mainly caused by fluctuations in the graphene-substrate distance.

cond-mat.mes-hall↗

arXiv2011

Hydrogen bonds and van der Waals forces in ice at ambient and high pressures

The first principles approaches, density functional theory (DFT) and quantum Monte Carlo, have been used to examine the balance between van der Waals (vdW) forces and hydrogen (H) bonding in ambient and high pressure phases of ice. At higher pressure, the contribution to the lattice energy from vdW increases and that from H bonding decreases, leading vdW to have a substantial effect on the transition pressures between the crystalline ice phases. An important consequence, likely to be of relevance to molecular crystals in general, is that transition pressures obtained from DFT functionals which neglect vdW forces are greatly overestimated.

cond-mat.mtrl-sci↗

arXiv2009

Temperature of the inner-core boundary of the Earth: Melting of iron at high pressure from first-principles coexistence simulations

The Earth's core consists of a solid ball with a radius of 1221 Km, surrounded by a liquid shell which extends up to 3480 Km from the centre of the planet, roughly half way towards the surface (the mean radius of the Earth is 6373 km). The main constituent of the core is iron, and therefore the melting temperature of iron at the pressure encountered at the boundary between the solid and the liquid (the ICB) provides an estimate of the temperature of the core. Here I report the melting temperature of Fe at pressures near that of the ICB, obtained with first principles techniques based on density functional theory. The calculations have been performed by directly simulating solid and liquid iron in coexistence, and show that and at a pressure of $\sim 328$ GPa iron melts at $\sim 6370\pm 100$ K. These findings are in good agreement with earlier simulations, which used exactly the same quantum mechanics techniques, but obtained melting properties from the calculation of the free energies of solid and liquid Fe

cond-mat.mtrl-sci↗

arXiv2005

The melting curve of MgO from first principles simulations

First principles calculations based on density functional theory, both with the local density approximation (LDA) and with generalised gradient corrections (GGA), and the projector augmented wave method, have been used to simulate solid and liquid MgO in direct coexistence in the range of pressure $0\le p \le 135$ GPa. The calculated LDA zero pressure melting temperature is $ T_m^{\rm LDA} = 3110 \pm 50 $ K, in good agreement with the experimental data. The calculated GGA zero pressure melting temperature $ T_m^{\rm GGA} = 2575 \pm 100 $ K is significantly lower than the LDA one, but the difference between the GGA and the LDA melting curves is greatly reduced at high pressure. The LDA calculated zero pressure melting slope is $dT/dp \sim 100$ K/GPa, which is more than three times higher than the currently available experimental one ((Zerr and Boehler, Nature {\bf 371}, 506 (1994)). As the pressure is increased, the melting curve deviates significantly from the experimental one. At the core mantle boundary pressure of 135 GPa MgO melts at $T_m = 8140 \pm 150$ K.

cond-mat.mtrl-sci↗

arXiv2003

First principles simulations of direct coexistence of solid and liquid aluminium

First principles calculations based on density functional theory, with generalised gradient corrections and ultrasoft pseudopotentials, have been used to simulate solid and liquid aluminium in direct coexistence at zero pressure. Simulations have been carried out on systems containing up to 1000 atoms for 15 ps. The points on the melting curve extracted from these simulations are in very good agreement with previous calculations, which employed the same electronic structure method but used an approach based on the explicit calculation of free energies [L. Vočadlo and D. Alfè, Phys. Rev. B, {\bf 65}, 214105 (2002).]

cond-mat.mtrl-sci↗